db/pgbackrest
1
0
Fork 0
mirror of https://github.com/pgbackrest/pgbackrest.git synced 2024-12-14 10:13:05 +02:00
Code Issues Releases Activity
9639a2c15f
pgbackrest/test/expect/mock-all-001.log

3212 lines
232 KiB
Plaintext
Raw Normal View History

Create test matrix for mock/all to increase coverage and reduce tests. The same test configurations are run on all four test VMs, which seems a real waste of resources. Vary the tests per VM to increase coverage while reducing the total number of tests. Be sure to include each major feature (remote, s3, encryption) in each VM at least once.
2019-03-02 15:01:02 +02:00
run 001 - rmt 0, s3 0, enc 0, delta 1
=====================================
Log testing can now be enabled for certain deterministic tests. This works by comparing the generated logs against a previous copy. Currently only enabled for the backup/synthetic tests.
2015-05-05 19:08:48 +02:00
Move help/version integration tests to mock/all. Help and version are covered by unit tests, so we really just to need to make sure there is output when called from the command line.
2018-04-08 18:16:32 +02:00
> [CONTAINER-EXEC] db-master [BACKREST-BIN] version
------------------------------------------------------------------------------------------------------------------------------------
[BACKREST-NAME-VERSION]
> [CONTAINER-EXEC] db-master [BACKREST-BIN] help version
------------------------------------------------------------------------------------------------------------------------------------
[BACKREST-NAME-VERSION] - 'version' command help
Get version.
Displays installed pgBackRest version.
Improved stanza-create command so that it can repair broken repositories in most cases and is robust enough to be made mandatory. Contributed by Cynthia Shang.
2016-12-20 23:52:20 +02:00
stanza-create db - create required data for stanza (db-master host)
Reduce log level for all expect tests to detail. The C code is designed to be efficient rather than deterministic at the debug log level. As we move more testing from integration to unit tests it makes less sense to try and maintain the expect logs at this log level. Most of the expect logs have already been moved to detail level but mock/all still had tests at debug level. Change the logging defaults in the config file and remove as many references to log-level-console as possible.
2019-05-23 00:23:44 +02:00
> [CONTAINER-EXEC] db-master [BACKREST-BIN] --config=[TEST_PATH]/db-master/pgbackrest.conf --stanza=db --no-online stanza-create
Improved stanza-create command so that it can repair broken repositories in most cases and is robust enough to be made mandatory. Contributed by Cynthia Shang.
2016-12-20 23:52:20 +02:00
------------------------------------------------------------------------------------------------------------------------------------
Remove compress/compress-level options from commands where unused. These commands (e.g. restore, archive-get) never used the compress options but allowed them to be passed on the command line. Now they will error when these options are passed on the command line. If these errors occur then remove the unused options.
2020-02-27 19:25:32 +02:00
P00 INFO: stanza-create command begin [BACKREST-VERSION]: --buffer-size=[BUFFER-SIZE] --config=[TEST_PATH]/db-master/pgbackrest.conf --db-timeout=45 --lock-path=[TEST_PATH]/db-master/lock --log-level-console=detail --log-level-file=[LOG-LEVEL-FILE] --log-level-stderr=off --log-path=[TEST_PATH]/db-master/log[] --no-log-timestamp --no-online --pg1-path=[TEST_PATH]/db-master/db/base --protocol-timeout=60 --repo1-path=[TEST_PATH]/db-master/repo --stanza=db
Improved stanza-create command so that it can repair broken repositories in most cases and is robust enough to be made mandatory. Contributed by Cynthia Shang.
2016-12-20 23:52:20 +02:00
P00 INFO: stanza-create command end: completed successfully
+ supplemental file: [TEST_PATH]/db-master/repo/backup/db/backup.info
---------------------------------------------------------------------
[backrest]
backrest-format=5
backrest-version="[VERSION-1]"
[db]
db-catalog-version=201409291
db-control-version=942
db-id=1
Automate generation of WAL and pg_control test files. The existing static files would not work with 32-bit or big-endian systems so create functions to generate these files dynamically rather than creating a bunch of new static files.
2017-11-19 03:02:54 +02:00
db-system-id=1000000000000000094
Improved stanza-create command so that it can repair broken repositories in most cases and is robust enough to be made mandatory. Contributed by Cynthia Shang.
2016-12-20 23:52:20 +02:00
db-version="9.4"
[db:history]
Automate generation of WAL and pg_control test files. The existing static files would not work with 32-bit or big-endian systems so create functions to generate these files dynamically rather than creating a bunch of new static files.
2017-11-19 03:02:54 +02:00
1={"db-catalog-version":201409291,"db-control-version":942,"db-system-id":1000000000000000094,"db-version":"9.4"}
Improved stanza-create command so that it can repair broken repositories in most cases and is robust enough to be made mandatory. Contributed by Cynthia Shang.
2016-12-20 23:52:20 +02:00
Improve performance of info file load/save. Info files required three copies in memory to be loaded (the original string, an ini representation, and the final info object). Not only was this memory inefficient but the Ini object does sequential scans when searching for keys making large files very slow to load. This has not been an issue since archive.info and backup.info are very small, but it becomes a big deal when loading manifests with hundreds of thousands of files. Instead of holding copies of the data in memory, use a callback to deliver the ini data directly to the object when loading. Use a similar method for save to avoid having an intermediate copy. Save is a bit complex because sections/keys must be written in alpha order or older versions of pgBackRest will not calculate the correct checksum. Also move the load retry logic to helper functions rather than embedding it in the Info object. This allows for more flexibility in loading and ensures that stack traces will be available when developing unit tests. Reviewed by Cynthia Shang.
2019-09-06 19:48:28 +02:00
[backrest]
backrest-checksum="[CHECKSUM]"
Improved stanza-create command so that it can repair broken repositories in most cases and is robust enough to be made mandatory. Contributed by Cynthia Shang.
2016-12-20 23:52:20 +02:00
+ supplemental file: [TEST_PATH]/db-master/repo/archive/db/archive.info
-----------------------------------------------------------------------
[backrest]
backrest-format=5
backrest-version="[VERSION-1]"
[db]
db-id=1
Automate generation of WAL and pg_control test files. The existing static files would not work with 32-bit or big-endian systems so create functions to generate these files dynamically rather than creating a bunch of new static files.
2017-11-19 03:02:54 +02:00
db-system-id=1000000000000000094
Improved stanza-create command so that it can repair broken repositories in most cases and is robust enough to be made mandatory. Contributed by Cynthia Shang.
2016-12-20 23:52:20 +02:00
db-version="9.4"
[db:history]
Automate generation of WAL and pg_control test files. The existing static files would not work with 32-bit or big-endian systems so create functions to generate these files dynamically rather than creating a bunch of new static files.
2017-11-19 03:02:54 +02:00
1={"db-id":1000000000000000094,"db-version":"9.4"}
Improved stanza-create command so that it can repair broken repositories in most cases and is robust enough to be made mandatory. Contributed by Cynthia Shang.
2016-12-20 23:52:20 +02:00
Improve performance of info file load/save. Info files required three copies in memory to be loaded (the original string, an ini representation, and the final info object). Not only was this memory inefficient but the Ini object does sequential scans when searching for keys making large files very slow to load. This has not been an issue since archive.info and backup.info are very small, but it becomes a big deal when loading manifests with hundreds of thousands of files. Instead of holding copies of the data in memory, use a callback to deliver the ini data directly to the object when loading. Use a similar method for save to avoid having an intermediate copy. Save is a bit complex because sections/keys must be written in alpha order or older versions of pgBackRest will not calculate the correct checksum. Also move the load retry logic to helper functions rather than embedding it in the Info object. This allows for more flexibility in loading and ensures that stack traces will be available when developing unit tests. Reviewed by Cynthia Shang.
2019-09-06 19:48:28 +02:00
[backrest]
backrest-checksum="[CHECKSUM]"
Major refactor of the test suite. * Make the code more modular and object-oriented. * Multiple Docker containers can now be created for a single test to simulate more realistic environments.
2016-06-24 14:12:58 +02:00
full backup - create pg_stat link, pg_clog dir (db-master host)
Integration test improvements for disk and memory efficiency. Set log-level-file=off when more that one test will run. In this case is it impossible to see the logs anyway since they will be automatically cleaned up after the test. This improves performance pretty dramatically since trace-level logging is expensive. If a singe integration test is run then log-level-file is trace by default but can be changed with the --log-level-test-file option. Reduce buffer-size to 64k to save memory during testing and allow more processes to run in parallel. Update log replacement rules so that these options can change without affecting expect logs.
2019-12-17 22:23:07 +02:00
> [CONTAINER-EXEC] db-master [BACKREST-BIN] --config=[TEST_PATH]/db-master/pgbackrest.conf --no-online --manifest-save-threshold=3 --buffer-size=[BUFFER-SIZE] --checksum-page --process-max=1 --repo1-type=cifs --type=full --stanza=db backup
Log testing can now be enabled for certain deterministic tests. This works by comparing the generated logs against a previous copy. Currently only enabled for the backup/synthetic tests.
2015-05-05 19:08:48 +02:00
------------------------------------------------------------------------------------------------------------------------------------
Add infrastructure for multiple compression type support. Add compress-type option and deprecate compress option. Since the compress option is boolean it won't work with multiple compression types. Add logic to cfgLoadUpdateOption() to update compress-type if it is not set directly. The compress option should no longer be referenced outside the cfgLoadUpdateOption() function. Add common/compress/helper module to contain interface functions that work with multiple compression types. Code outside this module should no longer call specific compression drivers, though it may be OK to reference a specific compression type using the new interface (e.g., saving backup history files in gz format). Unit tests only test compression using the gz format because other formats may not be available in all builds. It is the job of integration tests to exercise all compression types. Additional compression types will be added in future commits.
2020-03-06 21:41:03 +02:00
P00 INFO: backup command begin [BACKREST-VERSION]: --buffer-size=[BUFFER-SIZE] --checksum-page --compress-level=3 --compress-type=none --config=[TEST_PATH]/db-master/pgbackrest.conf --db-timeout=45 --lock-path=[TEST_PATH]/db-master/lock --log-level-console=detail --log-level-file=[LOG-LEVEL-FILE] --log-level-stderr=off --log-path=[TEST_PATH]/db-master/log[] --no-log-timestamp --manifest-save-threshold=3 --no-online --pg1-path=[TEST_PATH]/db-master/db/base --process-max=1 --protocol-timeout=60 --repo1-path=[TEST_PATH]/db-master/repo --repo1-type=cifs --stanza=db --start-fast --type=full
Add time-based retention for full backups. The --repo-retention-full-type option allows retention of full backups based on a time period, specified in days. The new option will default to 'count' and therefore will not affect current installations. Setting repo-retention-full-type to 'time' will allow the user to use a time period, in days, to indicate full backup retention. Using this method, a full backup can be expired only if the time the backup completed is older than the number of days set with repo-retention-full (calculated from the moment the 'expire' command is run) and at least one full backup meets the retention period. If archive retention has not been configured, then the default settings will expire archives that are prior to the oldest retained full backup. For example, if there are three full backups ending in times that are 25 days old (F1), 20 days old (F2) and 10 days old (F3), then if the full retention period is 15 days, then only F1 will be expired; F2 will be retained because F1 is not at least 15 days old.
2020-05-08 21:25:03 +02:00
P00 WARN: option 'repo1-retention-full' is not set for 'repo1-retention-full-type=count', the repository may run out of space
Rename retention-* options to repo-retention-*.
2018-02-19 22:07:24 +02:00
HINT: to retain full backups indefinitely (without warning), set option 'repo1-retention-full' to the maximum.
Exclude temporary and unlogged relation (table/index) files from backup. Implemented using the same logic as the patches adding this feature to PostgreSQL, 8694cc96 and 920a5e50. Temporary relation exclusion is enabled in PostgreSQL ≥ 9.0. Unlogged relation exclusion is enabled in PostgreSQL ≥ 9.1, where the feature was introduced. Contributed by Cynthia Shang.
2018-07-31 00:53:34 +02:00
P01 INFO: backup file [TEST_PATH]/db-master/db/base/base/32768/33001 (64KB, 33%) checksum 6bf316f11d28c28914ea9be92c00de9bea6d9a6b
Added the --checksum-page option. This option allows pgBackRest to validate page checksums in data files when checksums are enabled on PostgreSQL >= 9.3. Note that this functionality requires a C library which may not initially be available in OS packages. The option will automatically be enabled when the library is present and checksums are enabled on the cluster.
2016-12-13 01:54:07 +02:00
P00 WARN: invalid page checksums found in file [TEST_PATH]/db-master/db/base/base/32768/33001 at pages 0, 3-5, 7
Fix incorrect offline upper bound for ignoring page checksum errors. For offline backups the upper bound was being set to 0x0000FFFF0000FFFF rather than UINT64_MAX. This meant that page checksum errors might be ignored for databases with a lot of past WAL in offline mode. Online mode is not affected since the upper bound is retrieved from pg_start_backup().
2019-07-11 15:13:56 +02:00
P01 INFO: backup file [TEST_PATH]/db-master/db/base/base/32768/44000_init (32KB, 49%) checksum 7a16d165e4775f7c92e8cdf60c0af57313f0bf90
P01 INFO: backup file [TEST_PATH]/db-master/db/base/base/32768/33000.32767 (32KB, 66%) checksum 6e99b589e550e68e934fd235ccba59fe5b592a9e
P01 INFO: backup file [TEST_PATH]/db-master/db/base/base/32768/33000 (32KB, 83%) checksum 7a16d165e4775f7c92e8cdf60c0af57313f0bf90
Exclude temporary and unlogged relation (table/index) files from backup. Implemented using the same logic as the patches adding this feature to PostgreSQL, 8694cc96 and 920a5e50. Temporary relation exclusion is enabled in PostgreSQL ≥ 9.0. Unlogged relation exclusion is enabled in PostgreSQL ≥ 9.1, where the feature was introduced. Contributed by Cynthia Shang.
2018-07-31 00:53:34 +02:00
P01 INFO: backup file [TEST_PATH]/db-master/db/base/base/16384/17000 (16KB, 91%) checksum e0101dd8ffb910c9c202ca35b5f828bcb9697bed
Added the --checksum-page option. This option allows pgBackRest to validate page checksums in data files when checksums are enabled on PostgreSQL >= 9.3. Note that this functionality requires a C library which may not initially be available in OS packages. The option will automatically be enabled when the library is present and checksums are enabled on the cluster.
2016-12-13 01:54:07 +02:00
P00 WARN: invalid page checksum found in file [TEST_PATH]/db-master/db/base/base/16384/17000 at page 1
Support configurable WAL segment size. PostgreSQL 11 introduces configurable WAL segment sizes, from 1MB to 1GB. There are two areas that needed to be updated to support this: building the archive-get queue and checking that WAL has been archived after a backup. Both operations require the WAL segment size to properly build a list. Checking the archive after a backup is still implemented in Perl and has an active database connection, so just get the WAL segment size from the database. The archive-get command does not have a connection to the database, so get the WAL segment size from pg_control instead. This requires a deeper inspection of pg_control than has been done in the past, so it seemed best to copy the relevant data structures from each version of PostgreSQL and build a generic interface layer to address them. While this approach is a bit verbose, it has the advantage of being relatively simple, and can easily be updated for new versions of PostgreSQL. Since the integration tests generate pg_control files for testing, teach Perl how to generate files with the correct offsets for both 32-bit and 64-bit architectures.
2018-09-25 11:24:42 +02:00
P01 INFO: backup file [TEST_PATH]/db-master/db/base/global/pg_control (8KB, 95%) checksum 4c77c900f7af0d9ab13fa9982051a42e0b637f6c
Add more realistic data files to synthetic backup and restore tests.
2016-12-10 16:13:55 +02:00
P01 INFO: backup file [TEST_PATH]/db-master/db/base/base/1/12000 (8KB, 99%) checksum 22c98d248ff548311eda88559e4a8405ed77c003
For simplicity, the pg_control file is now copied with the rest of the files instead of by itself of at the end of the process. The backup command does not require this behavior and the restore copies to a temporary file which is renamed at the end of the restore.
2016-11-30 21:23:05 +02:00
P01 INFO: backup file [TEST_PATH]/db-master/db/base/postgresql.conf (21B, 99%) checksum 6721d92c9fcdf4248acff1f9a1377127d9064807
Fix failure in manifest build when two or more files in PGDATA are linked to the same directory. Reported by Vitaliy Kukharik.
2018-05-02 18:19:54 +02:00
P01 INFO: backup file [TEST_PATH]/db-master/db/base/pg_hba.conf (9B, 99%) checksum dd4cea0cae348309f9de28ad4ded8ee2cc2e6d5b
For simplicity, the pg_control file is now copied with the rest of the files instead of by itself of at the end of the process. The backup command does not require this behavior and the restore copies to a temporary file which is renamed at the end of the restore.
2016-11-30 21:23:05 +02:00
P01 INFO: backup file [TEST_PATH]/db-master/db/base/pg_stat/global.stat (5B, 99%) checksum e350d5ce0153f3e22d5db21cf2a4eff00f3ee877
P01 INFO: backup file [TEST_PATH]/db-master/db/base/base/32768/PG_VERSION (3B, 99%) checksum 184473f470864e067ee3a22e64b47b0a1c356f29
P01 INFO: backup file [TEST_PATH]/db-master/db/base/base/16384/PG_VERSION (3B, 99%) checksum 184473f470864e067ee3a22e64b47b0a1c356f29
P01 INFO: backup file [TEST_PATH]/db-master/db/base/base/1/PG_VERSION (3B, 99%) checksum 184473f470864e067ee3a22e64b47b0a1c356f29
P01 INFO: backup file [TEST_PATH]/db-master/db/base/PG_VERSION (3B, 100%) checksum 184473f470864e067ee3a22e64b47b0a1c356f29
Use minio for integration testing. ScalityS3 has not received any maintenance in years and is slow to start which is bad for testing. Replace it with minio which starts quickly and ships as a single executable or a tiny container. Minio has stricter limits on allowable characters but should still provide enough coverage to show that our encoding is working correctly. This commit also includes the upgrade to openssl 1.1.1 in the Ubuntu 18.04 container.
2019-07-03 04:20:35 +02:00
P01 INFO: backup file [TEST_PATH]/db-master/db/base/special-!_.*'()&!@;:+,? (0B, 100%)
Exclude temporary and unlogged relation (table/index) files from backup. Implemented using the same logic as the patches adding this feature to PostgreSQL, 8694cc96 and 920a5e50. Temporary relation exclusion is enabled in PostgreSQL ≥ 9.0. Unlogged relation exclusion is enabled in PostgreSQL ≥ 9.1, where the feature was introduced. Contributed by Cynthia Shang.
2018-07-31 00:53:34 +02:00
P00 INFO: full backup size = 192KB
Show Process ID in log output instead of filtering it out with the timestamp.
2016-09-14 23:30:57 +02:00
P00 INFO: new backup label = [BACKUP-FULL-1]
Added a predictable log entry to signal that a command has completed successfully. For example a backup ends successfully with: INFO: backup command end: completed successfully.
2016-12-04 21:44:53 +02:00
P00 INFO: backup command end: completed successfully
Integration test improvements for disk and memory efficiency. Set log-level-file=off when more that one test will run. In this case is it impossible to see the logs anyway since they will be automatically cleaned up after the test. This improves performance pretty dramatically since trace-level logging is expensive. If a singe integration test is run then log-level-file is trace by default but can be changed with the --log-level-test-file option. Reduce buffer-size to 64k to save memory during testing and allow more processes to run in parallel. Update log replacement rules so that these options can change without affecting expect logs.
2019-12-17 22:23:07 +02:00
P00 INFO: expire command begin [BACKREST-VERSION]: --buffer-size=[BUFFER-SIZE] --config=[TEST_PATH]/db-master/pgbackrest.conf --lock-path=[TEST_PATH]/db-master/lock --log-level-console=detail --log-level-file=[LOG-LEVEL-FILE] --log-level-stderr=off --log-path=[TEST_PATH]/db-master/log[] --no-log-timestamp --repo1-path=[TEST_PATH]/db-master/repo --repo1-type=cifs --stanza=db
Rename retention-* options to repo-retention-*.
2018-02-19 22:07:24 +02:00
P00 INFO: option 'repo1-retention-archive' is not set - archive logs will not be expired
Added a predictable log entry to signal that a command has completed successfully. For example a backup ends successfully with: INFO: backup command end: completed successfully.
2016-12-04 21:44:53 +02:00
P00 INFO: expire command end: completed successfully
Log testing can now be enabled for certain deterministic tests. This works by comparing the generated logs against a previous copy. Currently only enabled for the backup/synthetic tests.
2015-05-05 19:08:48 +02:00
Rename test paths for clarity. This was worked out as part of the test suite refactor [c8f806a] but not committed with it because of the large number of expect logs changes involved. Keeping them separate made it easier to audit the changes in the refactor.
2016-06-24 16:54:31 +02:00
+ supplemental file: [TEST_PATH]/db-master/pgbackrest.conf
----------------------------------------------------------
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
[db]
Rename db-* options to pg-* and backup-* options to repo-* to improve consistency. * repo-* options are now indexed although only one is allowed. * List deprecated option names in documentation and command-line help.
2018-02-04 01:27:38 +02:00
pg1-path=[TEST_PATH]/db-master/db/base
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
New simpler configuration and consistent project/exe/path naming. * The repo-path option now always refers to the repository where backups and archive are stored, whether local or remote, so the repo-remote-path option has been removed. The new spool-path option can be used to define a location for queueing WAL segments when archiving asynchronously. Otherwise, a local repository is no longer required. * Implemented a new config format which should be far simpler to use. See the User Guide and Configuration Reference for details but for a simple configuration all options can now be placed in the stanza section. Options that are shared between stanzas can be placed in the [global] section. More complex configurations can still make use of command sections though this should be a rare use case. * The default configuration filename is now pgbackrest.conf instead of pg_backrest.conf. This was done for consistency with other naming changes but also to prevent old config files from being loaded accidentally. * The default repository name was changed from /var/lib/backup to /var/lib/pgbackrest. * Lock files are now stored in /tmp/pgbackrest by default. These days /run/pgbackrest would be the preferred location but that would require init scripts which are not part of this release. The lock-path option can be used to configure the lock directory. * Log files are now stored in /var/log/pgbackrest by default and no longer have the date appended so they can be managed with logrotate. The log-path option can be used to configure the lock directory. * Executable filename changed from pg_backrest to pgbackrest.
2016-04-14 15:30:54 +02:00
[global]
Integration test improvements for disk and memory efficiency. Set log-level-file=off when more that one test will run. In this case is it impossible to see the logs anyway since they will be automatically cleaned up after the test. This improves performance pretty dramatically since trace-level logging is expensive. If a singe integration test is run then log-level-file is trace by default but can be changed with the --log-level-test-file option. Reduce buffer-size to 64k to save memory during testing and allow more processes to run in parallel. Update log replacement rules so that these options can change without affecting expect logs.
2019-12-17 22:23:07 +02:00
buffer-size=[BUFFER-SIZE]
Fix custom settings for compress-level option being ignored. Reported by Jens Wilke.
2017-10-23 00:05:46 +02:00
compress-level=3
Add infrastructure for multiple compression type support. Add compress-type option and deprecate compress option. Since the compress option is boolean it won't work with multiple compression types. Add logic to cfgLoadUpdateOption() to update compress-type if it is not set directly. The compress option should no longer be referenced outside the cfgLoadUpdateOption() function. Add common/compress/helper module to contain interface functions that work with multiple compression types. Code outside this module should no longer call specific compression drivers, though it may be OK to reference a specific compression type using the new interface (e.g., saving backup history files in gz format). Unit tests only test compression using the gz format because other formats may not be available in all builds. It is the job of integration tests to exercise all compression types. Additional compression types will be added in future commits.
2020-03-06 21:41:03 +02:00
compress-type=none
Storage and IO layer refactor: Refactor storage layer to allow for new repository filesystems using drivers. (Reviewed by Cynthia Shang.) Refactor IO layer to allow for new compression formats, checksum types, and other capabilities using filters. (Reviewed by Cynthia Shang.)
2017-06-09 23:51:41 +02:00
db-timeout=45
lock-path=[TEST_PATH]/db-master/lock
Reduce log level for all expect tests to detail. The C code is designed to be efficient rather than deterministic at the debug log level. As we move more testing from integration to unit tests it makes less sense to try and maintain the expect logs at this log level. Most of the expect logs have already been moved to detail level but mock/all still had tests at debug level. Change the logging defaults in the config file and remove as many references to log-level-console as possible.
2019-05-23 00:23:44 +02:00
log-level-console=detail
Integration test improvements for disk and memory efficiency. Set log-level-file=off when more that one test will run. In this case is it impossible to see the logs anyway since they will be automatically cleaned up after the test. This improves performance pretty dramatically since trace-level logging is expensive. If a singe integration test is run then log-level-file is trace by default but can be changed with the --log-level-test-file option. Reduce buffer-size to 64k to save memory during testing and allow more processes to run in parallel. Update log replacement rules so that these options can change without affecting expect logs.
2019-12-17 22:23:07 +02:00
log-level-file=[LOG-LEVEL-FILE]
Added the log-level-stderr option. Controls whether console log messages are sent to stderr or stdout. By default this is set to warn which represents a change in behavior from previous versions, even though it may be more intuitive. Setting log-level-stderr=off will preserve the old behavior. Suggested by Sascha Biberhofer.
2016-10-05 15:09:30 +02:00
log-level-stderr=off
Storage and IO layer refactor: Refactor storage layer to allow for new repository filesystems using drivers. (Reviewed by Cynthia Shang.) Refactor IO layer to allow for new compression formats, checksum types, and other capabilities using filters. (Reviewed by Cynthia Shang.)
2017-06-09 23:51:41 +02:00
log-path=[TEST_PATH]/db-master/log
Integration test improvements for disk and memory efficiency. Set log-level-file=off when more that one test will run. In this case is it impossible to see the logs anyway since they will be automatically cleaned up after the test. This improves performance pretty dramatically since trace-level logging is expensive. If a singe integration test is run then log-level-file is trace by default but can be changed with the --log-level-test-file option. Reduce buffer-size to 64k to save memory during testing and allow more processes to run in parallel. Update log replacement rules so that these options can change without affecting expect logs.
2019-12-17 22:23:07 +02:00
log-subprocess=[LOG-SUBPROCESS]
Set log-timestamp=n for integration tests. This means less filtering of logs needs to be done and new timestamps can be added without adding new filters.
2018-05-11 17:24:38 +02:00
log-timestamp=n
Storage and IO layer refactor: Refactor storage layer to allow for new repository filesystems using drivers. (Reviewed by Cynthia Shang.) Refactor IO layer to allow for new compression formats, checksum types, and other capabilities using filters. (Reviewed by Cynthia Shang.)
2017-06-09 23:51:41 +02:00
protocol-timeout=60
Rename db-* options to pg-* and backup-* options to repo-* to improve consistency. * repo-* options are now indexed although only one is allowed. * List deprecated option names in documentation and command-line help.
2018-02-04 01:27:38 +02:00
repo1-path=[TEST_PATH]/db-master/repo
Consolidate archive-push tests into new archive test.
2017-06-21 14:38:53 +02:00
spool-path=[TEST_PATH]/db-master/spool
New simpler configuration and consistent project/exe/path naming. * The repo-path option now always refers to the repository where backups and archive are stored, whether local or remote, so the repo-remote-path option has been removed. The new spool-path option can be used to define a location for queueing WAL segments when archiving asynchronously. Otherwise, a local repository is no longer required. * Implemented a new config format which should be far simpler to use. See the User Guide and Configuration Reference for details but for a simple configuration all options can now be placed in the stanza section. Options that are shared between stanzas can be placed in the [global] section. More complex configurations can still make use of command sections though this should be a rare use case. * The default configuration filename is now pgbackrest.conf instead of pg_backrest.conf. This was done for consistency with other naming changes but also to prevent old config files from being loaded accidentally. * The default repository name was changed from /var/lib/backup to /var/lib/pgbackrest. * Lock files are now stored in /tmp/pgbackrest by default. These days /run/pgbackrest would be the preferred location but that would require init scripts which are not part of this release. The lock-path option can be used to configure the lock directory. * Log files are now stored in /var/log/pgbackrest by default and no longer have the date appended so they can be managed with logrotate. The log-path option can be used to configure the lock directory. * Executable filename changed from pg_backrest to pgbackrest.
2016-04-14 15:30:54 +02:00
[global:backup]
archive-copy=y
start-fast=y
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
Rename test paths for clarity. This was worked out as part of the test suite refactor [c8f806a] but not committed with it because of the large number of expect logs changes involved. Keeping them separate made it easier to audit the changes in the refactor.
2016-06-24 16:54:31 +02:00
+ supplemental file: [TEST_PATH]/db-master/repo/backup/db/[BACKUP-FULL-1]/backup.manifest
-----------------------------------------------------------------------------------------
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
[backrest]
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
backrest-format=5
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
backrest-version="[VERSION-1]"
[backup]
backup-label="[BACKUP-FULL-1]"
backup-timestamp-copy-start=[TIMESTAMP]
backup-timestamp-start=[TIMESTAMP]
backup-timestamp-stop=[TIMESTAMP]
backup-type="full"
[backup:db]
Database version refactoring. * Refactor db version constants into a separate module. * Update synthetic backup tests to PostgreSQL 9.4.
2016-08-12 04:35:24 +02:00
db-catalog-version=201409291
db-control-version=942
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
db-id=1
Automate generation of WAL and pg_control test files. The existing static files would not work with 32-bit or big-endian systems so create functions to generate these files dynamically rather than creating a bunch of new static files.
2017-11-19 03:02:54 +02:00
db-system-id=1000000000000000094
Database version refactoring. * Refactor db version constants into a separate module. * Update synthetic backup tests to PostgreSQL 9.4.
2016-08-12 04:35:24 +02:00
db-version="9.4"
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
[backup:option]
Set archive-check option in manifest correctly when offline. Archive check does not run when in offline backup mode but the option was set to true in the manifest. It's harmless since these options are informational only but it could cause confusion when debugging.
2019-11-28 15:27:21 +02:00
option-archive-check=false
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
option-archive-copy=true
Backup from a standby cluster. A connection to the primary cluster is still required to start/stop the backup and copy files that are not replicated, but the vast majority of files are copied from the standby in order to reduce load on the master.
2016-08-25 17:25:46 +02:00
option-backup-standby=false
Integration test improvements for disk and memory efficiency. Set log-level-file=off when more that one test will run. In this case is it impossible to see the logs anyway since they will be automatically cleaned up after the test. This improves performance pretty dramatically since trace-level logging is expensive. If a singe integration test is run then log-level-file is trace by default but can be changed with the --log-level-test-file option. Reduce buffer-size to 64k to save memory during testing and allow more processes to run in parallel. Update log replacement rules so that these options can change without affecting expect logs.
2019-12-17 22:23:07 +02:00
option-buffer-size=[BUFFER-SIZE]
Added the --checksum-page option. This option allows pgBackRest to validate page checksums in data files when checksums are enabled on PostgreSQL >= 9.3. Note that this functionality requires a C library which may not initially be available in OS packages. The option will automatically be enabled when the library is present and checksums are enabled on the cluster.
2016-12-13 01:54:07 +02:00
option-checksum-page=true
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
option-compress=false
Add additional options to backup.manifest for debugging purposes. Add the buffer-size, compress-level, compress-level-network, and process-max options to the backup:option section in backup.manifest to aid in debugging. It may also make sense to propagate these options up to backup.info so they can be displayed in the info command, but for now this is deemed sufficient. Contributed by blogh.
2019-03-10 11:03:52 +02:00
option-compress-level=3
option-compress-level-network=3
Add infrastructure for multiple compression type support. Add compress-type option and deprecate compress option. Since the compress option is boolean it won't work with multiple compression types. Add logic to cfgLoadUpdateOption() to update compress-type if it is not set directly. The compress option should no longer be referenced outside the cfgLoadUpdateOption() function. Add common/compress/helper module to contain interface functions that work with multiple compression types. Code outside this module should no longer call specific compression drivers, though it may be OK to reference a specific compression type using the new interface (e.g., saving backup history files in gz format). Unit tests only test compression using the gz format because other formats may not be available in all builds. It is the job of integration tests to exercise all compression types. Additional compression types will be added in future commits.
2020-03-06 21:41:03 +02:00
option-compress-type="none"
Add checksum delta for incremental backups. Use checksums rather than timestamps to determine if files have changed. This is useful in cases where the timestamps may not be trustworthy, e.g. when performing an incremental after failing over to a standby. If checksum delta is enabled then checksums will be used for verification of resumed backups, even if they are full. Resumes have always used checksums to verify the files in the repository, enabling delta performs checksums on the database files as well. Note that the user must manually enable this feature in cases were it would be useful or just keep in enabled all the time. A future commit will address automatically enabling the feature in cases where it seems likely to be useful. Contributed by Cynthia Shang.
2018-09-19 17:12:45 +02:00
option-delta=false
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
option-hardlink=false
Closed #167: Rename `--no-start-stop` option to `--no-online`.
2016-02-20 02:32:02 +02:00
option-online=false
Add additional options to backup.manifest for debugging purposes. Add the buffer-size, compress-level, compress-level-network, and process-max options to the backup:option section in backup.manifest to aid in debugging. It may also make sense to propagate these options up to backup.info so they can be displayed in the info command, but for now this is deemed sufficient. Contributed by blogh.
2019-03-10 11:03:52 +02:00
option-process-max=1
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
[backup:target]
Rename test paths for clarity. This was worked out as part of the test suite refactor [c8f806a] but not committed with it because of the large number of expect logs changes involved. Keeping them separate made it easier to audit the changes in the refactor.
2016-06-24 16:54:31 +02:00
pg_data={"path":"[TEST_PATH]/db-master/db/base","type":"path"}
Fix failure in manifest build when two or more files in PGDATA are linked to the same directory. Reported by Vitaliy Kukharik.
2018-05-02 18:19:54 +02:00
pg_data/pg_hba.conf={"file":"pg_hba.conf","path":"../pg_config","type":"link"}
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
pg_data/pg_stat={"path":"../pg_stat","type":"link"}
pg_data/postgresql.conf={"file":"postgresql.conf","path":"../pg_config","type":"link"}
[target:file]
Fixed a issue where tablespaces were copied from the master during standby backup.
2016-09-04 15:19:44 +02:00
pg_data/PG_VERSION={"checksum":"184473f470864e067ee3a22e64b47b0a1c356f29","master":true,"size":[SIZE],"timestamp":[TIMESTAMP-1]}
Added the --checksum-page option. This option allows pgBackRest to validate page checksums in data files when checksums are enabled on PostgreSQL >= 9.3. Note that this functionality requires a C library which may not initially be available in OS packages. The option will automatically be enabled when the library is present and checksums are enabled on the cluster.
2016-12-13 01:54:07 +02:00
pg_data/base/1/12000={"checksum":"22c98d248ff548311eda88559e4a8405ed77c003","checksum-page":true,"size":[SIZE],"timestamp":[TIMESTAMP-1]}
Database version refactoring. * Refactor db version constants into a separate module. * Update synthetic backup tests to PostgreSQL 9.4.
2016-08-12 04:35:24 +02:00
pg_data/base/1/PG_VERSION={"checksum":"184473f470864e067ee3a22e64b47b0a1c356f29","mode":"0660","size":[SIZE],"timestamp":[TIMESTAMP-1],"user":false}
Added the --checksum-page option. This option allows pgBackRest to validate page checksums in data files when checksums are enabled on PostgreSQL >= 9.3. Note that this functionality requires a C library which may not initially be available in OS packages. The option will automatically be enabled when the library is present and checksums are enabled on the cluster.
2016-12-13 01:54:07 +02:00
pg_data/base/16384/17000={"checksum":"e0101dd8ffb910c9c202ca35b5f828bcb9697bed","checksum-page":false,"checksum-page-error":[1],"size":[SIZE],"timestamp":[TIMESTAMP-1]}
Database version refactoring. * Refactor db version constants into a separate module. * Update synthetic backup tests to PostgreSQL 9.4.
2016-08-12 04:35:24 +02:00
pg_data/base/16384/PG_VERSION={"checksum":"184473f470864e067ee3a22e64b47b0a1c356f29","group":false,"size":[SIZE],"timestamp":[TIMESTAMP-1]}
Fix incorrect offline upper bound for ignoring page checksum errors. For offline backups the upper bound was being set to 0x0000FFFF0000FFFF rather than UINT64_MAX. This meant that page checksum errors might be ignored for databases with a lot of past WAL in offline mode. Online mode is not affected since the upper bound is retrieved from pg_start_backup().
2019-07-11 15:13:56 +02:00
pg_data/base/32768/33000={"checksum":"7a16d165e4775f7c92e8cdf60c0af57313f0bf90","checksum-page":true,"size":[SIZE],"timestamp":[TIMESTAMP-1]}
pg_data/base/32768/33000.32767={"checksum":"6e99b589e550e68e934fd235ccba59fe5b592a9e","checksum-page":true,"size":[SIZE],"timestamp":[TIMESTAMP-1]}
Added the --checksum-page option. This option allows pgBackRest to validate page checksums in data files when checksums are enabled on PostgreSQL >= 9.3. Note that this functionality requires a C library which may not initially be available in OS packages. The option will automatically be enabled when the library is present and checksums are enabled on the cluster.
2016-12-13 01:54:07 +02:00
pg_data/base/32768/33001={"checksum":"6bf316f11d28c28914ea9be92c00de9bea6d9a6b","checksum-page":false,"checksum-page-error":[0,[3,5],7],"size":[SIZE],"timestamp":[TIMESTAMP-1]}
Fix incorrect offline upper bound for ignoring page checksum errors. For offline backups the upper bound was being set to 0x0000FFFF0000FFFF rather than UINT64_MAX. This meant that page checksum errors might be ignored for databases with a lot of past WAL in offline mode. Online mode is not affected since the upper bound is retrieved from pg_start_backup().
2019-07-11 15:13:56 +02:00
pg_data/base/32768/44000_init={"checksum":"7a16d165e4775f7c92e8cdf60c0af57313f0bf90","checksum-page":true,"size":[SIZE],"timestamp":[TIMESTAMP-1]}
Database version refactoring. * Refactor db version constants into a separate module. * Update synthetic backup tests to PostgreSQL 9.4.
2016-08-12 04:35:24 +02:00
pg_data/base/32768/PG_VERSION={"checksum":"184473f470864e067ee3a22e64b47b0a1c356f29","size":[SIZE],"timestamp":[TIMESTAMP-1]}
Support configurable WAL segment size. PostgreSQL 11 introduces configurable WAL segment sizes, from 1MB to 1GB. There are two areas that needed to be updated to support this: building the archive-get queue and checking that WAL has been archived after a backup. Both operations require the WAL segment size to properly build a list. Checking the archive after a backup is still implemented in Perl and has an active database connection, so just get the WAL segment size from the database. The archive-get command does not have a connection to the database, so get the WAL segment size from pg_control instead. This requires a deeper inspection of pg_control than has been done in the past, so it seemed best to copy the relevant data structures from each version of PostgreSQL and build a generic interface layer to address them. While this approach is a bit verbose, it has the advantage of being relatively simple, and can easily be updated for new versions of PostgreSQL. Since the integration tests generate pg_control files for testing, teach Perl how to generate files with the correct offsets for both 32-bit and 64-bit architectures.
2018-09-25 11:24:42 +02:00
pg_data/global/pg_control={"checksum":"4c77c900f7af0d9ab13fa9982051a42e0b637f6c","master":true,"size":[SIZE],"timestamp":[TIMESTAMP-2]}
Fix failure in manifest build when two or more files in PGDATA are linked to the same directory. Reported by Vitaliy Kukharik.
2018-05-02 18:19:54 +02:00
pg_data/pg_hba.conf={"checksum":"dd4cea0cae348309f9de28ad4ded8ee2cc2e6d5b","master":true,"size":[SIZE],"timestamp":[TIMESTAMP-2]}
Fixed a issue where tablespaces were copied from the master during standby backup.
2016-09-04 15:19:44 +02:00
pg_data/pg_stat/global.stat={"checksum":"e350d5ce0153f3e22d5db21cf2a4eff00f3ee877","master":true,"size":[SIZE],"timestamp":[TIMESTAMP-2]}
pg_data/postgresql.conf={"checksum":"6721d92c9fcdf4248acff1f9a1377127d9064807","master":true,"size":[SIZE],"timestamp":[TIMESTAMP-2]}
Use minio for integration testing. ScalityS3 has not received any maintenance in years and is slow to start which is bad for testing. Replace it with minio which starts quickly and ships as a single executable or a tiny container. Minio has stricter limits on allowable characters but should still provide enough coverage to show that our encoding is working correctly. This commit also includes the upgrade to openssl 1.1.1 in the Ubuntu 18.04 container.
2019-07-03 04:20:35 +02:00
pg_data/special-!_.*'()&!@;:+,?={"master":true,"size":[SIZE],"timestamp":[TIMESTAMP-1]}
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
[target:file:default]
Continuous integration using travis-ci.
2017-02-21 15:59:23 +02:00
group="[GROUP-1]"
Fixed a issue where tablespaces were copied from the master during standby backup.
2016-09-04 15:19:44 +02:00
master=false
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
mode="0600"
Regression tests can now be run as any properly-configured user, not just vagrant.
2016-10-01 19:39:44 +02:00
user="[USER-1]"
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
[target:link]
Fix failure in manifest build when two or more files in PGDATA are linked to the same directory. Reported by Vitaliy Kukharik.
2018-05-02 18:19:54 +02:00
pg_data/pg_hba.conf={"destination":"../pg_config/pg_hba.conf"}
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
pg_data/pg_stat={"destination":"../pg_stat"}
pg_data/postgresql.conf={"destination":"../pg_config/postgresql.conf"}
[target:link:default]
Continuous integration using travis-ci.
2017-02-21 15:59:23 +02:00
group="[GROUP-1]"
Regression tests can now be run as any properly-configured user, not just vagrant.
2016-10-01 19:39:44 +02:00
user="[USER-1]"
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
[target:path]
pg_data={}
pg_data/base={}
Allow selective restore of databases from a cluster backup. This feature can result in major space and time savings when only specific databases are restored. Unrestored databases will not be accessible but must be manually dropped before they will be removed from the shared catalogue.
2016-05-11 15:21:39 +02:00
pg_data/base/1={}
pg_data/base/16384={}
pg_data/base/32768={}
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
pg_data/global={}
pg_data/pg_clog={}
Exclude directories during backup that are cleaned, recreated, or zeroed by PostgreSQL at startup. These include (depending on the version where they were introduced): pgsql_tmp, pg_dynshmem, pg_notify, pg_replslot, pg_serial, pg_snapshots, pg_stat_tmp, pg_subtrans. The postgresql.auto.conf.tmp file is now excluded in addition to files that were already excluded: backup_label.old, postmaster.opts, postmaster.pid, recovery.conf, recovery.done.
2016-08-16 15:35:16 +02:00
pg_data/pg_dynshmem={}
pg_data/pg_notify={}
pg_data/pg_replslot={}
pg_data/pg_serial={}
pg_data/pg_snapshots={}
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
pg_data/pg_stat={}
Exclude directories during backup that are cleaned, recreated, or zeroed by PostgreSQL at startup. These include (depending on the version where they were introduced): pgsql_tmp, pg_dynshmem, pg_notify, pg_replslot, pg_serial, pg_snapshots, pg_stat_tmp, pg_subtrans. The postgresql.auto.conf.tmp file is now excluded in addition to files that were already excluded: backup_label.old, postmaster.opts, postmaster.pid, recovery.conf, recovery.done.
2016-08-16 15:35:16 +02:00
pg_data/pg_stat_tmp={}
pg_data/pg_subtrans={}
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
pg_data/pg_tblspc={}
[target:path:default]
Continuous integration using travis-ci.
2017-02-21 15:59:23 +02:00
group="[GROUP-1]"
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
mode="0700"
Regression tests can now be run as any properly-configured user, not just vagrant.
2016-10-01 19:39:44 +02:00
user="[USER-1]"
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
Move info file checksum to the end of the file. Putting the checksum at the beginning of the file made it impossible to stream the file out when saving. The entire file had to be held in memory while it was checksummed so the checksum could be written at the beginning. Instead place the checksum at the end. This does not break the existing Perl or C code since the read is not order dependent. There are no plans to improve the Perl code to take advantage of this change, but it will make the C implementation more efficient. Reviewed by Cynthia Shang.
2019-08-22 01:45:48 +02:00
[backrest]
backrest-checksum="[CHECKSUM]"
Rename test paths for clarity. This was worked out as part of the test suite refactor [c8f806a] but not committed with it because of the large number of expect logs changes involved. Keeping them separate made it easier to audit the changes in the refactor.
2016-06-24 16:54:31 +02:00
+ supplemental file: [TEST_PATH]/db-master/repo/backup/db/backup.info
---------------------------------------------------------------------
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
[backrest]
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
backrest-format=5
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
backrest-version="[VERSION-1]"
[backup:current]
Set archive-check option in manifest correctly when offline. Archive check does not run when in offline backup mode but the option was set to true in the manifest. It's harmless since these options are informational only but it could cause confusion when debugging.
2019-11-28 15:27:21 +02:00
[BACKUP-FULL-1]={"backrest-format":5,"backrest-version":"[VERSION-1]","backup-archive-start":null,"backup-archive-stop":null,"backup-info-repo-size":[SIZE],"backup-info-repo-size-delta":[DELTA],"backup-info-size":[SIZE],"backup-info-size-delta":[DELTA],"backup-timestamp-start":[TIMESTAMP],"backup-timestamp-stop":[TIMESTAMP],"backup-type":"full","db-id":1,"option-archive-check":false,"option-archive-copy":true,"option-backup-standby":false,"option-checksum-page":true,"option-compress":false,"option-hardlink":false,"option-online":false}
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
[db]
Database version refactoring. * Refactor db version constants into a separate module. * Update synthetic backup tests to PostgreSQL 9.4.
2016-08-12 04:35:24 +02:00
db-catalog-version=201409291
db-control-version=942
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
db-id=1
Automate generation of WAL and pg_control test files. The existing static files would not work with 32-bit or big-endian systems so create functions to generate these files dynamically rather than creating a bunch of new static files.
2017-11-19 03:02:54 +02:00
db-system-id=1000000000000000094
Database version refactoring. * Refactor db version constants into a separate module. * Update synthetic backup tests to PostgreSQL 9.4.
2016-08-12 04:35:24 +02:00
db-version="9.4"
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
[db:history]
Automate generation of WAL and pg_control test files. The existing static files would not work with 32-bit or big-endian systems so create functions to generate these files dynamically rather than creating a bunch of new static files.
2017-11-19 03:02:54 +02:00
1={"db-catalog-version":201409291,"db-control-version":942,"db-system-id":1000000000000000094,"db-version":"9.4"}
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
Improve performance of info file load/save. Info files required three copies in memory to be loaded (the original string, an ini representation, and the final info object). Not only was this memory inefficient but the Ini object does sequential scans when searching for keys making large files very slow to load. This has not been an issue since archive.info and backup.info are very small, but it becomes a big deal when loading manifests with hundreds of thousands of files. Instead of holding copies of the data in memory, use a callback to deliver the ini data directly to the object when loading. Use a similar method for save to avoid having an intermediate copy. Save is a bit complex because sections/keys must be written in alpha order or older versions of pgBackRest will not calculate the correct checksum. Also move the load retry logic to helper functions rather than embedding it in the Info object. This allows for more flexibility in loading and ensures that stack traces will be available when developing unit tests. Reviewed by Cynthia Shang.
2019-09-06 19:48:28 +02:00
[backrest]
backrest-checksum="[CHECKSUM]"
Major refactor of the test suite. * Make the code more modular and object-oriented. * Multiple Docker containers can now be created for a single test to simulate more realistic environments.
2016-06-24 14:12:58 +02:00
stop all stanzas (db-master host)
Rename test paths for clarity. This was worked out as part of the test suite refactor [c8f806a] but not committed with it because of the large number of expect logs changes involved. Keeping them separate made it easier to audit the changes in the refactor.
2016-06-24 16:54:31 +02:00
> [CONTAINER-EXEC] db-master [BACKREST-BIN] --config=[TEST_PATH]/db-master/pgbackrest.conf --force stop
Implemented issue #129: Stop/Start Commands. Implemented issue #136: Add stderr back into protocol layer. There were also miscellaneous bug fixes and improvements to buffering in IO.pm.
2015-10-08 17:43:56 +02:00
------------------------------------------------------------------------------------------------------------------------------------
Integration test improvements for disk and memory efficiency. Set log-level-file=off when more that one test will run. In this case is it impossible to see the logs anyway since they will be automatically cleaned up after the test. This improves performance pretty dramatically since trace-level logging is expensive. If a singe integration test is run then log-level-file is trace by default but can be changed with the --log-level-test-file option. Reduce buffer-size to 64k to save memory during testing and allow more processes to run in parallel. Update log replacement rules so that these options can change without affecting expect logs.
2019-12-17 22:23:07 +02:00
P00 INFO: stop command begin [BACKREST-VERSION]: --config=[TEST_PATH]/db-master/pgbackrest.conf --force --lock-path=[TEST_PATH]/db-master/lock --log-level-console=detail --log-level-file=[LOG-LEVEL-FILE] --log-level-stderr=off --log-path=[TEST_PATH]/db-master/log[] --no-log-timestamp --repo1-path=[TEST_PATH]/db-master/repo
Added a predictable log entry to signal that a command has completed successfully. For example a backup ends successfully with: INFO: backup command end: completed successfully.
2016-12-04 21:44:53 +02:00
P00 INFO: stop command end: completed successfully
Implemented issue #129: Stop/Start Commands. Implemented issue #136: Add stderr back into protocol layer. There were also miscellaneous bug fixes and improvements to buffering in IO.pm.
2015-10-08 17:43:56 +02:00
Major refactor of the test suite. * Make the code more modular and object-oriented. * Multiple Docker containers can now be created for a single test to simulate more realistic environments.
2016-06-24 14:12:58 +02:00
full backup - global stop (db-master host)
Rename test paths for clarity. This was worked out as part of the test suite refactor [c8f806a] but not committed with it because of the large number of expect logs changes involved. Keeping them separate made it easier to audit the changes in the refactor.
2016-06-24 16:54:31 +02:00
> [CONTAINER-EXEC] db-master [BACKREST-BIN] --config=[TEST_PATH]/db-master/pgbackrest.conf --no-online --type=full --stanza=db backup
Implemented issue #129: Stop/Start Commands. Implemented issue #136: Add stderr back into protocol layer. There were also miscellaneous bug fixes and improvements to buffering in IO.pm.
2015-10-08 17:43:56 +02:00
------------------------------------------------------------------------------------------------------------------------------------
Add infrastructure for multiple compression type support. Add compress-type option and deprecate compress option. Since the compress option is boolean it won't work with multiple compression types. Add logic to cfgLoadUpdateOption() to update compress-type if it is not set directly. The compress option should no longer be referenced outside the cfgLoadUpdateOption() function. Add common/compress/helper module to contain interface functions that work with multiple compression types. Code outside this module should no longer call specific compression drivers, though it may be OK to reference a specific compression type using the new interface (e.g., saving backup history files in gz format). Unit tests only test compression using the gz format because other formats may not be available in all builds. It is the job of integration tests to exercise all compression types. Additional compression types will be added in future commits.
2020-03-06 21:41:03 +02:00
P00 INFO: backup command begin [BACKREST-VERSION]: --buffer-size=[BUFFER-SIZE] --compress-level=3 --compress-type=none --config=[TEST_PATH]/db-master/pgbackrest.conf --db-timeout=45 --lock-path=[TEST_PATH]/db-master/lock --log-level-console=detail --log-level-file=[LOG-LEVEL-FILE] --log-level-stderr=off --log-path=[TEST_PATH]/db-master/log[] --no-log-timestamp --no-online --pg1-path=[TEST_PATH]/db-master/db/base --protocol-timeout=60 --repo1-path=[TEST_PATH]/db-master/repo --stanza=db --start-fast --type=full
Add time-based retention for full backups. The --repo-retention-full-type option allows retention of full backups based on a time period, specified in days. The new option will default to 'count' and therefore will not affect current installations. Setting repo-retention-full-type to 'time' will allow the user to use a time period, in days, to indicate full backup retention. Using this method, a full backup can be expired only if the time the backup completed is older than the number of days set with repo-retention-full (calculated from the moment the 'expire' command is run) and at least one full backup meets the retention period. If archive retention has not been configured, then the default settings will expire archives that are prior to the oldest retained full backup. For example, if there are three full backups ending in times that are 25 days old (F1), 20 days old (F2) and 10 days old (F3), then if the full retention period is 15 days, then only F1 will be expired; F2 will be retained because F1 is not at least 15 days old.
2020-05-08 21:25:03 +02:00
P00 WARN: option 'repo1-retention-full' is not set for 'repo1-retention-full-type=count', the repository may run out of space
Rename retention-* options to repo-retention-*.
2018-02-19 22:07:24 +02:00
HINT: to retain full backups indefinitely (without warning), set option 'repo1-retention-full' to the maximum.
Renumber return codes between 25 and 125 to avoid PostgreSQL interpreting some as fatal signal exceptions. Suggested by Yogesh Sharma.
2017-04-12 18:41:34 +02:00
P00 ERROR: [062]: stop file exists for all stanzas
P00 INFO: backup command end: aborted with exception [062]
Implemented issue #129: Stop/Start Commands. Implemented issue #136: Add stderr back into protocol layer. There were also miscellaneous bug fixes and improvements to buffering in IO.pm.
2015-10-08 17:43:56 +02:00
Major refactor of the test suite. * Make the code more modular and object-oriented. * Multiple Docker containers can now be created for a single test to simulate more realistic environments.
2016-06-24 14:12:58 +02:00
stop db stanza (db-master host)
Rename test paths for clarity. This was worked out as part of the test suite refactor [c8f806a] but not committed with it because of the large number of expect logs changes involved. Keeping them separate made it easier to audit the changes in the refactor.
2016-06-24 16:54:31 +02:00
> [CONTAINER-EXEC] db-master [BACKREST-BIN] --config=[TEST_PATH]/db-master/pgbackrest.conf --stanza=db stop
Implemented issue #129: Stop/Start Commands. Implemented issue #136: Add stderr back into protocol layer. There were also miscellaneous bug fixes and improvements to buffering in IO.pm.
2015-10-08 17:43:56 +02:00
------------------------------------------------------------------------------------------------------------------------------------
Integration test improvements for disk and memory efficiency. Set log-level-file=off when more that one test will run. In this case is it impossible to see the logs anyway since they will be automatically cleaned up after the test. This improves performance pretty dramatically since trace-level logging is expensive. If a singe integration test is run then log-level-file is trace by default but can be changed with the --log-level-test-file option. Reduce buffer-size to 64k to save memory during testing and allow more processes to run in parallel. Update log replacement rules so that these options can change without affecting expect logs.
2019-12-17 22:23:07 +02:00
P00 INFO: stop command begin [BACKREST-VERSION]: --config=[TEST_PATH]/db-master/pgbackrest.conf --lock-path=[TEST_PATH]/db-master/lock --log-level-console=detail --log-level-file=[LOG-LEVEL-FILE] --log-level-stderr=off --log-path=[TEST_PATH]/db-master/log[] --no-log-timestamp --repo1-path=[TEST_PATH]/db-master/repo --stanza=db
Added a predictable log entry to signal that a command has completed successfully. For example a backup ends successfully with: INFO: backup command end: completed successfully.
2016-12-04 21:44:53 +02:00
P00 INFO: stop command end: completed successfully
Implemented issue #129: Stop/Start Commands. Implemented issue #136: Add stderr back into protocol layer. There were also miscellaneous bug fixes and improvements to buffering in IO.pm.
2015-10-08 17:43:56 +02:00
Major refactor of the test suite. * Make the code more modular and object-oriented. * Multiple Docker containers can now be created for a single test to simulate more realistic environments.
2016-06-24 14:12:58 +02:00
stop db stanza (db-master host)
Rename test paths for clarity. This was worked out as part of the test suite refactor [c8f806a] but not committed with it because of the large number of expect logs changes involved. Keeping them separate made it easier to audit the changes in the refactor.
2016-06-24 16:54:31 +02:00
> [CONTAINER-EXEC] db-master [BACKREST-BIN] --config=[TEST_PATH]/db-master/pgbackrest.conf --stanza=db stop
Implemented issue #129: Stop/Start Commands. Implemented issue #136: Add stderr back into protocol layer. There were also miscellaneous bug fixes and improvements to buffering in IO.pm.
2015-10-08 17:43:56 +02:00
------------------------------------------------------------------------------------------------------------------------------------
Integration test improvements for disk and memory efficiency. Set log-level-file=off when more that one test will run. In this case is it impossible to see the logs anyway since they will be automatically cleaned up after the test. This improves performance pretty dramatically since trace-level logging is expensive. If a singe integration test is run then log-level-file is trace by default but can be changed with the --log-level-test-file option. Reduce buffer-size to 64k to save memory during testing and allow more processes to run in parallel. Update log replacement rules so that these options can change without affecting expect logs.
2019-12-17 22:23:07 +02:00
P00 INFO: stop command begin [BACKREST-VERSION]: --config=[TEST_PATH]/db-master/pgbackrest.conf --lock-path=[TEST_PATH]/db-master/lock --log-level-console=detail --log-level-file=[LOG-LEVEL-FILE] --log-level-stderr=off --log-path=[TEST_PATH]/db-master/log[] --no-log-timestamp --repo1-path=[TEST_PATH]/db-master/repo --stanza=db
Show Process ID in log output instead of filtering it out with the timestamp.
2016-09-14 23:30:57 +02:00
P00 WARN: stop file already exists for stanza db
Added a predictable log entry to signal that a command has completed successfully. For example a backup ends successfully with: INFO: backup command end: completed successfully.
2016-12-04 21:44:53 +02:00
P00 INFO: stop command end: completed successfully
Implemented issue #129: Stop/Start Commands. Implemented issue #136: Add stderr back into protocol layer. There were also miscellaneous bug fixes and improvements to buffering in IO.pm.
2015-10-08 17:43:56 +02:00
Major refactor of the test suite. * Make the code more modular and object-oriented. * Multiple Docker containers can now be created for a single test to simulate more realistic environments.
2016-06-24 14:12:58 +02:00
full backup - stanza stop (db-master host)
Rename test paths for clarity. This was worked out as part of the test suite refactor [c8f806a] but not committed with it because of the large number of expect logs changes involved. Keeping them separate made it easier to audit the changes in the refactor.
2016-06-24 16:54:31 +02:00
> [CONTAINER-EXEC] db-master [BACKREST-BIN] --config=[TEST_PATH]/db-master/pgbackrest.conf --no-online --type=full --stanza=db backup
Implemented issue #129: Stop/Start Commands. Implemented issue #136: Add stderr back into protocol layer. There were also miscellaneous bug fixes and improvements to buffering in IO.pm.
2015-10-08 17:43:56 +02:00
------------------------------------------------------------------------------------------------------------------------------------
Add infrastructure for multiple compression type support. Add compress-type option and deprecate compress option. Since the compress option is boolean it won't work with multiple compression types. Add logic to cfgLoadUpdateOption() to update compress-type if it is not set directly. The compress option should no longer be referenced outside the cfgLoadUpdateOption() function. Add common/compress/helper module to contain interface functions that work with multiple compression types. Code outside this module should no longer call specific compression drivers, though it may be OK to reference a specific compression type using the new interface (e.g., saving backup history files in gz format). Unit tests only test compression using the gz format because other formats may not be available in all builds. It is the job of integration tests to exercise all compression types. Additional compression types will be added in future commits.
2020-03-06 21:41:03 +02:00
P00 INFO: backup command begin [BACKREST-VERSION]: --buffer-size=[BUFFER-SIZE] --compress-level=3 --compress-type=none --config=[TEST_PATH]/db-master/pgbackrest.conf --db-timeout=45 --lock-path=[TEST_PATH]/db-master/lock --log-level-console=detail --log-level-file=[LOG-LEVEL-FILE] --log-level-stderr=off --log-path=[TEST_PATH]/db-master/log[] --no-log-timestamp --no-online --pg1-path=[TEST_PATH]/db-master/db/base --protocol-timeout=60 --repo1-path=[TEST_PATH]/db-master/repo --stanza=db --start-fast --type=full
Add time-based retention for full backups. The --repo-retention-full-type option allows retention of full backups based on a time period, specified in days. The new option will default to 'count' and therefore will not affect current installations. Setting repo-retention-full-type to 'time' will allow the user to use a time period, in days, to indicate full backup retention. Using this method, a full backup can be expired only if the time the backup completed is older than the number of days set with repo-retention-full (calculated from the moment the 'expire' command is run) and at least one full backup meets the retention period. If archive retention has not been configured, then the default settings will expire archives that are prior to the oldest retained full backup. For example, if there are three full backups ending in times that are 25 days old (F1), 20 days old (F2) and 10 days old (F3), then if the full retention period is 15 days, then only F1 will be expired; F2 will be retained because F1 is not at least 15 days old.
2020-05-08 21:25:03 +02:00
P00 WARN: option 'repo1-retention-full' is not set for 'repo1-retention-full-type=count', the repository may run out of space
Rename retention-* options to repo-retention-*.
2018-02-19 22:07:24 +02:00
HINT: to retain full backups indefinitely (without warning), set option 'repo1-retention-full' to the maximum.
Renumber return codes between 25 and 125 to avoid PostgreSQL interpreting some as fatal signal exceptions. Suggested by Yogesh Sharma.
2017-04-12 18:41:34 +02:00
P00 ERROR: [062]: stop file exists for stanza db
P00 INFO: backup command end: aborted with exception [062]
Implemented issue #129: Stop/Start Commands. Implemented issue #136: Add stderr back into protocol layer. There were also miscellaneous bug fixes and improvements to buffering in IO.pm.
2015-10-08 17:43:56 +02:00
Major refactor of the test suite. * Make the code more modular and object-oriented. * Multiple Docker containers can now be created for a single test to simulate more realistic environments.
2016-06-24 14:12:58 +02:00
start db stanza (db-master host)
Rename test paths for clarity. This was worked out as part of the test suite refactor [c8f806a] but not committed with it because of the large number of expect logs changes involved. Keeping them separate made it easier to audit the changes in the refactor.
2016-06-24 16:54:31 +02:00
> [CONTAINER-EXEC] db-master [BACKREST-BIN] --config=[TEST_PATH]/db-master/pgbackrest.conf --stanza=db start
Implemented issue #129: Stop/Start Commands. Implemented issue #136: Add stderr back into protocol layer. There were also miscellaneous bug fixes and improvements to buffering in IO.pm.
2015-10-08 17:43:56 +02:00
------------------------------------------------------------------------------------------------------------------------------------
Integration test improvements for disk and memory efficiency. Set log-level-file=off when more that one test will run. In this case is it impossible to see the logs anyway since they will be automatically cleaned up after the test. This improves performance pretty dramatically since trace-level logging is expensive. If a singe integration test is run then log-level-file is trace by default but can be changed with the --log-level-test-file option. Reduce buffer-size to 64k to save memory during testing and allow more processes to run in parallel. Update log replacement rules so that these options can change without affecting expect logs.
2019-12-17 22:23:07 +02:00
P00 INFO: start command begin [BACKREST-VERSION]: --config=[TEST_PATH]/db-master/pgbackrest.conf --lock-path=[TEST_PATH]/db-master/lock --log-level-console=detail --log-level-file=[LOG-LEVEL-FILE] --log-level-stderr=off --log-path=[TEST_PATH]/db-master/log[] --no-log-timestamp --repo1-path=[TEST_PATH]/db-master/repo --stanza=db
Added a predictable log entry to signal that a command has completed successfully. For example a backup ends successfully with: INFO: backup command end: completed successfully.
2016-12-04 21:44:53 +02:00
P00 INFO: start command end: completed successfully
Implemented issue #129: Stop/Start Commands. Implemented issue #136: Add stderr back into protocol layer. There were also miscellaneous bug fixes and improvements to buffering in IO.pm.
2015-10-08 17:43:56 +02:00
Major refactor of the test suite. * Make the code more modular and object-oriented. * Multiple Docker containers can now be created for a single test to simulate more realistic environments.
2016-06-24 14:12:58 +02:00
start all stanzas (db-master host)
Rename test paths for clarity. This was worked out as part of the test suite refactor [c8f806a] but not committed with it because of the large number of expect logs changes involved. Keeping them separate made it easier to audit the changes in the refactor.
2016-06-24 16:54:31 +02:00
> [CONTAINER-EXEC] db-master [BACKREST-BIN] --config=[TEST_PATH]/db-master/pgbackrest.conf start
Implemented issue #129: Stop/Start Commands. Implemented issue #136: Add stderr back into protocol layer. There were also miscellaneous bug fixes and improvements to buffering in IO.pm.
2015-10-08 17:43:56 +02:00
------------------------------------------------------------------------------------------------------------------------------------
Integration test improvements for disk and memory efficiency. Set log-level-file=off when more that one test will run. In this case is it impossible to see the logs anyway since they will be automatically cleaned up after the test. This improves performance pretty dramatically since trace-level logging is expensive. If a singe integration test is run then log-level-file is trace by default but can be changed with the --log-level-test-file option. Reduce buffer-size to 64k to save memory during testing and allow more processes to run in parallel. Update log replacement rules so that these options can change without affecting expect logs.
2019-12-17 22:23:07 +02:00
P00 INFO: start command begin [BACKREST-VERSION]: --config=[TEST_PATH]/db-master/pgbackrest.conf --lock-path=[TEST_PATH]/db-master/lock --log-level-console=detail --log-level-file=[LOG-LEVEL-FILE] --log-level-stderr=off --log-path=[TEST_PATH]/db-master/log[] --no-log-timestamp --repo1-path=[TEST_PATH]/db-master/repo
Added a predictable log entry to signal that a command has completed successfully. For example a backup ends successfully with: INFO: backup command end: completed successfully.
2016-12-04 21:44:53 +02:00
P00 INFO: start command end: completed successfully
Implemented issue #129: Stop/Start Commands. Implemented issue #136: Add stderr back into protocol layer. There were also miscellaneous bug fixes and improvements to buffering in IO.pm.
2015-10-08 17:43:56 +02:00
Major refactor of the test suite. * Make the code more modular and object-oriented. * Multiple Docker containers can now be created for a single test to simulate more realistic environments.
2016-06-24 14:12:58 +02:00
start all stanzas (db-master host)
Rename test paths for clarity. This was worked out as part of the test suite refactor [c8f806a] but not committed with it because of the large number of expect logs changes involved. Keeping them separate made it easier to audit the changes in the refactor.
2016-06-24 16:54:31 +02:00
> [CONTAINER-EXEC] db-master [BACKREST-BIN] --config=[TEST_PATH]/db-master/pgbackrest.conf start
Implemented issue #129: Stop/Start Commands. Implemented issue #136: Add stderr back into protocol layer. There were also miscellaneous bug fixes and improvements to buffering in IO.pm.
2015-10-08 17:43:56 +02:00
------------------------------------------------------------------------------------------------------------------------------------
Integration test improvements for disk and memory efficiency. Set log-level-file=off when more that one test will run. In this case is it impossible to see the logs anyway since they will be automatically cleaned up after the test. This improves performance pretty dramatically since trace-level logging is expensive. If a singe integration test is run then log-level-file is trace by default but can be changed with the --log-level-test-file option. Reduce buffer-size to 64k to save memory during testing and allow more processes to run in parallel. Update log replacement rules so that these options can change without affecting expect logs.
2019-12-17 22:23:07 +02:00
P00 INFO: start command begin [BACKREST-VERSION]: --config=[TEST_PATH]/db-master/pgbackrest.conf --lock-path=[TEST_PATH]/db-master/lock --log-level-console=detail --log-level-file=[LOG-LEVEL-FILE] --log-level-stderr=off --log-path=[TEST_PATH]/db-master/log[] --no-log-timestamp --repo1-path=[TEST_PATH]/db-master/repo
Show Process ID in log output instead of filtering it out with the timestamp.
2016-09-14 23:30:57 +02:00
P00 WARN: stop file does not exist
Added a predictable log entry to signal that a command has completed successfully. For example a backup ends successfully with: INFO: backup command end: completed successfully.
2016-12-04 21:44:53 +02:00
P00 INFO: start command end: completed successfully
Implemented issue #129: Stop/Start Commands. Implemented issue #136: Add stderr back into protocol layer. There were also miscellaneous bug fixes and improvements to buffering in IO.pm.
2015-10-08 17:43:56 +02:00
Major refactor of the test suite. * Make the code more modular and object-oriented. * Multiple Docker containers can now be created for a single test to simulate more realistic environments.
2016-06-24 14:12:58 +02:00
full backup - resume (db-master host)
Remove obsolete integration tests from mock/all. The protocol timeout tests have been superceded by unit tests. The TEST_BACKUP_RESUME test point was incorrectly included into a number of tests, probably a copy pasto. It didn't hurt anything but it did add 200ms to each test where it appeared. Catalog and control version tests were redundant. The database version and system id tests covered the important code paths and the C code gets these values from a lookup table. Finally, fix an incomplete update to the backup.info file while munging for tests.
2019-11-21 23:06:27 +02:00
> [CONTAINER-EXEC] db-master [BACKREST-BIN] --config=[TEST_PATH]/db-master/pgbackrest.conf --no-online --force --checksum-page --delta --type=full --stanza=db backup
Log testing can now be enabled for certain deterministic tests. This works by comparing the generated logs against a previous copy. Currently only enabled for the backup/synthetic tests.
2015-05-05 19:08:48 +02:00
------------------------------------------------------------------------------------------------------------------------------------
Add infrastructure for multiple compression type support. Add compress-type option and deprecate compress option. Since the compress option is boolean it won't work with multiple compression types. Add logic to cfgLoadUpdateOption() to update compress-type if it is not set directly. The compress option should no longer be referenced outside the cfgLoadUpdateOption() function. Add common/compress/helper module to contain interface functions that work with multiple compression types. Code outside this module should no longer call specific compression drivers, though it may be OK to reference a specific compression type using the new interface (e.g., saving backup history files in gz format). Unit tests only test compression using the gz format because other formats may not be available in all builds. It is the job of integration tests to exercise all compression types. Additional compression types will be added in future commits.
2020-03-06 21:41:03 +02:00
P00 INFO: backup command begin [BACKREST-VERSION]: --buffer-size=[BUFFER-SIZE] --checksum-page --compress-level=3 --compress-type=none --config=[TEST_PATH]/db-master/pgbackrest.conf --db-timeout=45 --delta --exclude=postgresql.auto.conf --exclude=pg_log/ --exclude=pg_log2 --exclude=apipe --force --lock-path=[TEST_PATH]/db-master/lock --log-level-console=detail --log-level-file=[LOG-LEVEL-FILE] --log-level-stderr=off --log-path=[TEST_PATH]/db-master/log[] --no-log-timestamp --no-online --pg1-path=[TEST_PATH]/db-master/db/base --protocol-timeout=60 --repo1-path=[TEST_PATH]/db-master/repo --stanza=db --start-fast --type=full
Add time-based retention for full backups. The --repo-retention-full-type option allows retention of full backups based on a time period, specified in days. The new option will default to 'count' and therefore will not affect current installations. Setting repo-retention-full-type to 'time' will allow the user to use a time period, in days, to indicate full backup retention. Using this method, a full backup can be expired only if the time the backup completed is older than the number of days set with repo-retention-full (calculated from the moment the 'expire' command is run) and at least one full backup meets the retention period. If archive retention has not been configured, then the default settings will expire archives that are prior to the oldest retained full backup. For example, if there are three full backups ending in times that are 25 days old (F1), 20 days old (F2) and 10 days old (F3), then if the full retention period is 15 days, then only F1 will be expired; F2 will be retained because F1 is not at least 15 days old.
2020-05-08 21:25:03 +02:00
P00 WARN: option 'repo1-retention-full' is not set for 'repo1-retention-full-type=count', the repository may run out of space
Rename retention-* options to repo-retention-*.
2018-02-19 22:07:24 +02:00
HINT: to retain full backups indefinitely (without warning), set option 'repo1-retention-full' to the maximum.
Update missing manifest warning in BackupInfo. This brings the Perl message in line with C to reduce expect log churn.
2019-11-25 15:51:28 +02:00
P00 WARN: backup '[BACKUP-FULL-1]' missing manifest removed from backup.info
Show Process ID in log output instead of filtering it out with the timestamp.
2016-09-14 23:30:57 +02:00
P00 WARN: --no-online passed and postmaster.pid exists but --force was passed so backup will continue though it looks like the postmaster is running and the backup will probably not be consistent
The backup command is implemented entirely in C. For the most part this is a direct migration of the Perl code into C except as noted below. A backup can now be initiated from a linked directory. The link will not be stored in the manifest or recreated on restore. If a link or directory does not already exist in the restore location then a directory will be created. The logic for creating backup labels has been improved and it should no longer be possible to get a backup label earlier than the latest backup even with timezone changes or clock skew. This has never been an issue in the field that we know of, but we found it in testing. For online backups all times are fetched from the PostgreSQL primary host (before only copy start was). This doesn't affect backup integrity but it does prevent clock skew between hosts affecting backup duration reporting. Archive copy now works as expected when the archive and backup have different compression settings, i.e. when one is compressed and the other is not. This was a long-standing bug in the Perl code. Resume will now work even if hardlink settings have been changed. Reviewed by Cynthia Shang.
2019-12-14 00:14:26 +02:00
P00 INFO: exclude '[TEST_PATH]/db-master/db/base/apipe' from backup using 'apipe' exclusion
P00 INFO: exclude contents of '[TEST_PATH]/db-master/db/base/pg_log' from backup using 'pg_log/' exclusion
P00 INFO: exclude '[TEST_PATH]/db-master/db/base/pg_log2' from backup using 'pg_log2' exclusion
P00 INFO: exclude '[TEST_PATH]/db-master/db/base/postgresql.auto.conf' from backup using 'postgresql.auto.conf' exclusion
P00 WARN: resumable backup [BACKUP-FULL-2] of same type exists -- remove invalid files and resume
P00 DETAIL: remove file '[TEST_PATH]/db-master/repo/backup/db/[BACKUP-FULL-2]/file.tmp' from resumed backup (missing in manifest)
P00 DETAIL: remove file '[TEST_PATH]/db-master/repo/backup/db/[BACKUP-FULL-2]/pg_data/PG_VERSION' from resumed backup (no checksum in resumed manifest)
P00 DETAIL: remove file '[TEST_PATH]/db-master/repo/backup/db/[BACKUP-FULL-2]/pg_data/special-!_.*'()&!@;:+,?' from resumed backup (zero size)
Exclude temporary and unlogged relation (table/index) files from backup. Implemented using the same logic as the patches adding this feature to PostgreSQL, 8694cc96 and 920a5e50. Temporary relation exclusion is enabled in PostgreSQL ≥ 9.0. Unlogged relation exclusion is enabled in PostgreSQL ≥ 9.1, where the feature was introduced. Contributed by Cynthia Shang.
2018-07-31 00:53:34 +02:00
P01 DETAIL: checksum resumed file [TEST_PATH]/db-master/db/base/base/32768/33001 (64KB, 33%) checksum 6bf316f11d28c28914ea9be92c00de9bea6d9a6b
Fix incorrect offline upper bound for ignoring page checksum errors. For offline backups the upper bound was being set to 0x0000FFFF0000FFFF rather than UINT64_MAX. This meant that page checksum errors might be ignored for databases with a lot of past WAL in offline mode. Online mode is not affected since the upper bound is retrieved from pg_start_backup().
2019-07-11 15:13:56 +02:00
P01 DETAIL: checksum resumed file [TEST_PATH]/db-master/db/base/base/32768/44000_init (32KB, 49%) checksum 7a16d165e4775f7c92e8cdf60c0af57313f0bf90
P01 DETAIL: checksum resumed file [TEST_PATH]/db-master/db/base/base/32768/33000.32767 (32KB, 66%) checksum 6e99b589e550e68e934fd235ccba59fe5b592a9e
P01 DETAIL: checksum resumed file [TEST_PATH]/db-master/db/base/base/32768/33000 (32KB, 83%) checksum 7a16d165e4775f7c92e8cdf60c0af57313f0bf90
Exclude temporary and unlogged relation (table/index) files from backup. Implemented using the same logic as the patches adding this feature to PostgreSQL, 8694cc96 and 920a5e50. Temporary relation exclusion is enabled in PostgreSQL ≥ 9.0. Unlogged relation exclusion is enabled in PostgreSQL ≥ 9.1, where the feature was introduced. Contributed by Cynthia Shang.
2018-07-31 00:53:34 +02:00
P01 DETAIL: checksum resumed file [TEST_PATH]/db-master/db/base/base/16384/17000 (16KB, 91%) checksum e0101dd8ffb910c9c202ca35b5f828bcb9697bed
Support configurable WAL segment size. PostgreSQL 11 introduces configurable WAL segment sizes, from 1MB to 1GB. There are two areas that needed to be updated to support this: building the archive-get queue and checking that WAL has been archived after a backup. Both operations require the WAL segment size to properly build a list. Checking the archive after a backup is still implemented in Perl and has an active database connection, so just get the WAL segment size from the database. The archive-get command does not have a connection to the database, so get the WAL segment size from pg_control instead. This requires a deeper inspection of pg_control than has been done in the past, so it seemed best to copy the relevant data structures from each version of PostgreSQL and build a generic interface layer to address them. While this approach is a bit verbose, it has the advantage of being relatively simple, and can easily be updated for new versions of PostgreSQL. Since the integration tests generate pg_control files for testing, teach Perl how to generate files with the correct offsets for both 32-bit and 64-bit architectures.
2018-09-25 11:24:42 +02:00
P01 DETAIL: checksum resumed file [TEST_PATH]/db-master/db/base/global/pg_control (8KB, 95%) checksum 4c77c900f7af0d9ab13fa9982051a42e0b637f6c
Add more realistic data files to synthetic backup and restore tests.
2016-12-10 16:13:55 +02:00
P01 DETAIL: checksum resumed file [TEST_PATH]/db-master/db/base/base/1/12000 (8KB, 99%) checksum 22c98d248ff548311eda88559e4a8405ed77c003
For simplicity, the pg_control file is now copied with the rest of the files instead of by itself of at the end of the process. The backup command does not require this behavior and the restore copies to a temporary file which is renamed at the end of the restore.
2016-11-30 21:23:05 +02:00
P01 DETAIL: checksum resumed file [TEST_PATH]/db-master/db/base/postgresql.conf (21B, 99%) checksum 6721d92c9fcdf4248acff1f9a1377127d9064807
Fix failure in manifest build when two or more files in PGDATA are linked to the same directory. Reported by Vitaliy Kukharik.
2018-05-02 18:19:54 +02:00
P01 DETAIL: checksum resumed file [TEST_PATH]/db-master/db/base/pg_hba.conf (9B, 99%) checksum dd4cea0cae348309f9de28ad4ded8ee2cc2e6d5b
Add checksum delta for incremental backups. Use checksums rather than timestamps to determine if files have changed. This is useful in cases where the timestamps may not be trustworthy, e.g. when performing an incremental after failing over to a standby. If checksum delta is enabled then checksums will be used for verification of resumed backups, even if they are full. Resumes have always used checksums to verify the files in the repository, enabling delta performs checksums on the database files as well. Note that the user must manually enable this feature in cases were it would be useful or just keep in enabled all the time. A future commit will address automatically enabling the feature in cases where it seems likely to be useful. Contributed by Cynthia Shang.
2018-09-19 17:12:45 +02:00
P01 INFO: backup file [TEST_PATH]/db-master/db/base/changecontent.txt (7B, 99%) checksum 238a131a3e8eb98d1fc5b27d882ca40b7618fd2a
Add more realistic data files to synthetic backup and restore tests.
2016-12-10 16:13:55 +02:00
P01 DETAIL: checksum resumed file [TEST_PATH]/db-master/db/base/pg_stat/global.stat (5B, 99%) checksum e350d5ce0153f3e22d5db21cf2a4eff00f3ee877
Add checksum delta for incremental backups. Use checksums rather than timestamps to determine if files have changed. This is useful in cases where the timestamps may not be trustworthy, e.g. when performing an incremental after failing over to a standby. If checksum delta is enabled then checksums will be used for verification of resumed backups, even if they are full. Resumes have always used checksums to verify the files in the repository, enabling delta performs checksums on the database files as well. Note that the user must manually enable this feature in cases were it would be useful or just keep in enabled all the time. A future commit will address automatically enabling the feature in cases where it seems likely to be useful. Contributed by Cynthia Shang.
2018-09-19 17:12:45 +02:00
P01 INFO: backup file [TEST_PATH]/db-master/db/base/changetime.txt (4B, 99%) checksum 88087292ed82e26f3eb824d0bffc05ccf7a30f8d
For simplicity, the pg_control file is now copied with the rest of the files instead of by itself of at the end of the process. The backup command does not require this behavior and the restore copies to a temporary file which is renamed at the end of the restore.
2016-11-30 21:23:05 +02:00
P01 DETAIL: checksum resumed file [TEST_PATH]/db-master/db/base/base/32768/PG_VERSION (3B, 99%) checksum 184473f470864e067ee3a22e64b47b0a1c356f29
P01 DETAIL: checksum resumed file [TEST_PATH]/db-master/db/base/base/16384/PG_VERSION (3B, 99%) checksum 184473f470864e067ee3a22e64b47b0a1c356f29
P01 DETAIL: checksum resumed file [TEST_PATH]/db-master/db/base/base/1/PG_VERSION (3B, 99%) checksum 184473f470864e067ee3a22e64b47b0a1c356f29
Allow zero-size files in backup manifest to reference a prior manifest regardless of timestamp delta. Contributed by Cynthia Shang.
2018-08-24 22:50:33 +02:00
P01 INFO: backup file [TEST_PATH]/db-master/db/base/PG_VERSION (3B, 100%) checksum 184473f470864e067ee3a22e64b47b0a1c356f29
Improve support for special characters in filenames. % characters caused issues in backup/restore due to filenames being appended directly into a format string. Reserved XML characters (<>&') caused issues in the S3 driver due to improper escaping. Add a file with all common special characters to regression testing.
2018-09-10 16:54:34 +02:00
P01 INFO: backup file [TEST_PATH]/db-master/db/base/zero_from_start (0B, 100%)
Use minio for integration testing. ScalityS3 has not received any maintenance in years and is slow to start which is bad for testing. Replace it with minio which starts quickly and ships as a single executable or a tiny container. Minio has stricter limits on allowable characters but should still provide enough coverage to show that our encoding is working correctly. This commit also includes the upgrade to openssl 1.1.1 in the Ubuntu 18.04 container.
2019-07-03 04:20:35 +02:00
P01 INFO: backup file [TEST_PATH]/db-master/db/base/special-!_.*'()&!@;:+,? (0B, 100%)
Exclude temporary and unlogged relation (table/index) files from backup. Implemented using the same logic as the patches adding this feature to PostgreSQL, 8694cc96 and 920a5e50. Temporary relation exclusion is enabled in PostgreSQL ≥ 9.0. Unlogged relation exclusion is enabled in PostgreSQL ≥ 9.1, where the feature was introduced. Contributed by Cynthia Shang.
2018-07-31 00:53:34 +02:00
P00 INFO: full backup size = 192KB
Show Process ID in log output instead of filtering it out with the timestamp.
2016-09-14 23:30:57 +02:00
P00 INFO: new backup label = [BACKUP-FULL-2]
Added a predictable log entry to signal that a command has completed successfully. For example a backup ends successfully with: INFO: backup command end: completed successfully.
2016-12-04 21:44:53 +02:00
P00 INFO: backup command end: completed successfully
Integration test improvements for disk and memory efficiency. Set log-level-file=off when more that one test will run. In this case is it impossible to see the logs anyway since they will be automatically cleaned up after the test. This improves performance pretty dramatically since trace-level logging is expensive. If a singe integration test is run then log-level-file is trace by default but can be changed with the --log-level-test-file option. Reduce buffer-size to 64k to save memory during testing and allow more processes to run in parallel. Update log replacement rules so that these options can change without affecting expect logs.
2019-12-17 22:23:07 +02:00
P00 INFO: expire command begin [BACKREST-VERSION]: --buffer-size=[BUFFER-SIZE] --config=[TEST_PATH]/db-master/pgbackrest.conf --lock-path=[TEST_PATH]/db-master/lock --log-level-console=detail --log-level-file=[LOG-LEVEL-FILE] --log-level-stderr=off --log-path=[TEST_PATH]/db-master/log[] --no-log-timestamp --repo1-path=[TEST_PATH]/db-master/repo --stanza=db
Rename retention-* options to repo-retention-*.
2018-02-19 22:07:24 +02:00
P00 INFO: option 'repo1-retention-archive' is not set - archive logs will not be expired
Added a predictable log entry to signal that a command has completed successfully. For example a backup ends successfully with: INFO: backup command end: completed successfully.
2016-12-04 21:44:53 +02:00
P00 INFO: expire command end: completed successfully
Log testing can now be enabled for certain deterministic tests. This works by comparing the generated logs against a previous copy. Currently only enabled for the backup/synthetic tests.
2015-05-05 19:08:48 +02:00
Rename test paths for clarity. This was worked out as part of the test suite refactor [c8f806a] but not committed with it because of the large number of expect logs changes involved. Keeping them separate made it easier to audit the changes in the refactor.
2016-06-24 16:54:31 +02:00
+ supplemental file: [TEST_PATH]/db-master/pgbackrest.conf
----------------------------------------------------------
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
[db]
Rename db-* options to pg-* and backup-* options to repo-* to improve consistency. * repo-* options are now indexed although only one is allowed. * List deprecated option names in documentation and command-line help.
2018-02-04 01:27:38 +02:00
pg1-path=[TEST_PATH]/db-master/db/base
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
New simpler configuration and consistent project/exe/path naming. * The repo-path option now always refers to the repository where backups and archive are stored, whether local or remote, so the repo-remote-path option has been removed. The new spool-path option can be used to define a location for queueing WAL segments when archiving asynchronously. Otherwise, a local repository is no longer required. * Implemented a new config format which should be far simpler to use. See the User Guide and Configuration Reference for details but for a simple configuration all options can now be placed in the stanza section. Options that are shared between stanzas can be placed in the [global] section. More complex configurations can still make use of command sections though this should be a rare use case. * The default configuration filename is now pgbackrest.conf instead of pg_backrest.conf. This was done for consistency with other naming changes but also to prevent old config files from being loaded accidentally. * The default repository name was changed from /var/lib/backup to /var/lib/pgbackrest. * Lock files are now stored in /tmp/pgbackrest by default. These days /run/pgbackrest would be the preferred location but that would require init scripts which are not part of this release. The lock-path option can be used to configure the lock directory. * Log files are now stored in /var/log/pgbackrest by default and no longer have the date appended so they can be managed with logrotate. The log-path option can be used to configure the lock directory. * Executable filename changed from pg_backrest to pgbackrest.
2016-04-14 15:30:54 +02:00
[global]
Integration test improvements for disk and memory efficiency. Set log-level-file=off when more that one test will run. In this case is it impossible to see the logs anyway since they will be automatically cleaned up after the test. This improves performance pretty dramatically since trace-level logging is expensive. If a singe integration test is run then log-level-file is trace by default but can be changed with the --log-level-test-file option. Reduce buffer-size to 64k to save memory during testing and allow more processes to run in parallel. Update log replacement rules so that these options can change without affecting expect logs.
2019-12-17 22:23:07 +02:00
buffer-size=[BUFFER-SIZE]
Fix custom settings for compress-level option being ignored. Reported by Jens Wilke.
2017-10-23 00:05:46 +02:00
compress-level=3
Add infrastructure for multiple compression type support. Add compress-type option and deprecate compress option. Since the compress option is boolean it won't work with multiple compression types. Add logic to cfgLoadUpdateOption() to update compress-type if it is not set directly. The compress option should no longer be referenced outside the cfgLoadUpdateOption() function. Add common/compress/helper module to contain interface functions that work with multiple compression types. Code outside this module should no longer call specific compression drivers, though it may be OK to reference a specific compression type using the new interface (e.g., saving backup history files in gz format). Unit tests only test compression using the gz format because other formats may not be available in all builds. It is the job of integration tests to exercise all compression types. Additional compression types will be added in future commits.
2020-03-06 21:41:03 +02:00
compress-type=none
Storage and IO layer refactor: Refactor storage layer to allow for new repository filesystems using drivers. (Reviewed by Cynthia Shang.) Refactor IO layer to allow for new compression formats, checksum types, and other capabilities using filters. (Reviewed by Cynthia Shang.)
2017-06-09 23:51:41 +02:00
db-timeout=45
lock-path=[TEST_PATH]/db-master/lock
Reduce log level for all expect tests to detail. The C code is designed to be efficient rather than deterministic at the debug log level. As we move more testing from integration to unit tests it makes less sense to try and maintain the expect logs at this log level. Most of the expect logs have already been moved to detail level but mock/all still had tests at debug level. Change the logging defaults in the config file and remove as many references to log-level-console as possible.
2019-05-23 00:23:44 +02:00
log-level-console=detail
Integration test improvements for disk and memory efficiency. Set log-level-file=off when more that one test will run. In this case is it impossible to see the logs anyway since they will be automatically cleaned up after the test. This improves performance pretty dramatically since trace-level logging is expensive. If a singe integration test is run then log-level-file is trace by default but can be changed with the --log-level-test-file option. Reduce buffer-size to 64k to save memory during testing and allow more processes to run in parallel. Update log replacement rules so that these options can change without affecting expect logs.
2019-12-17 22:23:07 +02:00
log-level-file=[LOG-LEVEL-FILE]
Added the log-level-stderr option. Controls whether console log messages are sent to stderr or stdout. By default this is set to warn which represents a change in behavior from previous versions, even though it may be more intuitive. Setting log-level-stderr=off will preserve the old behavior. Suggested by Sascha Biberhofer.
2016-10-05 15:09:30 +02:00
log-level-stderr=off
Storage and IO layer refactor: Refactor storage layer to allow for new repository filesystems using drivers. (Reviewed by Cynthia Shang.) Refactor IO layer to allow for new compression formats, checksum types, and other capabilities using filters. (Reviewed by Cynthia Shang.)
2017-06-09 23:51:41 +02:00
log-path=[TEST_PATH]/db-master/log
Integration test improvements for disk and memory efficiency. Set log-level-file=off when more that one test will run. In this case is it impossible to see the logs anyway since they will be automatically cleaned up after the test. This improves performance pretty dramatically since trace-level logging is expensive. If a singe integration test is run then log-level-file is trace by default but can be changed with the --log-level-test-file option. Reduce buffer-size to 64k to save memory during testing and allow more processes to run in parallel. Update log replacement rules so that these options can change without affecting expect logs.
2019-12-17 22:23:07 +02:00
log-subprocess=[LOG-SUBPROCESS]
Set log-timestamp=n for integration tests. This means less filtering of logs needs to be done and new timestamps can be added without adding new filters.
2018-05-11 17:24:38 +02:00
log-timestamp=n
Storage and IO layer refactor: Refactor storage layer to allow for new repository filesystems using drivers. (Reviewed by Cynthia Shang.) Refactor IO layer to allow for new compression formats, checksum types, and other capabilities using filters. (Reviewed by Cynthia Shang.)
2017-06-09 23:51:41 +02:00
protocol-timeout=60
Rename db-* options to pg-* and backup-* options to repo-* to improve consistency. * repo-* options are now indexed although only one is allowed. * List deprecated option names in documentation and command-line help.
2018-02-04 01:27:38 +02:00
repo1-path=[TEST_PATH]/db-master/repo
Consolidate archive-push tests into new archive test.
2017-06-21 14:38:53 +02:00
spool-path=[TEST_PATH]/db-master/spool
New simpler configuration and consistent project/exe/path naming. * The repo-path option now always refers to the repository where backups and archive are stored, whether local or remote, so the repo-remote-path option has been removed. The new spool-path option can be used to define a location for queueing WAL segments when archiving asynchronously. Otherwise, a local repository is no longer required. * Implemented a new config format which should be far simpler to use. See the User Guide and Configuration Reference for details but for a simple configuration all options can now be placed in the stanza section. Options that are shared between stanzas can be placed in the [global] section. More complex configurations can still make use of command sections though this should be a rare use case. * The default configuration filename is now pgbackrest.conf instead of pg_backrest.conf. This was done for consistency with other naming changes but also to prevent old config files from being loaded accidentally. * The default repository name was changed from /var/lib/backup to /var/lib/pgbackrest. * Lock files are now stored in /tmp/pgbackrest by default. These days /run/pgbackrest would be the preferred location but that would require init scripts which are not part of this release. The lock-path option can be used to configure the lock directory. * Log files are now stored in /var/log/pgbackrest by default and no longer have the date appended so they can be managed with logrotate. The log-path option can be used to configure the lock directory. * Executable filename changed from pg_backrest to pgbackrest.
2016-04-14 15:30:54 +02:00
[global:backup]
archive-copy=y
Allow arbitrary directories and/or files to be excluded from a backup. Misuse of this feature can lead to inconsistent backups so read the --exclude documentation carefully before using.
2018-08-27 21:51:05 +02:00
exclude=postgresql.auto.conf
exclude=pg_log/
exclude=pg_log2
Fix exclusions for special files. Prior to 2.16 the Perl manifest code would skip any file that began with a dot. This was not intentional but it allowed PostgreSQL socket files to be located in the data directory. The new C code in 2.16 did not have this unintentional exclusion so socket files in the data directory caused errors. Worse, the file type error was being thrown before the exclusion check so there was really no way around the issue except to move the socket files out of the data directory. Special file types (e.g. socket, pipe) will now be automatically skipped and a warning logged to notify the user of the exclusion. The warning can be suppressed with an explicit --exclude. Reported by CluelessTechnologist, Janis Puris, Rachid Broum.
2019-08-23 13:47:54 +02:00
exclude=apipe
New simpler configuration and consistent project/exe/path naming. * The repo-path option now always refers to the repository where backups and archive are stored, whether local or remote, so the repo-remote-path option has been removed. The new spool-path option can be used to define a location for queueing WAL segments when archiving asynchronously. Otherwise, a local repository is no longer required. * Implemented a new config format which should be far simpler to use. See the User Guide and Configuration Reference for details but for a simple configuration all options can now be placed in the stanza section. Options that are shared between stanzas can be placed in the [global] section. More complex configurations can still make use of command sections though this should be a rare use case. * The default configuration filename is now pgbackrest.conf instead of pg_backrest.conf. This was done for consistency with other naming changes but also to prevent old config files from being loaded accidentally. * The default repository name was changed from /var/lib/backup to /var/lib/pgbackrest. * Lock files are now stored in /tmp/pgbackrest by default. These days /run/pgbackrest would be the preferred location but that would require init scripts which are not part of this release. The lock-path option can be used to configure the lock directory. * Log files are now stored in /var/log/pgbackrest by default and no longer have the date appended so they can be managed with logrotate. The log-path option can be used to configure the lock directory. * Executable filename changed from pg_backrest to pgbackrest.
2016-04-14 15:30:54 +02:00
start-fast=y
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
Rename test paths for clarity. This was worked out as part of the test suite refactor [c8f806a] but not committed with it because of the large number of expect logs changes involved. Keeping them separate made it easier to audit the changes in the refactor.
2016-06-24 16:54:31 +02:00
+ supplemental file: [TEST_PATH]/db-master/repo/backup/db/[BACKUP-FULL-2]/backup.manifest
-----------------------------------------------------------------------------------------
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
[backrest]
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
backrest-format=5
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
backrest-version="[VERSION-1]"
[backup]
backup-label="[BACKUP-FULL-2]"
backup-timestamp-copy-start=[TIMESTAMP]
backup-timestamp-start=[TIMESTAMP]
backup-timestamp-stop=[TIMESTAMP]
backup-type="full"
[backup:db]
Database version refactoring. * Refactor db version constants into a separate module. * Update synthetic backup tests to PostgreSQL 9.4.
2016-08-12 04:35:24 +02:00
db-catalog-version=201409291
db-control-version=942
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
db-id=1
Automate generation of WAL and pg_control test files. The existing static files would not work with 32-bit or big-endian systems so create functions to generate these files dynamically rather than creating a bunch of new static files.
2017-11-19 03:02:54 +02:00
db-system-id=1000000000000000094
Database version refactoring. * Refactor db version constants into a separate module. * Update synthetic backup tests to PostgreSQL 9.4.
2016-08-12 04:35:24 +02:00
db-version="9.4"
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
[backup:option]
Set archive-check option in manifest correctly when offline. Archive check does not run when in offline backup mode but the option was set to true in the manifest. It's harmless since these options are informational only but it could cause confusion when debugging.
2019-11-28 15:27:21 +02:00
option-archive-check=false
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
option-archive-copy=true
Backup from a standby cluster. A connection to the primary cluster is still required to start/stop the backup and copy files that are not replicated, but the vast majority of files are copied from the standby in order to reduce load on the master.
2016-08-25 17:25:46 +02:00
option-backup-standby=false
Integration test improvements for disk and memory efficiency. Set log-level-file=off when more that one test will run. In this case is it impossible to see the logs anyway since they will be automatically cleaned up after the test. This improves performance pretty dramatically since trace-level logging is expensive. If a singe integration test is run then log-level-file is trace by default but can be changed with the --log-level-test-file option. Reduce buffer-size to 64k to save memory during testing and allow more processes to run in parallel. Update log replacement rules so that these options can change without affecting expect logs.
2019-12-17 22:23:07 +02:00
option-buffer-size=[BUFFER-SIZE]
Added the --checksum-page option. This option allows pgBackRest to validate page checksums in data files when checksums are enabled on PostgreSQL >= 9.3. Note that this functionality requires a C library which may not initially be available in OS packages. The option will automatically be enabled when the library is present and checksums are enabled on the cluster.
2016-12-13 01:54:07 +02:00
option-checksum-page=true
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
option-compress=false
Add additional options to backup.manifest for debugging purposes. Add the buffer-size, compress-level, compress-level-network, and process-max options to the backup:option section in backup.manifest to aid in debugging. It may also make sense to propagate these options up to backup.info so they can be displayed in the info command, but for now this is deemed sufficient. Contributed by blogh.
2019-03-10 11:03:52 +02:00
option-compress-level=3
option-compress-level-network=3
Add infrastructure for multiple compression type support. Add compress-type option and deprecate compress option. Since the compress option is boolean it won't work with multiple compression types. Add logic to cfgLoadUpdateOption() to update compress-type if it is not set directly. The compress option should no longer be referenced outside the cfgLoadUpdateOption() function. Add common/compress/helper module to contain interface functions that work with multiple compression types. Code outside this module should no longer call specific compression drivers, though it may be OK to reference a specific compression type using the new interface (e.g., saving backup history files in gz format). Unit tests only test compression using the gz format because other formats may not be available in all builds. It is the job of integration tests to exercise all compression types. Additional compression types will be added in future commits.
2020-03-06 21:41:03 +02:00
option-compress-type="none"
Add checksum delta for incremental backups. Use checksums rather than timestamps to determine if files have changed. This is useful in cases where the timestamps may not be trustworthy, e.g. when performing an incremental after failing over to a standby. If checksum delta is enabled then checksums will be used for verification of resumed backups, even if they are full. Resumes have always used checksums to verify the files in the repository, enabling delta performs checksums on the database files as well. Note that the user must manually enable this feature in cases were it would be useful or just keep in enabled all the time. A future commit will address automatically enabling the feature in cases where it seems likely to be useful. Contributed by Cynthia Shang.
2018-09-19 17:12:45 +02:00
option-delta=true
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
option-hardlink=false
Closed #167: Rename `--no-start-stop` option to `--no-online`.
2016-02-20 02:32:02 +02:00
option-online=false
Add additional options to backup.manifest for debugging purposes. Add the buffer-size, compress-level, compress-level-network, and process-max options to the backup:option section in backup.manifest to aid in debugging. It may also make sense to propagate these options up to backup.info so they can be displayed in the info command, but for now this is deemed sufficient. Contributed by blogh.
2019-03-10 11:03:52 +02:00
option-process-max=1
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
[backup:target]
Rename test paths for clarity. This was worked out as part of the test suite refactor [c8f806a] but not committed with it because of the large number of expect logs changes involved. Keeping them separate made it easier to audit the changes in the refactor.
2016-06-24 16:54:31 +02:00
pg_data={"path":"[TEST_PATH]/db-master/db/base","type":"path"}
Fix failure in manifest build when two or more files in PGDATA are linked to the same directory. Reported by Vitaliy Kukharik.
2018-05-02 18:19:54 +02:00
pg_data/pg_hba.conf={"file":"pg_hba.conf","path":"../pg_config","type":"link"}
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
pg_data/pg_stat={"path":"../pg_stat","type":"link"}
pg_data/postgresql.conf={"file":"postgresql.conf","path":"../pg_config","type":"link"}
[target:file]
Fixed a issue where tablespaces were copied from the master during standby backup.
2016-09-04 15:19:44 +02:00
pg_data/PG_VERSION={"checksum":"184473f470864e067ee3a22e64b47b0a1c356f29","master":true,"size":[SIZE],"timestamp":[TIMESTAMP-1]}
Added the --checksum-page option. This option allows pgBackRest to validate page checksums in data files when checksums are enabled on PostgreSQL >= 9.3. Note that this functionality requires a C library which may not initially be available in OS packages. The option will automatically be enabled when the library is present and checksums are enabled on the cluster.
2016-12-13 01:54:07 +02:00
pg_data/base/1/12000={"checksum":"22c98d248ff548311eda88559e4a8405ed77c003","checksum-page":true,"size":[SIZE],"timestamp":[TIMESTAMP-1]}
Database version refactoring. * Refactor db version constants into a separate module. * Update synthetic backup tests to PostgreSQL 9.4.
2016-08-12 04:35:24 +02:00
pg_data/base/1/PG_VERSION={"checksum":"184473f470864e067ee3a22e64b47b0a1c356f29","mode":"0660","size":[SIZE],"timestamp":[TIMESTAMP-1],"user":false}
Added the --checksum-page option. This option allows pgBackRest to validate page checksums in data files when checksums are enabled on PostgreSQL >= 9.3. Note that this functionality requires a C library which may not initially be available in OS packages. The option will automatically be enabled when the library is present and checksums are enabled on the cluster.
2016-12-13 01:54:07 +02:00
pg_data/base/16384/17000={"checksum":"e0101dd8ffb910c9c202ca35b5f828bcb9697bed","checksum-page":false,"checksum-page-error":[1],"size":[SIZE],"timestamp":[TIMESTAMP-1]}
Database version refactoring. * Refactor db version constants into a separate module. * Update synthetic backup tests to PostgreSQL 9.4.
2016-08-12 04:35:24 +02:00
pg_data/base/16384/PG_VERSION={"checksum":"184473f470864e067ee3a22e64b47b0a1c356f29","group":false,"size":[SIZE],"timestamp":[TIMESTAMP-1]}
Fix incorrect offline upper bound for ignoring page checksum errors. For offline backups the upper bound was being set to 0x0000FFFF0000FFFF rather than UINT64_MAX. This meant that page checksum errors might be ignored for databases with a lot of past WAL in offline mode. Online mode is not affected since the upper bound is retrieved from pg_start_backup().
2019-07-11 15:13:56 +02:00
pg_data/base/32768/33000={"checksum":"7a16d165e4775f7c92e8cdf60c0af57313f0bf90","checksum-page":true,"size":[SIZE],"timestamp":[TIMESTAMP-1]}
pg_data/base/32768/33000.32767={"checksum":"6e99b589e550e68e934fd235ccba59fe5b592a9e","checksum-page":true,"size":[SIZE],"timestamp":[TIMESTAMP-1]}
Added the --checksum-page option. This option allows pgBackRest to validate page checksums in data files when checksums are enabled on PostgreSQL >= 9.3. Note that this functionality requires a C library which may not initially be available in OS packages. The option will automatically be enabled when the library is present and checksums are enabled on the cluster.
2016-12-13 01:54:07 +02:00
pg_data/base/32768/33001={"checksum":"6bf316f11d28c28914ea9be92c00de9bea6d9a6b","checksum-page":false,"checksum-page-error":[0,[3,5],7],"size":[SIZE],"timestamp":[TIMESTAMP-1]}
Fix incorrect offline upper bound for ignoring page checksum errors. For offline backups the upper bound was being set to 0x0000FFFF0000FFFF rather than UINT64_MAX. This meant that page checksum errors might be ignored for databases with a lot of past WAL in offline mode. Online mode is not affected since the upper bound is retrieved from pg_start_backup().
2019-07-11 15:13:56 +02:00
pg_data/base/32768/44000_init={"checksum":"7a16d165e4775f7c92e8cdf60c0af57313f0bf90","checksum-page":true,"size":[SIZE],"timestamp":[TIMESTAMP-1]}
Database version refactoring. * Refactor db version constants into a separate module. * Update synthetic backup tests to PostgreSQL 9.4.
2016-08-12 04:35:24 +02:00
pg_data/base/32768/PG_VERSION={"checksum":"184473f470864e067ee3a22e64b47b0a1c356f29","size":[SIZE],"timestamp":[TIMESTAMP-1]}
Add checksum delta for incremental backups. Use checksums rather than timestamps to determine if files have changed. This is useful in cases where the timestamps may not be trustworthy, e.g. when performing an incremental after failing over to a standby. If checksum delta is enabled then checksums will be used for verification of resumed backups, even if they are full. Resumes have always used checksums to verify the files in the repository, enabling delta performs checksums on the database files as well. Note that the user must manually enable this feature in cases were it would be useful or just keep in enabled all the time. A future commit will address automatically enabling the feature in cases where it seems likely to be useful. Contributed by Cynthia Shang.
2018-09-19 17:12:45 +02:00
pg_data/changecontent.txt={"checksum":"238a131a3e8eb98d1fc5b27d882ca40b7618fd2a","master":true,"size":[SIZE],"timestamp":[TIMESTAMP-1]}
pg_data/changetime.txt={"checksum":"88087292ed82e26f3eb824d0bffc05ccf7a30f8d","master":true,"size":[SIZE],"timestamp":[TIMESTAMP-1]}
Support configurable WAL segment size. PostgreSQL 11 introduces configurable WAL segment sizes, from 1MB to 1GB. There are two areas that needed to be updated to support this: building the archive-get queue and checking that WAL has been archived after a backup. Both operations require the WAL segment size to properly build a list. Checking the archive after a backup is still implemented in Perl and has an active database connection, so just get the WAL segment size from the database. The archive-get command does not have a connection to the database, so get the WAL segment size from pg_control instead. This requires a deeper inspection of pg_control than has been done in the past, so it seemed best to copy the relevant data structures from each version of PostgreSQL and build a generic interface layer to address them. While this approach is a bit verbose, it has the advantage of being relatively simple, and can easily be updated for new versions of PostgreSQL. Since the integration tests generate pg_control files for testing, teach Perl how to generate files with the correct offsets for both 32-bit and 64-bit architectures.
2018-09-25 11:24:42 +02:00
pg_data/global/pg_control={"checksum":"4c77c900f7af0d9ab13fa9982051a42e0b637f6c","master":true,"size":[SIZE],"timestamp":[TIMESTAMP-2]}
Fix failure in manifest build when two or more files in PGDATA are linked to the same directory. Reported by Vitaliy Kukharik.
2018-05-02 18:19:54 +02:00
pg_data/pg_hba.conf={"checksum":"dd4cea0cae348309f9de28ad4ded8ee2cc2e6d5b","master":true,"size":[SIZE],"timestamp":[TIMESTAMP-2]}
Fixed a issue where tablespaces were copied from the master during standby backup.
2016-09-04 15:19:44 +02:00
pg_data/pg_stat/global.stat={"checksum":"e350d5ce0153f3e22d5db21cf2a4eff00f3ee877","master":true,"size":[SIZE],"timestamp":[TIMESTAMP-2]}
pg_data/postgresql.conf={"checksum":"6721d92c9fcdf4248acff1f9a1377127d9064807","master":true,"size":[SIZE],"timestamp":[TIMESTAMP-2]}
Use minio for integration testing. ScalityS3 has not received any maintenance in years and is slow to start which is bad for testing. Replace it with minio which starts quickly and ships as a single executable or a tiny container. Minio has stricter limits on allowable characters but should still provide enough coverage to show that our encoding is working correctly. This commit also includes the upgrade to openssl 1.1.1 in the Ubuntu 18.04 container.
2019-07-03 04:20:35 +02:00
pg_data/special-!_.*'()&!@;:+,?={"master":true,"size":[SIZE],"timestamp":[TIMESTAMP-1]}
Allow zero-size files in backup manifest to reference a prior manifest regardless of timestamp delta. Contributed by Cynthia Shang.
2018-08-24 22:50:33 +02:00
pg_data/zero_from_start={"master":true,"size":[SIZE],"timestamp":[TIMESTAMP-1]}
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
[target:file:default]
Continuous integration using travis-ci.
2017-02-21 15:59:23 +02:00
group="[GROUP-1]"
Fixed a issue where tablespaces were copied from the master during standby backup.
2016-09-04 15:19:44 +02:00
master=false
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
mode="0600"
Regression tests can now be run as any properly-configured user, not just vagrant.
2016-10-01 19:39:44 +02:00
user="[USER-1]"
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
[target:link]
Fix failure in manifest build when two or more files in PGDATA are linked to the same directory. Reported by Vitaliy Kukharik.
2018-05-02 18:19:54 +02:00
pg_data/pg_hba.conf={"destination":"../pg_config/pg_hba.conf"}
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
pg_data/pg_stat={"destination":"../pg_stat"}
pg_data/postgresql.conf={"destination":"../pg_config/postgresql.conf"}
[target:link:default]
Continuous integration using travis-ci.
2017-02-21 15:59:23 +02:00
group="[GROUP-1]"
Regression tests can now be run as any properly-configured user, not just vagrant.
2016-10-01 19:39:44 +02:00
user="[USER-1]"
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
[target:path]
Fixed an issue with invalid backup groups being set correctly on restore. If the backup cannot map a group to a name it stores the group in the manifest as false then uses either the owner of $PGDATA to set the group during restore or failing that the group of the current user. This logic was not working correctly because the selected group was overwriting the user on restore leaving the group undefined and the user incorrectly set to the group. (Reported by Jeff McCormick.)
2017-11-20 19:46:35 +02:00
pg_data={"group":false,"mode":"0777","user":false}
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
pg_data/base={}
Allow selective restore of databases from a cluster backup. This feature can result in major space and time savings when only specific databases are restored. Unrestored databases will not be accessible but must be manually dropped before they will be removed from the shared catalogue.
2016-05-11 15:21:39 +02:00
pg_data/base/1={}
pg_data/base/16384={}
pg_data/base/32768={}
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
pg_data/global={}
pg_data/pg_clog={}
Exclude directories during backup that are cleaned, recreated, or zeroed by PostgreSQL at startup. These include (depending on the version where they were introduced): pgsql_tmp, pg_dynshmem, pg_notify, pg_replslot, pg_serial, pg_snapshots, pg_stat_tmp, pg_subtrans. The postgresql.auto.conf.tmp file is now excluded in addition to files that were already excluded: backup_label.old, postmaster.opts, postmaster.pid, recovery.conf, recovery.done.
2016-08-16 15:35:16 +02:00
pg_data/pg_dynshmem={}
Allow arbitrary directories and/or files to be excluded from a backup. Misuse of this feature can lead to inconsistent backups so read the --exclude documentation carefully before using.
2018-08-27 21:51:05 +02:00
pg_data/pg_log={}
Exclude directories during backup that are cleaned, recreated, or zeroed by PostgreSQL at startup. These include (depending on the version where they were introduced): pgsql_tmp, pg_dynshmem, pg_notify, pg_replslot, pg_serial, pg_snapshots, pg_stat_tmp, pg_subtrans. The postgresql.auto.conf.tmp file is now excluded in addition to files that were already excluded: backup_label.old, postmaster.opts, postmaster.pid, recovery.conf, recovery.done.
2016-08-16 15:35:16 +02:00
pg_data/pg_notify={}
pg_data/pg_replslot={}
pg_data/pg_serial={}
pg_data/pg_snapshots={}
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
pg_data/pg_stat={}
Exclude directories during backup that are cleaned, recreated, or zeroed by PostgreSQL at startup. These include (depending on the version where they were introduced): pgsql_tmp, pg_dynshmem, pg_notify, pg_replslot, pg_serial, pg_snapshots, pg_stat_tmp, pg_subtrans. The postgresql.auto.conf.tmp file is now excluded in addition to files that were already excluded: backup_label.old, postmaster.opts, postmaster.pid, recovery.conf, recovery.done.
2016-08-16 15:35:16 +02:00
pg_data/pg_stat_tmp={}
pg_data/pg_subtrans={}
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
pg_data/pg_tblspc={}
[target:path:default]
Continuous integration using travis-ci.
2017-02-21 15:59:23 +02:00
group="[GROUP-1]"
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
mode="0700"
Regression tests can now be run as any properly-configured user, not just vagrant.
2016-10-01 19:39:44 +02:00
user="[USER-1]"
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
Move info file checksum to the end of the file. Putting the checksum at the beginning of the file made it impossible to stream the file out when saving. The entire file had to be held in memory while it was checksummed so the checksum could be written at the beginning. Instead place the checksum at the end. This does not break the existing Perl or C code since the read is not order dependent. There are no plans to improve the Perl code to take advantage of this change, but it will make the C implementation more efficient. Reviewed by Cynthia Shang.
2019-08-22 01:45:48 +02:00
[backrest]
backrest-checksum="[CHECKSUM]"
Rename test paths for clarity. This was worked out as part of the test suite refactor [c8f806a] but not committed with it because of the large number of expect logs changes involved. Keeping them separate made it easier to audit the changes in the refactor.
2016-06-24 16:54:31 +02:00
+ supplemental file: [TEST_PATH]/db-master/repo/backup/db/backup.info
---------------------------------------------------------------------
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
[backrest]
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
backrest-format=5
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
backrest-version="[VERSION-1]"
[backup:current]
Set archive-check option in manifest correctly when offline. Archive check does not run when in offline backup mode but the option was set to true in the manifest. It's harmless since these options are informational only but it could cause confusion when debugging.
2019-11-28 15:27:21 +02:00
[BACKUP-FULL-2]={"backrest-format":5,"backrest-version":"[VERSION-1]","backup-archive-start":null,"backup-archive-stop":null,"backup-info-repo-size":[SIZE],"backup-info-repo-size-delta":[DELTA],"backup-info-size":[SIZE],"backup-info-size-delta":[DELTA],"backup-timestamp-start":[TIMESTAMP],"backup-timestamp-stop":[TIMESTAMP],"backup-type":"full","db-id":1,"option-archive-check":false,"option-archive-copy":true,"option-backup-standby":false,"option-checksum-page":true,"option-compress":false,"option-hardlink":false,"option-online":false}
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
[db]
Database version refactoring. * Refactor db version constants into a separate module. * Update synthetic backup tests to PostgreSQL 9.4.
2016-08-12 04:35:24 +02:00
db-catalog-version=201409291
db-control-version=942
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
db-id=1
Automate generation of WAL and pg_control test files. The existing static files would not work with 32-bit or big-endian systems so create functions to generate these files dynamically rather than creating a bunch of new static files.
2017-11-19 03:02:54 +02:00
db-system-id=1000000000000000094
Database version refactoring. * Refactor db version constants into a separate module. * Update synthetic backup tests to PostgreSQL 9.4.
2016-08-12 04:35:24 +02:00
db-version="9.4"
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
[db:history]
Automate generation of WAL and pg_control test files. The existing static files would not work with 32-bit or big-endian systems so create functions to generate these files dynamically rather than creating a bunch of new static files.
2017-11-19 03:02:54 +02:00
1={"db-catalog-version":201409291,"db-control-version":942,"db-system-id":1000000000000000094,"db-version":"9.4"}
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
Improve performance of info file load/save. Info files required three copies in memory to be loaded (the original string, an ini representation, and the final info object). Not only was this memory inefficient but the Ini object does sequential scans when searching for keys making large files very slow to load. This has not been an issue since archive.info and backup.info are very small, but it becomes a big deal when loading manifests with hundreds of thousands of files. Instead of holding copies of the data in memory, use a callback to deliver the ini data directly to the object when loading. Use a similar method for save to avoid having an intermediate copy. Save is a bit complex because sections/keys must be written in alpha order or older versions of pgBackRest will not calculate the correct checksum. Also move the load retry logic to helper functions rather than embedding it in the Info object. This allows for more flexibility in loading and ensures that stack traces will be available when developing unit tests. Reviewed by Cynthia Shang.
2019-09-06 19:48:28 +02:00
[backrest]
backrest-checksum="[CHECKSUM]"
Major refactor of the test suite. * Make the code more modular and object-oriented. * Multiple Docker containers can now be created for a single test to simulate more realistic environments.
2016-06-24 14:12:58 +02:00
restore delta, backup '[BACKUP-FULL-2]' - add and delete files (db-master host)
Rename test paths for clarity. This was worked out as part of the test suite refactor [c8f806a] but not committed with it because of the large number of expect logs changes involved. Keeping them separate made it easier to audit the changes in the refactor.
2016-06-24 16:54:31 +02:00
> [CONTAINER-EXEC] db-master [BACKREST-BIN] --config=[TEST_PATH]/db-master/pgbackrest.conf --delta --set=[BACKUP-FULL-2] --link-all --stanza=db restore
Log testing can now be enabled for certain deterministic tests. This works by comparing the generated logs against a previous copy. Currently only enabled for the backup/synthetic tests.
2015-05-05 19:08:48 +02:00
------------------------------------------------------------------------------------------------------------------------------------
Remove compress/compress-level options from commands where unused. These commands (e.g. restore, archive-get) never used the compress options but allowed them to be passed on the command line. Now they will error when these options are passed on the command line. If these errors occur then remove the unused options.
2020-02-27 19:25:32 +02:00
P00 INFO: restore command begin [BACKREST-VERSION]: --buffer-size=[BUFFER-SIZE] --config=[TEST_PATH]/db-master/pgbackrest.conf --delta --link-all --lock-path=[TEST_PATH]/db-master/lock --log-level-console=detail --log-level-file=[LOG-LEVEL-FILE] --log-level-stderr=off --log-path=[TEST_PATH]/db-master/log[] --no-log-timestamp --pg1-path=[TEST_PATH]/db-master/db/base --protocol-timeout=60 --repo1-path=[TEST_PATH]/db-master/repo --set=[BACKUP-FULL-2] --stanza=db
Show Process ID in log output instead of filtering it out with the timestamp.
2016-09-14 23:30:57 +02:00
P00 INFO: restore backup set [BACKUP-FULL-2]
The restore command is implemented entirely in C. For the most part this is a direct migration of the Perl code into C. There is one important behavioral change with regard to how file permissions are handled. The Perl code tried to set ownership as it was in the manifest even when running as an unprivileged user. This usually just led to errors and frustration. The C code works like this: If a restore is run as a non-root user (the typical scenario) then all files restored will belong to the user/group executing pgBackRest. If existing files are not owned by the executing user/group then an error will result if the ownership cannot be updated to the executing user/group. In that case the file ownership will need to be updated by a privileged user before the restore can be retried. If a restore is run as the root user then pgBackRest will attempt to recreate the ownership recorded in the manifest when the backup was made. Only user/group names are stored in the manifest so the same names must exist on the restore host for this to work. If the user/group name cannot be found locally then the user/group of the PostgreSQL data directory will be used and finally root if the data directory user/group cannot be mapped to a name. Reviewed by Cynthia Shang.
2019-09-26 13:52:02 +02:00
P00 WARN: unknown user in backup manifest mapped to '[USER-2]'
P00 WARN: unknown group in backup manifest mapped to '[GROUP-2]'
P00 DETAIL: check '[TEST_PATH]/db-master/db/base' exists
P00 DETAIL: check '[TEST_PATH]/db-master/db/pg_config' exists
P00 DETAIL: check '[TEST_PATH]/db-master/db/pg_stat' exists
P00 DETAIL: remove 'global/pg_control' so cluster will not start if restore does not complete
P00 INFO: remove invalid files/links/paths from '[TEST_PATH]/db-master/db/base'
P00 DETAIL: update ownership for '[TEST_PATH]/db-master/db/base'
P00 DETAIL: remove special file '[TEST_PATH]/db-master/db/base/apipe'
P00 DETAIL: remove invalid file '[TEST_PATH]/db-master/db/base/backup_label.old'
P00 DETAIL: update mode for '[TEST_PATH]/db-master/db/base/base' to 0700
P00 DETAIL: update ownership for '[TEST_PATH]/db-master/db/base/base/1/PG_VERSION'
P00 DETAIL: update mode for '[TEST_PATH]/db-master/db/base/base/1/PG_VERSION' to 0660
P00 DETAIL: update ownership for '[TEST_PATH]/db-master/db/base/base/16384/PG_VERSION'
P00 DETAIL: remove invalid file '[TEST_PATH]/db-master/db/base/base/32768/44000'
P00 DETAIL: remove invalid file '[TEST_PATH]/db-master/db/base/base/32768/t333_44000'
P00 DETAIL: remove invalid path '[TEST_PATH]/db-master/db/base/base/pgsql_tmp'
P00 DETAIL: remove invalid path '[TEST_PATH]/db-master/db/base/deleteme'
P00 DETAIL: remove invalid file '[TEST_PATH]/db-master/db/base/global/pg_internal.init'
P00 DETAIL: remove invalid file '[TEST_PATH]/db-master/db/base/pg_dynshmem/anything.tmp'
P00 DETAIL: remove invalid file '[TEST_PATH]/db-master/db/base/pg_log/logfile'
P00 DETAIL: remove invalid path '[TEST_PATH]/db-master/db/base/pg_log2'
P00 DETAIL: remove invalid file '[TEST_PATH]/db-master/db/base/pg_notify/anything.tmp'
P00 DETAIL: remove invalid file '[TEST_PATH]/db-master/db/base/pg_replslot/anything.tmp'
P00 DETAIL: remove invalid file '[TEST_PATH]/db-master/db/base/pg_serial/anything.tmp'
P00 DETAIL: remove invalid file '[TEST_PATH]/db-master/db/base/pg_snapshots/anything.tmp'
P00 DETAIL: remove invalid file '[TEST_PATH]/db-master/db/base/pg_stat_tmp/anything.tmp'
P00 DETAIL: remove invalid file '[TEST_PATH]/db-master/db/base/pg_subtrans/anything.tmp'
P00 DETAIL: remove invalid link '[TEST_PATH]/db-master/db/base/postgresql.auto.conf'
P00 DETAIL: remove invalid file '[TEST_PATH]/db-master/db/base/postgresql.auto.conf.tmp'
P00 DETAIL: remove invalid file '[TEST_PATH]/db-master/db/base/postmaster.opts'
P00 DETAIL: remove invalid file '[TEST_PATH]/db-master/db/base/recovery.conf'
P00 DETAIL: remove invalid file '[TEST_PATH]/db-master/db/base/recovery.done'
P00 INFO: remove invalid files/links/paths from '[TEST_PATH]/db-master/db/pg_stat'
P00 DETAIL: create path '[TEST_PATH]/db-master/db/base/pg_clog'
Exclude temporary and unlogged relation (table/index) files from backup. Implemented using the same logic as the patches adding this feature to PostgreSQL, 8694cc96 and 920a5e50. Temporary relation exclusion is enabled in PostgreSQL ≥ 9.0. Unlogged relation exclusion is enabled in PostgreSQL ≥ 9.1, where the feature was introduced. Contributed by Cynthia Shang.
2018-07-31 00:53:34 +02:00
P01 DETAIL: restore file [TEST_PATH]/db-master/db/base/base/32768/33001 - exists and matches backup (64KB, 33%) checksum 6bf316f11d28c28914ea9be92c00de9bea6d9a6b
Fix incorrect offline upper bound for ignoring page checksum errors. For offline backups the upper bound was being set to 0x0000FFFF0000FFFF rather than UINT64_MAX. This meant that page checksum errors might be ignored for databases with a lot of past WAL in offline mode. Online mode is not affected since the upper bound is retrieved from pg_start_backup().
2019-07-11 15:13:56 +02:00
P01 DETAIL: restore file [TEST_PATH]/db-master/db/base/base/32768/44000_init - exists and matches backup (32KB, 49%) checksum 7a16d165e4775f7c92e8cdf60c0af57313f0bf90
P01 DETAIL: restore file [TEST_PATH]/db-master/db/base/base/32768/33000.32767 - exists and matches backup (32KB, 66%) checksum 6e99b589e550e68e934fd235ccba59fe5b592a9e
P01 DETAIL: restore file [TEST_PATH]/db-master/db/base/base/32768/33000 - exists and matches backup (32KB, 83%) checksum 7a16d165e4775f7c92e8cdf60c0af57313f0bf90
Exclude temporary and unlogged relation (table/index) files from backup. Implemented using the same logic as the patches adding this feature to PostgreSQL, 8694cc96 and 920a5e50. Temporary relation exclusion is enabled in PostgreSQL ≥ 9.0. Unlogged relation exclusion is enabled in PostgreSQL ≥ 9.1, where the feature was introduced. Contributed by Cynthia Shang.
2018-07-31 00:53:34 +02:00
P01 INFO: restore file [TEST_PATH]/db-master/db/base/base/16384/17000 (16KB, 91%) checksum e0101dd8ffb910c9c202ca35b5f828bcb9697bed
Support configurable WAL segment size. PostgreSQL 11 introduces configurable WAL segment sizes, from 1MB to 1GB. There are two areas that needed to be updated to support this: building the archive-get queue and checking that WAL has been archived after a backup. Both operations require the WAL segment size to properly build a list. Checking the archive after a backup is still implemented in Perl and has an active database connection, so just get the WAL segment size from the database. The archive-get command does not have a connection to the database, so get the WAL segment size from pg_control instead. This requires a deeper inspection of pg_control than has been done in the past, so it seemed best to copy the relevant data structures from each version of PostgreSQL and build a generic interface layer to address them. While this approach is a bit verbose, it has the advantage of being relatively simple, and can easily be updated for new versions of PostgreSQL. Since the integration tests generate pg_control files for testing, teach Perl how to generate files with the correct offsets for both 32-bit and 64-bit architectures.
2018-09-25 11:24:42 +02:00
P01 INFO: restore file [TEST_PATH]/db-master/db/base/global/pg_control.pgbackrest.tmp (8KB, 95%) checksum 4c77c900f7af0d9ab13fa9982051a42e0b637f6c
Add more realistic data files to synthetic backup and restore tests.
2016-12-10 16:13:55 +02:00
P01 DETAIL: restore file [TEST_PATH]/db-master/db/base/base/1/12000 - exists and matches backup (8KB, 99%) checksum 22c98d248ff548311eda88559e4a8405ed77c003
For simplicity, the pg_control file is now copied with the rest of the files instead of by itself of at the end of the process. The backup command does not require this behavior and the restore copies to a temporary file which is renamed at the end of the restore.
2016-11-30 21:23:05 +02:00
P01 DETAIL: restore file [TEST_PATH]/db-master/db/base/postgresql.conf - exists and matches backup (21B, 99%) checksum 6721d92c9fcdf4248acff1f9a1377127d9064807
Fix failure in manifest build when two or more files in PGDATA are linked to the same directory. Reported by Vitaliy Kukharik.
2018-05-02 18:19:54 +02:00
P01 DETAIL: restore file [TEST_PATH]/db-master/db/base/pg_hba.conf - exists and matches backup (9B, 99%) checksum dd4cea0cae348309f9de28ad4ded8ee2cc2e6d5b
Add checksum delta for incremental backups. Use checksums rather than timestamps to determine if files have changed. This is useful in cases where the timestamps may not be trustworthy, e.g. when performing an incremental after failing over to a standby. If checksum delta is enabled then checksums will be used for verification of resumed backups, even if they are full. Resumes have always used checksums to verify the files in the repository, enabling delta performs checksums on the database files as well. Note that the user must manually enable this feature in cases were it would be useful or just keep in enabled all the time. A future commit will address automatically enabling the feature in cases where it seems likely to be useful. Contributed by Cynthia Shang.
2018-09-19 17:12:45 +02:00
P01 DETAIL: restore file [TEST_PATH]/db-master/db/base/changecontent.txt - exists and matches backup (7B, 99%) checksum 238a131a3e8eb98d1fc5b27d882ca40b7618fd2a
Add more realistic data files to synthetic backup and restore tests.
2016-12-10 16:13:55 +02:00
P01 DETAIL: restore file [TEST_PATH]/db-master/db/base/pg_stat/global.stat - exists and matches backup (5B, 99%) checksum e350d5ce0153f3e22d5db21cf2a4eff00f3ee877
Add checksum delta for incremental backups. Use checksums rather than timestamps to determine if files have changed. This is useful in cases where the timestamps may not be trustworthy, e.g. when performing an incremental after failing over to a standby. If checksum delta is enabled then checksums will be used for verification of resumed backups, even if they are full. Resumes have always used checksums to verify the files in the repository, enabling delta performs checksums on the database files as well. Note that the user must manually enable this feature in cases were it would be useful or just keep in enabled all the time. A future commit will address automatically enabling the feature in cases where it seems likely to be useful. Contributed by Cynthia Shang.
2018-09-19 17:12:45 +02:00
P01 DETAIL: restore file [TEST_PATH]/db-master/db/base/changetime.txt - exists and matches backup (4B, 99%) checksum 88087292ed82e26f3eb824d0bffc05ccf7a30f8d
For simplicity, the pg_control file is now copied with the rest of the files instead of by itself of at the end of the process. The backup command does not require this behavior and the restore copies to a temporary file which is renamed at the end of the restore.
2016-11-30 21:23:05 +02:00
P01 DETAIL: restore file [TEST_PATH]/db-master/db/base/base/32768/PG_VERSION - exists and matches backup (3B, 99%) checksum 184473f470864e067ee3a22e64b47b0a1c356f29
P01 DETAIL: restore file [TEST_PATH]/db-master/db/base/base/16384/PG_VERSION - exists and matches backup (3B, 99%) checksum 184473f470864e067ee3a22e64b47b0a1c356f29
P01 DETAIL: restore file [TEST_PATH]/db-master/db/base/base/1/PG_VERSION - exists and matches backup (3B, 99%) checksum 184473f470864e067ee3a22e64b47b0a1c356f29
P01 DETAIL: restore file [TEST_PATH]/db-master/db/base/PG_VERSION - exists and matches backup (3B, 100%) checksum 184473f470864e067ee3a22e64b47b0a1c356f29
Allow zero-size files in backup manifest to reference a prior manifest regardless of timestamp delta. Contributed by Cynthia Shang.
2018-08-24 22:50:33 +02:00
P01 DETAIL: restore file [TEST_PATH]/db-master/db/base/zero_from_start - exists and is zero size (0B, 100%)
Use minio for integration testing. ScalityS3 has not received any maintenance in years and is slow to start which is bad for testing. Replace it with minio which starts quickly and ships as a single executable or a tiny container. Minio has stricter limits on allowable characters but should still provide enough coverage to show that our encoding is working correctly. This commit also includes the upgrade to openssl 1.1.1 in the Ubuntu 18.04 container.
2019-07-03 04:20:35 +02:00
P01 DETAIL: restore file [TEST_PATH]/db-master/db/base/special-!_.*'()&!@;:+,? - exists and is zero size (0B, 100%)
For simplicity, the pg_control file is now copied with the rest of the files instead of by itself of at the end of the process. The backup command does not require this behavior and the restore copies to a temporary file which is renamed at the end of the restore.
2016-11-30 21:23:05 +02:00
P00 INFO: write [TEST_PATH]/db-master/db/base/recovery.conf
The restore command is implemented entirely in C. For the most part this is a direct migration of the Perl code into C. There is one important behavioral change with regard to how file permissions are handled. The Perl code tried to set ownership as it was in the manifest even when running as an unprivileged user. This usually just led to errors and frustration. The C code works like this: If a restore is run as a non-root user (the typical scenario) then all files restored will belong to the user/group executing pgBackRest. If existing files are not owned by the executing user/group then an error will result if the ownership cannot be updated to the executing user/group. In that case the file ownership will need to be updated by a privileged user before the restore can be retried. If a restore is run as the root user then pgBackRest will attempt to recreate the ownership recorded in the manifest when the backup was made. Only user/group names are stored in the manifest so the same names must exist on the restore host for this to work. If the user/group name cannot be found locally then the user/group of the PostgreSQL data directory will be used and finally root if the data directory user/group cannot be mapped to a name. Reviewed by Cynthia Shang.
2019-09-26 13:52:02 +02:00
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/pg_config'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base/base'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base/base/1'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base/base/16384'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base/base/32768'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base/pg_clog'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base/pg_dynshmem'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base/pg_log'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base/pg_notify'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base/pg_replslot'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base/pg_serial'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base/pg_snapshots'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base/pg_stat'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base/pg_stat_tmp'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base/pg_subtrans'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base/pg_tblspc'
Storage and IO layer refactor: Refactor storage layer to allow for new repository filesystems using drivers. (Reviewed by Cynthia Shang.) Refactor IO layer to allow for new compression formats, checksum types, and other capabilities using filters. (Reviewed by Cynthia Shang.)
2017-06-09 23:51:41 +02:00
P00 INFO: restore global/pg_control (performed last to ensure aborted restores cannot be started)
The restore command is implemented entirely in C. For the most part this is a direct migration of the Perl code into C. There is one important behavioral change with regard to how file permissions are handled. The Perl code tried to set ownership as it was in the manifest even when running as an unprivileged user. This usually just led to errors and frustration. The C code works like this: If a restore is run as a non-root user (the typical scenario) then all files restored will belong to the user/group executing pgBackRest. If existing files are not owned by the executing user/group then an error will result if the ownership cannot be updated to the executing user/group. In that case the file ownership will need to be updated by a privileged user before the restore can be retried. If a restore is run as the root user then pgBackRest will attempt to recreate the ownership recorded in the manifest when the backup was made. Only user/group names are stored in the manifest so the same names must exist on the restore host for this to work. If the user/group name cannot be found locally then the user/group of the PostgreSQL data directory will be used and finally root if the data directory user/group cannot be mapped to a name. Reviewed by Cynthia Shang.
2019-09-26 13:52:02 +02:00
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base/global'
Added a predictable log entry to signal that a command has completed successfully. For example a backup ends successfully with: INFO: backup command end: completed successfully.
2016-12-04 21:44:53 +02:00
P00 INFO: restore command end: completed successfully
Log testing can now be enabled for certain deterministic tests. This works by comparing the generated logs against a previous copy. Currently only enabled for the backup/synthetic tests.
2015-05-05 19:08:48 +02:00
Rename test paths for clarity. This was worked out as part of the test suite refactor [c8f806a] but not committed with it because of the large number of expect logs changes involved. Keeping them separate made it easier to audit the changes in the refactor.
2016-06-24 16:54:31 +02:00
+ supplemental file: [TEST_PATH]/db-master/db/base/recovery.conf
----------------------------------------------------------------
PostgreSQL 12 support. Recovery settings are now written into postgresql.auto.conf instead of recovery.conf. Existing recovery_target* settings will be commented out to help avoid conflicts. A comment is added before recovery settings to identify them as written by pgBackRest since it is unclear how, in general, old settings will be removed. recovery.signal and standby.signal are automatically created based on the recovery settings.
2019-10-01 19:20:43 +02:00
# Recovery settings generated by pgBackRest restore on [TIMESTAMP]
Rename test paths for clarity. This was worked out as part of the test suite refactor [c8f806a] but not committed with it because of the large number of expect logs changes involved. Keeping them separate made it easier to audit the changes in the refactor.
2016-06-24 16:54:31 +02:00
restore_command = '[BACKREST-BIN] --config=[TEST_PATH]/db-master/pgbackrest.conf --stanza=db archive-get %f "%p"'
Added vagrant configuration for Ubuntu 12.04. Changes to regression tests to work better with VMs.
2015-07-07 03:00:48 +02:00
Fixed an issue with invalid backup groups being set correctly on restore. If the backup cannot map a group to a name it stores the group in the manifest as false then uses either the owner of $PGDATA to set the group during restore or failing that the group of the current user. This logic was not working correctly because the selected group was overwriting the user on restore leaving the group undefined and the user incorrectly set to the group. (Reported by Jeff McCormick.)
2017-11-20 19:46:35 +02:00
restore delta, backup '[BACKUP-FULL-2]' - fix permissions (db-master host)
> [CONTAINER-EXEC] db-master [BACKREST-BIN] --config=[TEST_PATH]/db-master/pgbackrest.conf --delta --set=[BACKUP-FULL-2] --link-all --log-level-console=detail --stanza=db restore
------------------------------------------------------------------------------------------------------------------------------------
Remove compress/compress-level options from commands where unused. These commands (e.g. restore, archive-get) never used the compress options but allowed them to be passed on the command line. Now they will error when these options are passed on the command line. If these errors occur then remove the unused options.
2020-02-27 19:25:32 +02:00
P00 INFO: restore command begin [BACKREST-VERSION]: --buffer-size=[BUFFER-SIZE] --config=[TEST_PATH]/db-master/pgbackrest.conf --delta --link-all --lock-path=[TEST_PATH]/db-master/lock --log-level-console=detail --log-level-file=[LOG-LEVEL-FILE] --log-level-stderr=off --log-path=[TEST_PATH]/db-master/log[] --no-log-timestamp --pg1-path=[TEST_PATH]/db-master/db/base --protocol-timeout=60 --repo1-path=[TEST_PATH]/db-master/repo --set=[BACKUP-FULL-2] --stanza=db
Fixed an issue with invalid backup groups being set correctly on restore. If the backup cannot map a group to a name it stores the group in the manifest as false then uses either the owner of $PGDATA to set the group during restore or failing that the group of the current user. This logic was not working correctly because the selected group was overwriting the user on restore leaving the group undefined and the user incorrectly set to the group. (Reported by Jeff McCormick.)
2017-11-20 19:46:35 +02:00
P00 INFO: restore backup set [BACKUP-FULL-2]
The restore command is implemented entirely in C. For the most part this is a direct migration of the Perl code into C. There is one important behavioral change with regard to how file permissions are handled. The Perl code tried to set ownership as it was in the manifest even when running as an unprivileged user. This usually just led to errors and frustration. The C code works like this: If a restore is run as a non-root user (the typical scenario) then all files restored will belong to the user/group executing pgBackRest. If existing files are not owned by the executing user/group then an error will result if the ownership cannot be updated to the executing user/group. In that case the file ownership will need to be updated by a privileged user before the restore can be retried. If a restore is run as the root user then pgBackRest will attempt to recreate the ownership recorded in the manifest when the backup was made. Only user/group names are stored in the manifest so the same names must exist on the restore host for this to work. If the user/group name cannot be found locally then the user/group of the PostgreSQL data directory will be used and finally root if the data directory user/group cannot be mapped to a name. Reviewed by Cynthia Shang.
2019-09-26 13:52:02 +02:00
P00 WARN: unknown user in backup manifest mapped to '[USER-1]'
P00 WARN: unknown group in backup manifest mapped to '[GROUP-1]'
P00 DETAIL: check '[TEST_PATH]/db-master/db/base' exists
P00 DETAIL: check '[TEST_PATH]/db-master/db/pg_config' exists
P00 DETAIL: check '[TEST_PATH]/db-master/db/pg_stat' exists
P00 DETAIL: remove 'global/pg_control' so cluster will not start if restore does not complete
P00 INFO: remove invalid files/links/paths from '[TEST_PATH]/db-master/db/base'
P00 DETAIL: update ownership for '[TEST_PATH]/db-master/db/base/base/1/PG_VERSION'
P00 DETAIL: update ownership for '[TEST_PATH]/db-master/db/base/base/16384/PG_VERSION'
P00 DETAIL: remove invalid file '[TEST_PATH]/db-master/db/base/recovery.conf'
P00 INFO: remove invalid files/links/paths from '[TEST_PATH]/db-master/db/pg_stat'
Exclude temporary and unlogged relation (table/index) files from backup. Implemented using the same logic as the patches adding this feature to PostgreSQL, 8694cc96 and 920a5e50. Temporary relation exclusion is enabled in PostgreSQL ≥ 9.0. Unlogged relation exclusion is enabled in PostgreSQL ≥ 9.1, where the feature was introduced. Contributed by Cynthia Shang.
2018-07-31 00:53:34 +02:00
P01 DETAIL: restore file [TEST_PATH]/db-master/db/base/base/32768/33001 - exists and matches backup (64KB, 33%) checksum 6bf316f11d28c28914ea9be92c00de9bea6d9a6b
Fix incorrect offline upper bound for ignoring page checksum errors. For offline backups the upper bound was being set to 0x0000FFFF0000FFFF rather than UINT64_MAX. This meant that page checksum errors might be ignored for databases with a lot of past WAL in offline mode. Online mode is not affected since the upper bound is retrieved from pg_start_backup().
2019-07-11 15:13:56 +02:00
P01 DETAIL: restore file [TEST_PATH]/db-master/db/base/base/32768/44000_init - exists and matches backup (32KB, 49%) checksum 7a16d165e4775f7c92e8cdf60c0af57313f0bf90
P01 DETAIL: restore file [TEST_PATH]/db-master/db/base/base/32768/33000.32767 - exists and matches backup (32KB, 66%) checksum 6e99b589e550e68e934fd235ccba59fe5b592a9e
P01 DETAIL: restore file [TEST_PATH]/db-master/db/base/base/32768/33000 - exists and matches backup (32KB, 83%) checksum 7a16d165e4775f7c92e8cdf60c0af57313f0bf90
Exclude temporary and unlogged relation (table/index) files from backup. Implemented using the same logic as the patches adding this feature to PostgreSQL, 8694cc96 and 920a5e50. Temporary relation exclusion is enabled in PostgreSQL ≥ 9.0. Unlogged relation exclusion is enabled in PostgreSQL ≥ 9.1, where the feature was introduced. Contributed by Cynthia Shang.
2018-07-31 00:53:34 +02:00
P01 DETAIL: restore file [TEST_PATH]/db-master/db/base/base/16384/17000 - exists and matches backup (16KB, 91%) checksum e0101dd8ffb910c9c202ca35b5f828bcb9697bed
Support configurable WAL segment size. PostgreSQL 11 introduces configurable WAL segment sizes, from 1MB to 1GB. There are two areas that needed to be updated to support this: building the archive-get queue and checking that WAL has been archived after a backup. Both operations require the WAL segment size to properly build a list. Checking the archive after a backup is still implemented in Perl and has an active database connection, so just get the WAL segment size from the database. The archive-get command does not have a connection to the database, so get the WAL segment size from pg_control instead. This requires a deeper inspection of pg_control than has been done in the past, so it seemed best to copy the relevant data structures from each version of PostgreSQL and build a generic interface layer to address them. While this approach is a bit verbose, it has the advantage of being relatively simple, and can easily be updated for new versions of PostgreSQL. Since the integration tests generate pg_control files for testing, teach Perl how to generate files with the correct offsets for both 32-bit and 64-bit architectures.
2018-09-25 11:24:42 +02:00
P01 INFO: restore file [TEST_PATH]/db-master/db/base/global/pg_control.pgbackrest.tmp (8KB, 95%) checksum 4c77c900f7af0d9ab13fa9982051a42e0b637f6c
Fixed an issue with invalid backup groups being set correctly on restore. If the backup cannot map a group to a name it stores the group in the manifest as false then uses either the owner of $PGDATA to set the group during restore or failing that the group of the current user. This logic was not working correctly because the selected group was overwriting the user on restore leaving the group undefined and the user incorrectly set to the group. (Reported by Jeff McCormick.)
2017-11-20 19:46:35 +02:00
P01 DETAIL: restore file [TEST_PATH]/db-master/db/base/base/1/12000 - exists and matches backup (8KB, 99%) checksum 22c98d248ff548311eda88559e4a8405ed77c003
P01 DETAIL: restore file [TEST_PATH]/db-master/db/base/postgresql.conf - exists and matches backup (21B, 99%) checksum 6721d92c9fcdf4248acff1f9a1377127d9064807
Fix failure in manifest build when two or more files in PGDATA are linked to the same directory. Reported by Vitaliy Kukharik.
2018-05-02 18:19:54 +02:00
P01 DETAIL: restore file [TEST_PATH]/db-master/db/base/pg_hba.conf - exists and matches backup (9B, 99%) checksum dd4cea0cae348309f9de28ad4ded8ee2cc2e6d5b
Add checksum delta for incremental backups. Use checksums rather than timestamps to determine if files have changed. This is useful in cases where the timestamps may not be trustworthy, e.g. when performing an incremental after failing over to a standby. If checksum delta is enabled then checksums will be used for verification of resumed backups, even if they are full. Resumes have always used checksums to verify the files in the repository, enabling delta performs checksums on the database files as well. Note that the user must manually enable this feature in cases were it would be useful or just keep in enabled all the time. A future commit will address automatically enabling the feature in cases where it seems likely to be useful. Contributed by Cynthia Shang.
2018-09-19 17:12:45 +02:00
P01 DETAIL: restore file [TEST_PATH]/db-master/db/base/changecontent.txt - exists and matches backup (7B, 99%) checksum 238a131a3e8eb98d1fc5b27d882ca40b7618fd2a
Fixed an issue with invalid backup groups being set correctly on restore. If the backup cannot map a group to a name it stores the group in the manifest as false then uses either the owner of $PGDATA to set the group during restore or failing that the group of the current user. This logic was not working correctly because the selected group was overwriting the user on restore leaving the group undefined and the user incorrectly set to the group. (Reported by Jeff McCormick.)
2017-11-20 19:46:35 +02:00
P01 DETAIL: restore file [TEST_PATH]/db-master/db/base/pg_stat/global.stat - exists and matches backup (5B, 99%) checksum e350d5ce0153f3e22d5db21cf2a4eff00f3ee877
Add checksum delta for incremental backups. Use checksums rather than timestamps to determine if files have changed. This is useful in cases where the timestamps may not be trustworthy, e.g. when performing an incremental after failing over to a standby. If checksum delta is enabled then checksums will be used for verification of resumed backups, even if they are full. Resumes have always used checksums to verify the files in the repository, enabling delta performs checksums on the database files as well. Note that the user must manually enable this feature in cases were it would be useful or just keep in enabled all the time. A future commit will address automatically enabling the feature in cases where it seems likely to be useful. Contributed by Cynthia Shang.
2018-09-19 17:12:45 +02:00
P01 DETAIL: restore file [TEST_PATH]/db-master/db/base/changetime.txt - exists and matches backup (4B, 99%) checksum 88087292ed82e26f3eb824d0bffc05ccf7a30f8d
Fixed an issue with invalid backup groups being set correctly on restore. If the backup cannot map a group to a name it stores the group in the manifest as false then uses either the owner of $PGDATA to set the group during restore or failing that the group of the current user. This logic was not working correctly because the selected group was overwriting the user on restore leaving the group undefined and the user incorrectly set to the group. (Reported by Jeff McCormick.)
2017-11-20 19:46:35 +02:00
P01 DETAIL: restore file [TEST_PATH]/db-master/db/base/base/32768/PG_VERSION - exists and matches backup (3B, 99%) checksum 184473f470864e067ee3a22e64b47b0a1c356f29
P01 DETAIL: restore file [TEST_PATH]/db-master/db/base/base/16384/PG_VERSION - exists and matches backup (3B, 99%) checksum 184473f470864e067ee3a22e64b47b0a1c356f29
P01 DETAIL: restore file [TEST_PATH]/db-master/db/base/base/1/PG_VERSION - exists and matches backup (3B, 99%) checksum 184473f470864e067ee3a22e64b47b0a1c356f29
P01 DETAIL: restore file [TEST_PATH]/db-master/db/base/PG_VERSION - exists and matches backup (3B, 100%) checksum 184473f470864e067ee3a22e64b47b0a1c356f29
Allow zero-size files in backup manifest to reference a prior manifest regardless of timestamp delta. Contributed by Cynthia Shang.
2018-08-24 22:50:33 +02:00
P01 DETAIL: restore file [TEST_PATH]/db-master/db/base/zero_from_start - exists and is zero size (0B, 100%)
Use minio for integration testing. ScalityS3 has not received any maintenance in years and is slow to start which is bad for testing. Replace it with minio which starts quickly and ships as a single executable or a tiny container. Minio has stricter limits on allowable characters but should still provide enough coverage to show that our encoding is working correctly. This commit also includes the upgrade to openssl 1.1.1 in the Ubuntu 18.04 container.
2019-07-03 04:20:35 +02:00
P01 DETAIL: restore file [TEST_PATH]/db-master/db/base/special-!_.*'()&!@;:+,? - exists and is zero size (0B, 100%)
Fixed an issue with invalid backup groups being set correctly on restore. If the backup cannot map a group to a name it stores the group in the manifest as false then uses either the owner of $PGDATA to set the group during restore or failing that the group of the current user. This logic was not working correctly because the selected group was overwriting the user on restore leaving the group undefined and the user incorrectly set to the group. (Reported by Jeff McCormick.)
2017-11-20 19:46:35 +02:00
P00 INFO: write [TEST_PATH]/db-master/db/base/recovery.conf
The restore command is implemented entirely in C. For the most part this is a direct migration of the Perl code into C. There is one important behavioral change with regard to how file permissions are handled. The Perl code tried to set ownership as it was in the manifest even when running as an unprivileged user. This usually just led to errors and frustration. The C code works like this: If a restore is run as a non-root user (the typical scenario) then all files restored will belong to the user/group executing pgBackRest. If existing files are not owned by the executing user/group then an error will result if the ownership cannot be updated to the executing user/group. In that case the file ownership will need to be updated by a privileged user before the restore can be retried. If a restore is run as the root user then pgBackRest will attempt to recreate the ownership recorded in the manifest when the backup was made. Only user/group names are stored in the manifest so the same names must exist on the restore host for this to work. If the user/group name cannot be found locally then the user/group of the PostgreSQL data directory will be used and finally root if the data directory user/group cannot be mapped to a name. Reviewed by Cynthia Shang.
2019-09-26 13:52:02 +02:00
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/pg_config'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base/base'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base/base/1'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base/base/16384'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base/base/32768'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base/pg_clog'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base/pg_dynshmem'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base/pg_log'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base/pg_notify'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base/pg_replslot'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base/pg_serial'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base/pg_snapshots'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base/pg_stat'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base/pg_stat_tmp'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base/pg_subtrans'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base/pg_tblspc'
Fixed an issue with invalid backup groups being set correctly on restore. If the backup cannot map a group to a name it stores the group in the manifest as false then uses either the owner of $PGDATA to set the group during restore or failing that the group of the current user. This logic was not working correctly because the selected group was overwriting the user on restore leaving the group undefined and the user incorrectly set to the group. (Reported by Jeff McCormick.)
2017-11-20 19:46:35 +02:00
P00 INFO: restore global/pg_control (performed last to ensure aborted restores cannot be started)
The restore command is implemented entirely in C. For the most part this is a direct migration of the Perl code into C. There is one important behavioral change with regard to how file permissions are handled. The Perl code tried to set ownership as it was in the manifest even when running as an unprivileged user. This usually just led to errors and frustration. The C code works like this: If a restore is run as a non-root user (the typical scenario) then all files restored will belong to the user/group executing pgBackRest. If existing files are not owned by the executing user/group then an error will result if the ownership cannot be updated to the executing user/group. In that case the file ownership will need to be updated by a privileged user before the restore can be retried. If a restore is run as the root user then pgBackRest will attempt to recreate the ownership recorded in the manifest when the backup was made. Only user/group names are stored in the manifest so the same names must exist on the restore host for this to work. If the user/group name cannot be found locally then the user/group of the PostgreSQL data directory will be used and finally root if the data directory user/group cannot be mapped to a name. Reviewed by Cynthia Shang.
2019-09-26 13:52:02 +02:00
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base/global'
Fixed an issue with invalid backup groups being set correctly on restore. If the backup cannot map a group to a name it stores the group in the manifest as false then uses either the owner of $PGDATA to set the group during restore or failing that the group of the current user. This logic was not working correctly because the selected group was overwriting the user on restore leaving the group undefined and the user incorrectly set to the group. (Reported by Jeff McCormick.)
2017-11-20 19:46:35 +02:00
P00 INFO: restore command end: completed successfully
+ supplemental file: [TEST_PATH]/db-master/db/base/recovery.conf
----------------------------------------------------------------
PostgreSQL 12 support. Recovery settings are now written into postgresql.auto.conf instead of recovery.conf. Existing recovery_target* settings will be commented out to help avoid conflicts. A comment is added before recovery settings to identify them as written by pgBackRest since it is unclear how, in general, old settings will be removed. recovery.signal and standby.signal are automatically created based on the recovery settings.
2019-10-01 19:20:43 +02:00
# Recovery settings generated by pgBackRest restore on [TIMESTAMP]
The restore command is implemented entirely in C. For the most part this is a direct migration of the Perl code into C. There is one important behavioral change with regard to how file permissions are handled. The Perl code tried to set ownership as it was in the manifest even when running as an unprivileged user. This usually just led to errors and frustration. The C code works like this: If a restore is run as a non-root user (the typical scenario) then all files restored will belong to the user/group executing pgBackRest. If existing files are not owned by the executing user/group then an error will result if the ownership cannot be updated to the executing user/group. In that case the file ownership will need to be updated by a privileged user before the restore can be retried. If a restore is run as the root user then pgBackRest will attempt to recreate the ownership recorded in the manifest when the backup was made. Only user/group names are stored in the manifest so the same names must exist on the restore host for this to work. If the user/group name cannot be found locally then the user/group of the PostgreSQL data directory will be used and finally root if the data directory user/group cannot be mapped to a name. Reviewed by Cynthia Shang.
2019-09-26 13:52:02 +02:00
restore_command = '[BACKREST-BIN] --config=[TEST_PATH]/db-master/pgbackrest.conf --stanza=db archive-get %f "%p"'
Fixed an issue with invalid backup groups being set correctly on restore. If the backup cannot map a group to a name it stores the group in the manifest as false then uses either the owner of $PGDATA to set the group during restore or failing that the group of the current user. This logic was not working correctly because the selected group was overwriting the user on restore leaving the group undefined and the user incorrectly set to the group. (Reported by Jeff McCormick.)
2017-11-20 19:46:35 +02:00
User/group permissions improvements. Improved handling of users/groups captured during backup that do not exist on the restore host. Also explicitly handle the case where user/group is not mapped to a name.
2016-06-27 03:01:20 +02:00
restore delta, backup '[BACKUP-FULL-2]' - fix broken symlink (db-master host)
Reduce log level for all expect tests to detail. The C code is designed to be efficient rather than deterministic at the debug log level. As we move more testing from integration to unit tests it makes less sense to try and maintain the expect logs at this log level. Most of the expect logs have already been moved to detail level but mock/all still had tests at debug level. Change the logging defaults in the config file and remove as many references to log-level-console as possible.
2019-05-23 00:23:44 +02:00
> [CONTAINER-EXEC] db-master [BACKREST-BIN] --config=[TEST_PATH]/db-master/pgbackrest.conf --delta --set=[BACKUP-FULL-2] --link-all --stanza=db restore
User/group permissions improvements. Improved handling of users/groups captured during backup that do not exist on the restore host. Also explicitly handle the case where user/group is not mapped to a name.
2016-06-27 03:01:20 +02:00
------------------------------------------------------------------------------------------------------------------------------------
Remove compress/compress-level options from commands where unused. These commands (e.g. restore, archive-get) never used the compress options but allowed them to be passed on the command line. Now they will error when these options are passed on the command line. If these errors occur then remove the unused options.
2020-02-27 19:25:32 +02:00
P00 INFO: restore command begin [BACKREST-VERSION]: --buffer-size=[BUFFER-SIZE] --config=[TEST_PATH]/db-master/pgbackrest.conf --delta --link-all --lock-path=[TEST_PATH]/db-master/lock --log-level-console=detail --log-level-file=[LOG-LEVEL-FILE] --log-level-stderr=off --log-path=[TEST_PATH]/db-master/log[] --no-log-timestamp --pg1-path=[TEST_PATH]/db-master/db/base --protocol-timeout=60 --repo1-path=[TEST_PATH]/db-master/repo --set=[BACKUP-FULL-2] --stanza=db
Show Process ID in log output instead of filtering it out with the timestamp.
2016-09-14 23:30:57 +02:00
P00 INFO: restore backup set [BACKUP-FULL-2]
The restore command is implemented entirely in C. For the most part this is a direct migration of the Perl code into C. There is one important behavioral change with regard to how file permissions are handled. The Perl code tried to set ownership as it was in the manifest even when running as an unprivileged user. This usually just led to errors and frustration. The C code works like this: If a restore is run as a non-root user (the typical scenario) then all files restored will belong to the user/group executing pgBackRest. If existing files are not owned by the executing user/group then an error will result if the ownership cannot be updated to the executing user/group. In that case the file ownership will need to be updated by a privileged user before the restore can be retried. If a restore is run as the root user then pgBackRest will attempt to recreate the ownership recorded in the manifest when the backup was made. Only user/group names are stored in the manifest so the same names must exist on the restore host for this to work. If the user/group name cannot be found locally then the user/group of the PostgreSQL data directory will be used and finally root if the data directory user/group cannot be mapped to a name. Reviewed by Cynthia Shang.
2019-09-26 13:52:02 +02:00
P00 WARN: unknown user in backup manifest mapped to current user
P00 WARN: unknown group in backup manifest mapped to current group
P00 DETAIL: check '[TEST_PATH]/db-master/db/base' exists
P00 DETAIL: check '[TEST_PATH]/db-master/db/pg_config' exists
P00 DETAIL: check '[TEST_PATH]/db-master/db/pg_stat' exists
P00 DETAIL: remove 'global/pg_control' so cluster will not start if restore does not complete
P00 INFO: remove invalid files/links/paths from '[TEST_PATH]/db-master/db/base'
P00 DETAIL: remove link '[TEST_PATH]/db-master/db/base/pg_stat' because destination changed
P00 DETAIL: remove invalid file '[TEST_PATH]/db-master/db/base/recovery.conf'
P00 INFO: remove invalid files/links/paths from '[TEST_PATH]/db-master/db/pg_stat'
P00 DETAIL: create symlink '[TEST_PATH]/db-master/db/base/pg_stat' to '../pg_stat'
Exclude temporary and unlogged relation (table/index) files from backup. Implemented using the same logic as the patches adding this feature to PostgreSQL, 8694cc96 and 920a5e50. Temporary relation exclusion is enabled in PostgreSQL ≥ 9.0. Unlogged relation exclusion is enabled in PostgreSQL ≥ 9.1, where the feature was introduced. Contributed by Cynthia Shang.
2018-07-31 00:53:34 +02:00
P01 DETAIL: restore file [TEST_PATH]/db-master/db/base/base/32768/33001 - exists and matches backup (64KB, 33%) checksum 6bf316f11d28c28914ea9be92c00de9bea6d9a6b
Fix incorrect offline upper bound for ignoring page checksum errors. For offline backups the upper bound was being set to 0x0000FFFF0000FFFF rather than UINT64_MAX. This meant that page checksum errors might be ignored for databases with a lot of past WAL in offline mode. Online mode is not affected since the upper bound is retrieved from pg_start_backup().
2019-07-11 15:13:56 +02:00
P01 DETAIL: restore file [TEST_PATH]/db-master/db/base/base/32768/44000_init - exists and matches backup (32KB, 49%) checksum 7a16d165e4775f7c92e8cdf60c0af57313f0bf90
P01 DETAIL: restore file [TEST_PATH]/db-master/db/base/base/32768/33000.32767 - exists and matches backup (32KB, 66%) checksum 6e99b589e550e68e934fd235ccba59fe5b592a9e
P01 DETAIL: restore file [TEST_PATH]/db-master/db/base/base/32768/33000 - exists and matches backup (32KB, 83%) checksum 7a16d165e4775f7c92e8cdf60c0af57313f0bf90
Exclude temporary and unlogged relation (table/index) files from backup. Implemented using the same logic as the patches adding this feature to PostgreSQL, 8694cc96 and 920a5e50. Temporary relation exclusion is enabled in PostgreSQL ≥ 9.0. Unlogged relation exclusion is enabled in PostgreSQL ≥ 9.1, where the feature was introduced. Contributed by Cynthia Shang.
2018-07-31 00:53:34 +02:00
P01 DETAIL: restore file [TEST_PATH]/db-master/db/base/base/16384/17000 - exists and matches backup (16KB, 91%) checksum e0101dd8ffb910c9c202ca35b5f828bcb9697bed
Support configurable WAL segment size. PostgreSQL 11 introduces configurable WAL segment sizes, from 1MB to 1GB. There are two areas that needed to be updated to support this: building the archive-get queue and checking that WAL has been archived after a backup. Both operations require the WAL segment size to properly build a list. Checking the archive after a backup is still implemented in Perl and has an active database connection, so just get the WAL segment size from the database. The archive-get command does not have a connection to the database, so get the WAL segment size from pg_control instead. This requires a deeper inspection of pg_control than has been done in the past, so it seemed best to copy the relevant data structures from each version of PostgreSQL and build a generic interface layer to address them. While this approach is a bit verbose, it has the advantage of being relatively simple, and can easily be updated for new versions of PostgreSQL. Since the integration tests generate pg_control files for testing, teach Perl how to generate files with the correct offsets for both 32-bit and 64-bit architectures.
2018-09-25 11:24:42 +02:00
P01 INFO: restore file [TEST_PATH]/db-master/db/base/global/pg_control.pgbackrest.tmp (8KB, 95%) checksum 4c77c900f7af0d9ab13fa9982051a42e0b637f6c
Add more realistic data files to synthetic backup and restore tests.
2016-12-10 16:13:55 +02:00
P01 DETAIL: restore file [TEST_PATH]/db-master/db/base/base/1/12000 - exists and matches backup (8KB, 99%) checksum 22c98d248ff548311eda88559e4a8405ed77c003
For simplicity, the pg_control file is now copied with the rest of the files instead of by itself of at the end of the process. The backup command does not require this behavior and the restore copies to a temporary file which is renamed at the end of the restore.
2016-11-30 21:23:05 +02:00
P01 DETAIL: restore file [TEST_PATH]/db-master/db/base/postgresql.conf - exists and matches backup (21B, 99%) checksum 6721d92c9fcdf4248acff1f9a1377127d9064807
Fix failure in manifest build when two or more files in PGDATA are linked to the same directory. Reported by Vitaliy Kukharik.
2018-05-02 18:19:54 +02:00
P01 DETAIL: restore file [TEST_PATH]/db-master/db/base/pg_hba.conf - exists and matches backup (9B, 99%) checksum dd4cea0cae348309f9de28ad4ded8ee2cc2e6d5b
Add checksum delta for incremental backups. Use checksums rather than timestamps to determine if files have changed. This is useful in cases where the timestamps may not be trustworthy, e.g. when performing an incremental after failing over to a standby. If checksum delta is enabled then checksums will be used for verification of resumed backups, even if they are full. Resumes have always used checksums to verify the files in the repository, enabling delta performs checksums on the database files as well. Note that the user must manually enable this feature in cases were it would be useful or just keep in enabled all the time. A future commit will address automatically enabling the feature in cases where it seems likely to be useful. Contributed by Cynthia Shang.
2018-09-19 17:12:45 +02:00
P01 DETAIL: restore file [TEST_PATH]/db-master/db/base/changecontent.txt - exists and matches backup (7B, 99%) checksum 238a131a3e8eb98d1fc5b27d882ca40b7618fd2a
For simplicity, the pg_control file is now copied with the rest of the files instead of by itself of at the end of the process. The backup command does not require this behavior and the restore copies to a temporary file which is renamed at the end of the restore.
2016-11-30 21:23:05 +02:00
P01 DETAIL: restore file [TEST_PATH]/db-master/db/base/pg_stat/global.stat - exists and matches backup (5B, 99%) checksum e350d5ce0153f3e22d5db21cf2a4eff00f3ee877
Add checksum delta for incremental backups. Use checksums rather than timestamps to determine if files have changed. This is useful in cases where the timestamps may not be trustworthy, e.g. when performing an incremental after failing over to a standby. If checksum delta is enabled then checksums will be used for verification of resumed backups, even if they are full. Resumes have always used checksums to verify the files in the repository, enabling delta performs checksums on the database files as well. Note that the user must manually enable this feature in cases were it would be useful or just keep in enabled all the time. A future commit will address automatically enabling the feature in cases where it seems likely to be useful. Contributed by Cynthia Shang.
2018-09-19 17:12:45 +02:00
P01 DETAIL: restore file [TEST_PATH]/db-master/db/base/changetime.txt - exists and matches backup (4B, 99%) checksum 88087292ed82e26f3eb824d0bffc05ccf7a30f8d
For simplicity, the pg_control file is now copied with the rest of the files instead of by itself of at the end of the process. The backup command does not require this behavior and the restore copies to a temporary file which is renamed at the end of the restore.
2016-11-30 21:23:05 +02:00
P01 DETAIL: restore file [TEST_PATH]/db-master/db/base/base/32768/PG_VERSION - exists and matches backup (3B, 99%) checksum 184473f470864e067ee3a22e64b47b0a1c356f29
P01 DETAIL: restore file [TEST_PATH]/db-master/db/base/base/16384/PG_VERSION - exists and matches backup (3B, 99%) checksum 184473f470864e067ee3a22e64b47b0a1c356f29
P01 DETAIL: restore file [TEST_PATH]/db-master/db/base/base/1/PG_VERSION - exists and matches backup (3B, 99%) checksum 184473f470864e067ee3a22e64b47b0a1c356f29
P01 DETAIL: restore file [TEST_PATH]/db-master/db/base/PG_VERSION - exists and matches backup (3B, 100%) checksum 184473f470864e067ee3a22e64b47b0a1c356f29
Allow zero-size files in backup manifest to reference a prior manifest regardless of timestamp delta. Contributed by Cynthia Shang.
2018-08-24 22:50:33 +02:00
P01 DETAIL: restore file [TEST_PATH]/db-master/db/base/zero_from_start - exists and is zero size (0B, 100%)
Use minio for integration testing. ScalityS3 has not received any maintenance in years and is slow to start which is bad for testing. Replace it with minio which starts quickly and ships as a single executable or a tiny container. Minio has stricter limits on allowable characters but should still provide enough coverage to show that our encoding is working correctly. This commit also includes the upgrade to openssl 1.1.1 in the Ubuntu 18.04 container.
2019-07-03 04:20:35 +02:00
P01 DETAIL: restore file [TEST_PATH]/db-master/db/base/special-!_.*'()&!@;:+,? - exists and is zero size (0B, 100%)
For simplicity, the pg_control file is now copied with the rest of the files instead of by itself of at the end of the process. The backup command does not require this behavior and the restore copies to a temporary file which is renamed at the end of the restore.
2016-11-30 21:23:05 +02:00
P00 INFO: write [TEST_PATH]/db-master/db/base/recovery.conf
The restore command is implemented entirely in C. For the most part this is a direct migration of the Perl code into C. There is one important behavioral change with regard to how file permissions are handled. The Perl code tried to set ownership as it was in the manifest even when running as an unprivileged user. This usually just led to errors and frustration. The C code works like this: If a restore is run as a non-root user (the typical scenario) then all files restored will belong to the user/group executing pgBackRest. If existing files are not owned by the executing user/group then an error will result if the ownership cannot be updated to the executing user/group. In that case the file ownership will need to be updated by a privileged user before the restore can be retried. If a restore is run as the root user then pgBackRest will attempt to recreate the ownership recorded in the manifest when the backup was made. Only user/group names are stored in the manifest so the same names must exist on the restore host for this to work. If the user/group name cannot be found locally then the user/group of the PostgreSQL data directory will be used and finally root if the data directory user/group cannot be mapped to a name. Reviewed by Cynthia Shang.
2019-09-26 13:52:02 +02:00
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/pg_config'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base/base'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base/base/1'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base/base/16384'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base/base/32768'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base/pg_clog'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base/pg_dynshmem'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base/pg_log'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base/pg_notify'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base/pg_replslot'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base/pg_serial'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base/pg_snapshots'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base/pg_stat'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base/pg_stat_tmp'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base/pg_subtrans'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base/pg_tblspc'
Storage and IO layer refactor: Refactor storage layer to allow for new repository filesystems using drivers. (Reviewed by Cynthia Shang.) Refactor IO layer to allow for new compression formats, checksum types, and other capabilities using filters. (Reviewed by Cynthia Shang.)
2017-06-09 23:51:41 +02:00
P00 INFO: restore global/pg_control (performed last to ensure aborted restores cannot be started)
The restore command is implemented entirely in C. For the most part this is a direct migration of the Perl code into C. There is one important behavioral change with regard to how file permissions are handled. The Perl code tried to set ownership as it was in the manifest even when running as an unprivileged user. This usually just led to errors and frustration. The C code works like this: If a restore is run as a non-root user (the typical scenario) then all files restored will belong to the user/group executing pgBackRest. If existing files are not owned by the executing user/group then an error will result if the ownership cannot be updated to the executing user/group. In that case the file ownership will need to be updated by a privileged user before the restore can be retried. If a restore is run as the root user then pgBackRest will attempt to recreate the ownership recorded in the manifest when the backup was made. Only user/group names are stored in the manifest so the same names must exist on the restore host for this to work. If the user/group name cannot be found locally then the user/group of the PostgreSQL data directory will be used and finally root if the data directory user/group cannot be mapped to a name. Reviewed by Cynthia Shang.
2019-09-26 13:52:02 +02:00
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base/global'
Added a predictable log entry to signal that a command has completed successfully. For example a backup ends successfully with: INFO: backup command end: completed successfully.
2016-12-04 21:44:53 +02:00
P00 INFO: restore command end: completed successfully
User/group permissions improvements. Improved handling of users/groups captured during backup that do not exist on the restore host. Also explicitly handle the case where user/group is not mapped to a name.
2016-06-27 03:01:20 +02:00
+ supplemental file: [TEST_PATH]/db-master/db/base/recovery.conf
----------------------------------------------------------------
PostgreSQL 12 support. Recovery settings are now written into postgresql.auto.conf instead of recovery.conf. Existing recovery_target* settings will be commented out to help avoid conflicts. A comment is added before recovery settings to identify them as written by pgBackRest since it is unclear how, in general, old settings will be removed. recovery.signal and standby.signal are automatically created based on the recovery settings.
2019-10-01 19:20:43 +02:00
# Recovery settings generated by pgBackRest restore on [TIMESTAMP]
Reduce log level for all expect tests to detail. The C code is designed to be efficient rather than deterministic at the debug log level. As we move more testing from integration to unit tests it makes less sense to try and maintain the expect logs at this log level. Most of the expect logs have already been moved to detail level but mock/all still had tests at debug level. Change the logging defaults in the config file and remove as many references to log-level-console as possible.
2019-05-23 00:23:44 +02:00
restore_command = '[BACKREST-BIN] --config=[TEST_PATH]/db-master/pgbackrest.conf --stanza=db archive-get %f "%p"'
User/group permissions improvements. Improved handling of users/groups captured during backup that do not exist on the restore host. Also explicitly handle the case where user/group is not mapped to a name.
2016-06-27 03:01:20 +02:00
Remove integration tests that are now covered in the unit tests. Most of these tests are just checking that errors are thrown when required. These are well covered in various unit tests. The "cannot resume" tests are also well covered in the backup unit tests. Finally, config warnings are well covered in the config unit tests. There is more to be done here, but this accounts for the low-hanging fruit.
2019-12-18 03:14:45 +02:00
restore delta, force, backup '[BACKUP-FULL-2]' - restore links as directories (db-master host)
Reduce log level for all expect tests to detail. The C code is designed to be efficient rather than deterministic at the debug log level. As we move more testing from integration to unit tests it makes less sense to try and maintain the expect logs at this log level. Most of the expect logs have already been moved to detail level but mock/all still had tests at debug level. Change the logging defaults in the config file and remove as many references to log-level-console as possible.
2019-05-23 00:23:44 +02:00
> [CONTAINER-EXEC] db-master [BACKREST-BIN] --config=[TEST_PATH]/db-master/pgbackrest.conf --delta --force --set=[BACKUP-FULL-2] --stanza=db restore
Closed #127: More sanity checking for --delta restores Added checks for `--delta` and `--force` restore options to ensure that the destination is a valid $PGDATA directory. pgBackRest will check for the presence of `PG_VERSION` or `backup.manifest` (left over from an aborted restore). If neither is found then `--delta` and `--force` will be disabled but the restore will proceed unless there are files in the $PGDATA directory (or any tablespace directories) in which case the operation will be aborted.
2016-02-06 06:03:29 +02:00
------------------------------------------------------------------------------------------------------------------------------------
Remove compress/compress-level options from commands where unused. These commands (e.g. restore, archive-get) never used the compress options but allowed them to be passed on the command line. Now they will error when these options are passed on the command line. If these errors occur then remove the unused options.
2020-02-27 19:25:32 +02:00
P00 INFO: restore command begin [BACKREST-VERSION]: --buffer-size=[BUFFER-SIZE] --config=[TEST_PATH]/db-master/pgbackrest.conf --delta --force --lock-path=[TEST_PATH]/db-master/lock --log-level-console=detail --log-level-file=[LOG-LEVEL-FILE] --log-level-stderr=off --log-path=[TEST_PATH]/db-master/log[] --no-log-timestamp --pg1-path=[TEST_PATH]/db-master/db/base --protocol-timeout=60 --repo1-path=[TEST_PATH]/db-master/repo --set=[BACKUP-FULL-2] --stanza=db
Show Process ID in log output instead of filtering it out with the timestamp.
2016-09-14 23:30:57 +02:00
P00 INFO: restore backup set [BACKUP-FULL-2]
The restore command is implemented entirely in C. For the most part this is a direct migration of the Perl code into C. There is one important behavioral change with regard to how file permissions are handled. The Perl code tried to set ownership as it was in the manifest even when running as an unprivileged user. This usually just led to errors and frustration. The C code works like this: If a restore is run as a non-root user (the typical scenario) then all files restored will belong to the user/group executing pgBackRest. If existing files are not owned by the executing user/group then an error will result if the ownership cannot be updated to the executing user/group. In that case the file ownership will need to be updated by a privileged user before the restore can be retried. If a restore is run as the root user then pgBackRest will attempt to recreate the ownership recorded in the manifest when the backup was made. Only user/group names are stored in the manifest so the same names must exist on the restore host for this to work. If the user/group name cannot be found locally then the user/group of the PostgreSQL data directory will be used and finally root if the data directory user/group cannot be mapped to a name. Reviewed by Cynthia Shang.
2019-09-26 13:52:02 +02:00
P00 WARN: file link 'pg_hba.conf' will be restored as a file at the same location
P00 WARN: contents of directory link 'pg_stat' will be restored in a directory at the same location
P00 WARN: file link 'postgresql.conf' will be restored as a file at the same location
P00 WARN: unknown user in backup manifest mapped to current user
P00 WARN: unknown user 'bogus' in backup manifest mapped to current user
P00 WARN: unknown group in backup manifest mapped to current group
P00 WARN: unknown group 'bogus' in backup manifest mapped to current group
P00 DETAIL: check '[TEST_PATH]/db-master/db/base' exists
P00 DETAIL: remove 'global/pg_control' so cluster will not start if restore does not complete
P00 INFO: remove invalid files/links/paths from '[TEST_PATH]/db-master/db/base'
P00 DETAIL: remove invalid link '[TEST_PATH]/db-master/db/base/pg_hba.conf'
P00 DETAIL: remove invalid link '[TEST_PATH]/db-master/db/base/pg_stat'
P00 DETAIL: remove invalid link '[TEST_PATH]/db-master/db/base/postgresql.conf'
P00 DETAIL: remove invalid file '[TEST_PATH]/db-master/db/base/recovery.conf'
P00 DETAIL: create path '[TEST_PATH]/db-master/db/base/pg_stat'
Reduce log level for all expect tests to detail. The C code is designed to be efficient rather than deterministic at the debug log level. As we move more testing from integration to unit tests it makes less sense to try and maintain the expect logs at this log level. Most of the expect logs have already been moved to detail level but mock/all still had tests at debug level. Change the logging defaults in the config file and remove as many references to log-level-console as possible.
2019-05-23 00:23:44 +02:00
P01 DETAIL: restore file [TEST_PATH]/db-master/db/base/base/32768/33001 - exists and matches size 65536 and modification time [MODIFICATION-TIME-1] (64KB, 33%) checksum 6bf316f11d28c28914ea9be92c00de9bea6d9a6b
Fix incorrect offline upper bound for ignoring page checksum errors. For offline backups the upper bound was being set to 0x0000FFFF0000FFFF rather than UINT64_MAX. This meant that page checksum errors might be ignored for databases with a lot of past WAL in offline mode. Online mode is not affected since the upper bound is retrieved from pg_start_backup().
2019-07-11 15:13:56 +02:00
P01 DETAIL: restore file [TEST_PATH]/db-master/db/base/base/32768/44000_init - exists and matches size 32768 and modification time [MODIFICATION-TIME-1] (32KB, 49%) checksum 7a16d165e4775f7c92e8cdf60c0af57313f0bf90
P01 DETAIL: restore file [TEST_PATH]/db-master/db/base/base/32768/33000.32767 - exists and matches size 32768 and modification time [MODIFICATION-TIME-1] (32KB, 66%) checksum 6e99b589e550e68e934fd235ccba59fe5b592a9e
P01 DETAIL: restore file [TEST_PATH]/db-master/db/base/base/32768/33000 - exists and matches size 32768 and modification time [MODIFICATION-TIME-1] (32KB, 83%) checksum 7a16d165e4775f7c92e8cdf60c0af57313f0bf90
Reduce log level for all expect tests to detail. The C code is designed to be efficient rather than deterministic at the debug log level. As we move more testing from integration to unit tests it makes less sense to try and maintain the expect logs at this log level. Most of the expect logs have already been moved to detail level but mock/all still had tests at debug level. Change the logging defaults in the config file and remove as many references to log-level-console as possible.
2019-05-23 00:23:44 +02:00
P01 DETAIL: restore file [TEST_PATH]/db-master/db/base/base/16384/17000 - exists and matches size 16384 and modification time [MODIFICATION-TIME-1] (16KB, 91%) checksum e0101dd8ffb910c9c202ca35b5f828bcb9697bed
Support configurable WAL segment size. PostgreSQL 11 introduces configurable WAL segment sizes, from 1MB to 1GB. There are two areas that needed to be updated to support this: building the archive-get queue and checking that WAL has been archived after a backup. Both operations require the WAL segment size to properly build a list. Checking the archive after a backup is still implemented in Perl and has an active database connection, so just get the WAL segment size from the database. The archive-get command does not have a connection to the database, so get the WAL segment size from pg_control instead. This requires a deeper inspection of pg_control than has been done in the past, so it seemed best to copy the relevant data structures from each version of PostgreSQL and build a generic interface layer to address them. While this approach is a bit verbose, it has the advantage of being relatively simple, and can easily be updated for new versions of PostgreSQL. Since the integration tests generate pg_control files for testing, teach Perl how to generate files with the correct offsets for both 32-bit and 64-bit architectures.
2018-09-25 11:24:42 +02:00
P01 INFO: restore file [TEST_PATH]/db-master/db/base/global/pg_control.pgbackrest.tmp (8KB, 95%) checksum 4c77c900f7af0d9ab13fa9982051a42e0b637f6c
Reduce log level for all expect tests to detail. The C code is designed to be efficient rather than deterministic at the debug log level. As we move more testing from integration to unit tests it makes less sense to try and maintain the expect logs at this log level. Most of the expect logs have already been moved to detail level but mock/all still had tests at debug level. Change the logging defaults in the config file and remove as many references to log-level-console as possible.
2019-05-23 00:23:44 +02:00
P01 DETAIL: restore file [TEST_PATH]/db-master/db/base/base/1/12000 - exists and matches size 8192 and modification time [MODIFICATION-TIME-1] (8KB, 99%) checksum 22c98d248ff548311eda88559e4a8405ed77c003
For simplicity, the pg_control file is now copied with the rest of the files instead of by itself of at the end of the process. The backup command does not require this behavior and the restore copies to a temporary file which is renamed at the end of the restore.
2016-11-30 21:23:05 +02:00
P01 INFO: restore file [TEST_PATH]/db-master/db/base/postgresql.conf (21B, 99%) checksum 6721d92c9fcdf4248acff1f9a1377127d9064807
Fix failure in manifest build when two or more files in PGDATA are linked to the same directory. Reported by Vitaliy Kukharik.
2018-05-02 18:19:54 +02:00
P01 INFO: restore file [TEST_PATH]/db-master/db/base/pg_hba.conf (9B, 99%) checksum dd4cea0cae348309f9de28ad4ded8ee2cc2e6d5b
Reduce log level for all expect tests to detail. The C code is designed to be efficient rather than deterministic at the debug log level. As we move more testing from integration to unit tests it makes less sense to try and maintain the expect logs at this log level. Most of the expect logs have already been moved to detail level but mock/all still had tests at debug level. Change the logging defaults in the config file and remove as many references to log-level-console as possible.
2019-05-23 00:23:44 +02:00
P01 DETAIL: restore file [TEST_PATH]/db-master/db/base/changecontent.txt - exists and matches size 7 and modification time [MODIFICATION-TIME-1] (7B, 99%) checksum 238a131a3e8eb98d1fc5b27d882ca40b7618fd2a
For simplicity, the pg_control file is now copied with the rest of the files instead of by itself of at the end of the process. The backup command does not require this behavior and the restore copies to a temporary file which is renamed at the end of the restore.
2016-11-30 21:23:05 +02:00
P01 INFO: restore file [TEST_PATH]/db-master/db/base/pg_stat/global.stat (5B, 99%) checksum e350d5ce0153f3e22d5db21cf2a4eff00f3ee877
Reduce log level for all expect tests to detail. The C code is designed to be efficient rather than deterministic at the debug log level. As we move more testing from integration to unit tests it makes less sense to try and maintain the expect logs at this log level. Most of the expect logs have already been moved to detail level but mock/all still had tests at debug level. Change the logging defaults in the config file and remove as many references to log-level-console as possible.
2019-05-23 00:23:44 +02:00
P01 DETAIL: restore file [TEST_PATH]/db-master/db/base/changetime.txt - exists and matches size 4 and modification time [MODIFICATION-TIME-1] (4B, 99%) checksum 88087292ed82e26f3eb824d0bffc05ccf7a30f8d
P01 DETAIL: restore file [TEST_PATH]/db-master/db/base/base/32768/PG_VERSION - exists and matches size 3 and modification time [MODIFICATION-TIME-1] (3B, 99%) checksum 184473f470864e067ee3a22e64b47b0a1c356f29
P01 DETAIL: restore file [TEST_PATH]/db-master/db/base/base/16384/PG_VERSION - exists and matches size 3 and modification time [MODIFICATION-TIME-1] (3B, 99%) checksum 184473f470864e067ee3a22e64b47b0a1c356f29
P01 DETAIL: restore file [TEST_PATH]/db-master/db/base/base/1/PG_VERSION - exists and matches size 3 and modification time [MODIFICATION-TIME-1] (3B, 99%) checksum 184473f470864e067ee3a22e64b47b0a1c356f29
Remove integration tests that are now covered in the unit tests. Most of these tests are just checking that errors are thrown when required. These are well covered in various unit tests. The "cannot resume" tests are also well covered in the backup unit tests. Finally, config warnings are well covered in the config unit tests. There is more to be done here, but this accounts for the low-hanging fruit.
2019-12-18 03:14:45 +02:00
P01 DETAIL: restore file [TEST_PATH]/db-master/db/base/PG_VERSION - exists and matches size 3 and modification time [MODIFICATION-TIME-1] (3B, 100%) checksum 184473f470864e067ee3a22e64b47b0a1c356f29
Reduce log level for all expect tests to detail. The C code is designed to be efficient rather than deterministic at the debug log level. As we move more testing from integration to unit tests it makes less sense to try and maintain the expect logs at this log level. Most of the expect logs have already been moved to detail level but mock/all still had tests at debug level. Change the logging defaults in the config file and remove as many references to log-level-console as possible.
2019-05-23 00:23:44 +02:00
P01 DETAIL: restore file [TEST_PATH]/db-master/db/base/zero_from_start - exists and matches size 0 and modification time [MODIFICATION-TIME-1] (0B, 100%)
Use minio for integration testing. ScalityS3 has not received any maintenance in years and is slow to start which is bad for testing. Replace it with minio which starts quickly and ships as a single executable or a tiny container. Minio has stricter limits on allowable characters but should still provide enough coverage to show that our encoding is working correctly. This commit also includes the upgrade to openssl 1.1.1 in the Ubuntu 18.04 container.
2019-07-03 04:20:35 +02:00
P01 DETAIL: restore file [TEST_PATH]/db-master/db/base/special-!_.*'()&!@;:+,? - exists and matches size 0 and modification time [MODIFICATION-TIME-1] (0B, 100%)
For simplicity, the pg_control file is now copied with the rest of the files instead of by itself of at the end of the process. The backup command does not require this behavior and the restore copies to a temporary file which is renamed at the end of the restore.
2016-11-30 21:23:05 +02:00
P00 INFO: write [TEST_PATH]/db-master/db/base/recovery.conf
The restore command is implemented entirely in C. For the most part this is a direct migration of the Perl code into C. There is one important behavioral change with regard to how file permissions are handled. The Perl code tried to set ownership as it was in the manifest even when running as an unprivileged user. This usually just led to errors and frustration. The C code works like this: If a restore is run as a non-root user (the typical scenario) then all files restored will belong to the user/group executing pgBackRest. If existing files are not owned by the executing user/group then an error will result if the ownership cannot be updated to the executing user/group. In that case the file ownership will need to be updated by a privileged user before the restore can be retried. If a restore is run as the root user then pgBackRest will attempt to recreate the ownership recorded in the manifest when the backup was made. Only user/group names are stored in the manifest so the same names must exist on the restore host for this to work. If the user/group name cannot be found locally then the user/group of the PostgreSQL data directory will be used and finally root if the data directory user/group cannot be mapped to a name. Reviewed by Cynthia Shang.
2019-09-26 13:52:02 +02:00
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base/base'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base/base/1'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base/base/16384'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base/base/32768'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base/pg_clog'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base/pg_dynshmem'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base/pg_log'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base/pg_notify'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base/pg_replslot'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base/pg_serial'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base/pg_snapshots'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base/pg_stat'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base/pg_stat_tmp'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base/pg_subtrans'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base/pg_tblspc'
Storage and IO layer refactor: Refactor storage layer to allow for new repository filesystems using drivers. (Reviewed by Cynthia Shang.) Refactor IO layer to allow for new compression formats, checksum types, and other capabilities using filters. (Reviewed by Cynthia Shang.)
2017-06-09 23:51:41 +02:00
P00 INFO: restore global/pg_control (performed last to ensure aborted restores cannot be started)
The restore command is implemented entirely in C. For the most part this is a direct migration of the Perl code into C. There is one important behavioral change with regard to how file permissions are handled. The Perl code tried to set ownership as it was in the manifest even when running as an unprivileged user. This usually just led to errors and frustration. The C code works like this: If a restore is run as a non-root user (the typical scenario) then all files restored will belong to the user/group executing pgBackRest. If existing files are not owned by the executing user/group then an error will result if the ownership cannot be updated to the executing user/group. In that case the file ownership will need to be updated by a privileged user before the restore can be retried. If a restore is run as the root user then pgBackRest will attempt to recreate the ownership recorded in the manifest when the backup was made. Only user/group names are stored in the manifest so the same names must exist on the restore host for this to work. If the user/group name cannot be found locally then the user/group of the PostgreSQL data directory will be used and finally root if the data directory user/group cannot be mapped to a name. Reviewed by Cynthia Shang.
2019-09-26 13:52:02 +02:00
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base/global'
Added a predictable log entry to signal that a command has completed successfully. For example a backup ends successfully with: INFO: backup command end: completed successfully.
2016-12-04 21:44:53 +02:00
P00 INFO: restore command end: completed successfully
Closed #127: More sanity checking for --delta restores Added checks for `--delta` and `--force` restore options to ensure that the destination is a valid $PGDATA directory. pgBackRest will check for the presence of `PG_VERSION` or `backup.manifest` (left over from an aborted restore). If neither is found then `--delta` and `--force` will be disabled but the restore will proceed unless there are files in the $PGDATA directory (or any tablespace directories) in which case the operation will be aborted.
2016-02-06 06:03:29 +02:00
Rename test paths for clarity. This was worked out as part of the test suite refactor [c8f806a] but not committed with it because of the large number of expect logs changes involved. Keeping them separate made it easier to audit the changes in the refactor.
2016-06-24 16:54:31 +02:00
+ supplemental file: [TEST_PATH]/db-master/db/base/recovery.conf
----------------------------------------------------------------
PostgreSQL 12 support. Recovery settings are now written into postgresql.auto.conf instead of recovery.conf. Existing recovery_target* settings will be commented out to help avoid conflicts. A comment is added before recovery settings to identify them as written by pgBackRest since it is unclear how, in general, old settings will be removed. recovery.signal and standby.signal are automatically created based on the recovery settings.
2019-10-01 19:20:43 +02:00
# Recovery settings generated by pgBackRest restore on [TIMESTAMP]
Reduce log level for all expect tests to detail. The C code is designed to be efficient rather than deterministic at the debug log level. As we move more testing from integration to unit tests it makes less sense to try and maintain the expect logs at this log level. Most of the expect logs have already been moved to detail level but mock/all still had tests at debug level. Change the logging defaults in the config file and remove as many references to log-level-console as possible.
2019-05-23 00:23:44 +02:00
restore_command = '[BACKREST-BIN] --config=[TEST_PATH]/db-master/pgbackrest.conf --stanza=db archive-get %f "%p"'
Closed #127: More sanity checking for --delta restores Added checks for `--delta` and `--force` restore options to ensure that the destination is a valid $PGDATA directory. pgBackRest will check for the presence of `PG_VERSION` or `backup.manifest` (left over from an aborted restore). If neither is found then `--delta` and `--force` will be disabled but the restore will proceed unless there are files in the $PGDATA directory (or any tablespace directories) in which case the operation will be aborted.
2016-02-06 06:03:29 +02:00
Major refactor of the test suite. * Make the code more modular and object-oriented. * Multiple Docker containers can now be created for a single test to simulate more realistic environments.
2016-06-24 14:12:58 +02:00
incr backup - add tablespace 1 (db-master host)
Remove test point infrastructure. 82df7e6f and 9856fef5 updated tests that used test points in preparation for the feature not being available in the C code. Since tests points are no longer used remove the infrastructure. Also remove one stray --test option in mock/all that was essentially a noop but no longer works now that the option has been removed.
2019-12-10 20:16:47 +02:00
> [CONTAINER-EXEC] db-master [BACKREST-BIN] --config=[TEST_PATH]/db-master/pgbackrest.conf --no-online --stanza=db backup
Log testing can now be enabled for certain deterministic tests. This works by comparing the generated logs against a previous copy. Currently only enabled for the backup/synthetic tests.
2015-05-05 19:08:48 +02:00
------------------------------------------------------------------------------------------------------------------------------------
Add infrastructure for multiple compression type support. Add compress-type option and deprecate compress option. Since the compress option is boolean it won't work with multiple compression types. Add logic to cfgLoadUpdateOption() to update compress-type if it is not set directly. The compress option should no longer be referenced outside the cfgLoadUpdateOption() function. Add common/compress/helper module to contain interface functions that work with multiple compression types. Code outside this module should no longer call specific compression drivers, though it may be OK to reference a specific compression type using the new interface (e.g., saving backup history files in gz format). Unit tests only test compression using the gz format because other formats may not be available in all builds. It is the job of integration tests to exercise all compression types. Additional compression types will be added in future commits.
2020-03-06 21:41:03 +02:00
P00 INFO: backup command begin [BACKREST-VERSION]: --buffer-size=[BUFFER-SIZE] --compress-level=3 --compress-type=none --config=[TEST_PATH]/db-master/pgbackrest.conf --db-timeout=45 --lock-path=[TEST_PATH]/db-master/lock --log-level-console=detail --log-level-file=[LOG-LEVEL-FILE] --log-level-stderr=off --log-path=[TEST_PATH]/db-master/log[] --no-log-timestamp --no-online --pg1-path=[TEST_PATH]/db-master/db/base --protocol-timeout=60 --repo1-path=[TEST_PATH]/db-master/repo --stanza=db --start-fast
Add time-based retention for full backups. The --repo-retention-full-type option allows retention of full backups based on a time period, specified in days. The new option will default to 'count' and therefore will not affect current installations. Setting repo-retention-full-type to 'time' will allow the user to use a time period, in days, to indicate full backup retention. Using this method, a full backup can be expired only if the time the backup completed is older than the number of days set with repo-retention-full (calculated from the moment the 'expire' command is run) and at least one full backup meets the retention period. If archive retention has not been configured, then the default settings will expire archives that are prior to the oldest retained full backup. For example, if there are three full backups ending in times that are 25 days old (F1), 20 days old (F2) and 10 days old (F3), then if the full retention period is 15 days, then only F1 will be expired; F2 will be retained because F1 is not at least 15 days old.
2020-05-08 21:25:03 +02:00
P00 WARN: option 'repo1-retention-full' is not set for 'repo1-retention-full-type=count', the repository may run out of space
Rename retention-* options to repo-retention-*.
2018-02-19 22:07:24 +02:00
HINT: to retain full backups indefinitely (without warning), set option 'repo1-retention-full' to the maximum.
Improvements and additional tests for stanza-upgrade. Contributed by Cynthia Shang.
2017-04-07 01:36:37 +02:00
P00 INFO: last backup label = [BACKUP-FULL-2], version = [VERSION-1]
Added the --checksum-page option. This option allows pgBackRest to validate page checksums in data files when checksums are enabled on PostgreSQL >= 9.3. Note that this functionality requires a C library which may not initially be available in OS packages. The option will automatically be enabled when the library is present and checksums are enabled on the cluster.
2016-12-13 01:54:07 +02:00
P00 WARN: incr backup cannot alter 'checksum-page' option to 'false', reset to 'true' from [BACKUP-FULL-2]
Fix critical bug in resume that resulted in inconsistent backups. A regression in v0.82 removed the timestamp comparison when deciding which files from the aborted backup to keep on resume. All resumed backups should be considered inconsistent. A resumed backup can be identified by checking the log for the message "aborted backup of same type exists, will be cleaned to remove invalid files and resumed". Reported by David Youatt, Yogesh Sharma, Stephen Frost.
2018-07-03 20:01:57 +02:00
P01 INFO: backup file [TEST_PATH]/db-master/db/base/badchecksum.txt (11B, 50%) checksum f927212cd08d11a42a666b2f04235398e9ceeb51
P01 INFO: backup file [TEST_PATH]/db-master/db/base/changesize.txt (4B, 68%) checksum 88087292ed82e26f3eb824d0bffc05ccf7a30f8d
Add zero-length file to mock/all test.
2018-07-05 21:40:50 +02:00
P01 INFO: backup file [TEST_PATH]/db-master/db/base/zerosize.txt (0B, 68%)
Fix critical bug in resume that resulted in inconsistent backups. A regression in v0.82 removed the timestamp comparison when deciding which files from the aborted backup to keep on resume. All resumed backups should be considered inconsistent. A resumed backup can be identified by checking the log for the message "aborted backup of same type exists, will be cleaned to remove invalid files and resumed". Reported by David Youatt, Yogesh Sharma, Stephen Frost.
2018-07-03 20:01:57 +02:00
P01 INFO: backup file [TEST_PATH]/db-master/db/base/pg_tblspc/1/[TS_PATH-1]/16384/tablespace1.txt (7B, 100%) checksum 14c44cef6287269b08d41de489fd492bb9fc795d
Added the --checksum-page option. This option allows pgBackRest to validate page checksums in data files when checksums are enabled on PostgreSQL >= 9.3. Note that this functionality requires a C library which may not initially be available in OS packages. The option will automatically be enabled when the library is present and checksums are enabled on the cluster.
2016-12-13 01:54:07 +02:00
P00 WARN: page misalignment in file [TEST_PATH]/db-master/db/base/pg_tblspc/1/[TS_PATH-1]/16384/tablespace1.txt: file size 7 is not divisible by page size 8192
The backup command is implemented entirely in C. For the most part this is a direct migration of the Perl code into C except as noted below. A backup can now be initiated from a linked directory. The link will not be stored in the manifest or recreated on restore. If a link or directory does not already exist in the restore location then a directory will be created. The logic for creating backup labels has been improved and it should no longer be possible to get a backup label earlier than the latest backup even with timezone changes or clock skew. This has never been an issue in the field that we know of, but we found it in testing. For online backups all times are fetched from the PostgreSQL primary host (before only copy start was). This doesn't affect backup integrity but it does prevent clock skew between hosts affecting backup duration reporting. Archive copy now works as expected when the archive and backup have different compression settings, i.e. when one is compressed and the other is not. This was a long-standing bug in the Perl code. Resume will now work even if hardlink settings have been changed. Reviewed by Cynthia Shang.
2019-12-14 00:14:26 +02:00
P00 DETAIL: reference pg_data/PG_VERSION to [BACKUP-FULL-2]
P00 DETAIL: reference pg_data/base/1/12000 to [BACKUP-FULL-2]
P00 DETAIL: reference pg_data/base/1/PG_VERSION to [BACKUP-FULL-2]
P00 DETAIL: reference pg_data/base/16384/17000 to [BACKUP-FULL-2]
P00 DETAIL: reference pg_data/base/16384/PG_VERSION to [BACKUP-FULL-2]
P00 DETAIL: reference pg_data/base/32768/33000 to [BACKUP-FULL-2]
P00 DETAIL: reference pg_data/base/32768/33000.32767 to [BACKUP-FULL-2]
P00 DETAIL: reference pg_data/base/32768/33001 to [BACKUP-FULL-2]
P00 DETAIL: reference pg_data/base/32768/44000_init to [BACKUP-FULL-2]
P00 DETAIL: reference pg_data/base/32768/PG_VERSION to [BACKUP-FULL-2]
P00 DETAIL: reference pg_data/changecontent.txt to [BACKUP-FULL-2]
P00 DETAIL: reference pg_data/changetime.txt to [BACKUP-FULL-2]
P00 DETAIL: reference pg_data/global/pg_control to [BACKUP-FULL-2]
P00 DETAIL: reference pg_data/pg_stat/global.stat to [BACKUP-FULL-2]
P00 DETAIL: reference pg_data/postgresql.conf to [BACKUP-FULL-2]
P00 DETAIL: reference pg_data/special-!_.*'()&!@;:+,? to [BACKUP-FULL-2]
P00 DETAIL: reference pg_data/zero_from_start to [BACKUP-FULL-2]
Fix critical bug in resume that resulted in inconsistent backups. A regression in v0.82 removed the timestamp comparison when deciding which files from the aborted backup to keep on resume. All resumed backups should be considered inconsistent. A resumed backup can be identified by checking the log for the message "aborted backup of same type exists, will be cleaned to remove invalid files and resumed". Reported by David Youatt, Yogesh Sharma, Stephen Frost.
2018-07-03 20:01:57 +02:00
P00 INFO: incr backup size = 22B
Show Process ID in log output instead of filtering it out with the timestamp.
2016-09-14 23:30:57 +02:00
P00 INFO: new backup label = [BACKUP-INCR-1]
Added a predictable log entry to signal that a command has completed successfully. For example a backup ends successfully with: INFO: backup command end: completed successfully.
2016-12-04 21:44:53 +02:00
P00 INFO: backup command end: completed successfully
Integration test improvements for disk and memory efficiency. Set log-level-file=off when more that one test will run. In this case is it impossible to see the logs anyway since they will be automatically cleaned up after the test. This improves performance pretty dramatically since trace-level logging is expensive. If a singe integration test is run then log-level-file is trace by default but can be changed with the --log-level-test-file option. Reduce buffer-size to 64k to save memory during testing and allow more processes to run in parallel. Update log replacement rules so that these options can change without affecting expect logs.
2019-12-17 22:23:07 +02:00
P00 INFO: expire command begin [BACKREST-VERSION]: --buffer-size=[BUFFER-SIZE] --config=[TEST_PATH]/db-master/pgbackrest.conf --lock-path=[TEST_PATH]/db-master/lock --log-level-console=detail --log-level-file=[LOG-LEVEL-FILE] --log-level-stderr=off --log-path=[TEST_PATH]/db-master/log[] --no-log-timestamp --repo1-path=[TEST_PATH]/db-master/repo --stanza=db
Rename retention-* options to repo-retention-*.
2018-02-19 22:07:24 +02:00
P00 INFO: option 'repo1-retention-archive' is not set - archive logs will not be expired
Added a predictable log entry to signal that a command has completed successfully. For example a backup ends successfully with: INFO: backup command end: completed successfully.
2016-12-04 21:44:53 +02:00
P00 INFO: expire command end: completed successfully
Log testing can now be enabled for certain deterministic tests. This works by comparing the generated logs against a previous copy. Currently only enabled for the backup/synthetic tests.
2015-05-05 19:08:48 +02:00
Rename test paths for clarity. This was worked out as part of the test suite refactor [c8f806a] but not committed with it because of the large number of expect logs changes involved. Keeping them separate made it easier to audit the changes in the refactor.
2016-06-24 16:54:31 +02:00
+ supplemental file: [TEST_PATH]/db-master/pgbackrest.conf
----------------------------------------------------------
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
[db]
Rename db-* options to pg-* and backup-* options to repo-* to improve consistency. * repo-* options are now indexed although only one is allowed. * List deprecated option names in documentation and command-line help.
2018-02-04 01:27:38 +02:00
pg1-path=[TEST_PATH]/db-master/db/base
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
New simpler configuration and consistent project/exe/path naming. * The repo-path option now always refers to the repository where backups and archive are stored, whether local or remote, so the repo-remote-path option has been removed. The new spool-path option can be used to define a location for queueing WAL segments when archiving asynchronously. Otherwise, a local repository is no longer required. * Implemented a new config format which should be far simpler to use. See the User Guide and Configuration Reference for details but for a simple configuration all options can now be placed in the stanza section. Options that are shared between stanzas can be placed in the [global] section. More complex configurations can still make use of command sections though this should be a rare use case. * The default configuration filename is now pgbackrest.conf instead of pg_backrest.conf. This was done for consistency with other naming changes but also to prevent old config files from being loaded accidentally. * The default repository name was changed from /var/lib/backup to /var/lib/pgbackrest. * Lock files are now stored in /tmp/pgbackrest by default. These days /run/pgbackrest would be the preferred location but that would require init scripts which are not part of this release. The lock-path option can be used to configure the lock directory. * Log files are now stored in /var/log/pgbackrest by default and no longer have the date appended so they can be managed with logrotate. The log-path option can be used to configure the lock directory. * Executable filename changed from pg_backrest to pgbackrest.
2016-04-14 15:30:54 +02:00
[global]
Integration test improvements for disk and memory efficiency. Set log-level-file=off when more that one test will run. In this case is it impossible to see the logs anyway since they will be automatically cleaned up after the test. This improves performance pretty dramatically since trace-level logging is expensive. If a singe integration test is run then log-level-file is trace by default but can be changed with the --log-level-test-file option. Reduce buffer-size to 64k to save memory during testing and allow more processes to run in parallel. Update log replacement rules so that these options can change without affecting expect logs.
2019-12-17 22:23:07 +02:00
buffer-size=[BUFFER-SIZE]
Fix custom settings for compress-level option being ignored. Reported by Jens Wilke.
2017-10-23 00:05:46 +02:00
compress-level=3
Add infrastructure for multiple compression type support. Add compress-type option and deprecate compress option. Since the compress option is boolean it won't work with multiple compression types. Add logic to cfgLoadUpdateOption() to update compress-type if it is not set directly. The compress option should no longer be referenced outside the cfgLoadUpdateOption() function. Add common/compress/helper module to contain interface functions that work with multiple compression types. Code outside this module should no longer call specific compression drivers, though it may be OK to reference a specific compression type using the new interface (e.g., saving backup history files in gz format). Unit tests only test compression using the gz format because other formats may not be available in all builds. It is the job of integration tests to exercise all compression types. Additional compression types will be added in future commits.
2020-03-06 21:41:03 +02:00
compress-type=none
Storage and IO layer refactor: Refactor storage layer to allow for new repository filesystems using drivers. (Reviewed by Cynthia Shang.) Refactor IO layer to allow for new compression formats, checksum types, and other capabilities using filters. (Reviewed by Cynthia Shang.)
2017-06-09 23:51:41 +02:00
db-timeout=45
lock-path=[TEST_PATH]/db-master/lock
Reduce log level for all expect tests to detail. The C code is designed to be efficient rather than deterministic at the debug log level. As we move more testing from integration to unit tests it makes less sense to try and maintain the expect logs at this log level. Most of the expect logs have already been moved to detail level but mock/all still had tests at debug level. Change the logging defaults in the config file and remove as many references to log-level-console as possible.
2019-05-23 00:23:44 +02:00
log-level-console=detail
Integration test improvements for disk and memory efficiency. Set log-level-file=off when more that one test will run. In this case is it impossible to see the logs anyway since they will be automatically cleaned up after the test. This improves performance pretty dramatically since trace-level logging is expensive. If a singe integration test is run then log-level-file is trace by default but can be changed with the --log-level-test-file option. Reduce buffer-size to 64k to save memory during testing and allow more processes to run in parallel. Update log replacement rules so that these options can change without affecting expect logs.
2019-12-17 22:23:07 +02:00
log-level-file=[LOG-LEVEL-FILE]
Added the log-level-stderr option. Controls whether console log messages are sent to stderr or stdout. By default this is set to warn which represents a change in behavior from previous versions, even though it may be more intuitive. Setting log-level-stderr=off will preserve the old behavior. Suggested by Sascha Biberhofer.
2016-10-05 15:09:30 +02:00
log-level-stderr=off
Storage and IO layer refactor: Refactor storage layer to allow for new repository filesystems using drivers. (Reviewed by Cynthia Shang.) Refactor IO layer to allow for new compression formats, checksum types, and other capabilities using filters. (Reviewed by Cynthia Shang.)
2017-06-09 23:51:41 +02:00
log-path=[TEST_PATH]/db-master/log
Integration test improvements for disk and memory efficiency. Set log-level-file=off when more that one test will run. In this case is it impossible to see the logs anyway since they will be automatically cleaned up after the test. This improves performance pretty dramatically since trace-level logging is expensive. If a singe integration test is run then log-level-file is trace by default but can be changed with the --log-level-test-file option. Reduce buffer-size to 64k to save memory during testing and allow more processes to run in parallel. Update log replacement rules so that these options can change without affecting expect logs.
2019-12-17 22:23:07 +02:00
log-subprocess=[LOG-SUBPROCESS]
Set log-timestamp=n for integration tests. This means less filtering of logs needs to be done and new timestamps can be added without adding new filters.
2018-05-11 17:24:38 +02:00
log-timestamp=n
Storage and IO layer refactor: Refactor storage layer to allow for new repository filesystems using drivers. (Reviewed by Cynthia Shang.) Refactor IO layer to allow for new compression formats, checksum types, and other capabilities using filters. (Reviewed by Cynthia Shang.)
2017-06-09 23:51:41 +02:00
protocol-timeout=60
Rename db-* options to pg-* and backup-* options to repo-* to improve consistency. * repo-* options are now indexed although only one is allowed. * List deprecated option names in documentation and command-line help.
2018-02-04 01:27:38 +02:00
repo1-path=[TEST_PATH]/db-master/repo
Consolidate archive-push tests into new archive test.
2017-06-21 14:38:53 +02:00
spool-path=[TEST_PATH]/db-master/spool
New simpler configuration and consistent project/exe/path naming. * The repo-path option now always refers to the repository where backups and archive are stored, whether local or remote, so the repo-remote-path option has been removed. The new spool-path option can be used to define a location for queueing WAL segments when archiving asynchronously. Otherwise, a local repository is no longer required. * Implemented a new config format which should be far simpler to use. See the User Guide and Configuration Reference for details but for a simple configuration all options can now be placed in the stanza section. Options that are shared between stanzas can be placed in the [global] section. More complex configurations can still make use of command sections though this should be a rare use case. * The default configuration filename is now pgbackrest.conf instead of pg_backrest.conf. This was done for consistency with other naming changes but also to prevent old config files from being loaded accidentally. * The default repository name was changed from /var/lib/backup to /var/lib/pgbackrest. * Lock files are now stored in /tmp/pgbackrest by default. These days /run/pgbackrest would be the preferred location but that would require init scripts which are not part of this release. The lock-path option can be used to configure the lock directory. * Log files are now stored in /var/log/pgbackrest by default and no longer have the date appended so they can be managed with logrotate. The log-path option can be used to configure the lock directory. * Executable filename changed from pg_backrest to pgbackrest.
2016-04-14 15:30:54 +02:00
[global:backup]
archive-copy=y
start-fast=y
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
Rename test paths for clarity. This was worked out as part of the test suite refactor [c8f806a] but not committed with it because of the large number of expect logs changes involved. Keeping them separate made it easier to audit the changes in the refactor.
2016-06-24 16:54:31 +02:00
+ supplemental file: [TEST_PATH]/db-master/repo/backup/db/[BACKUP-INCR-1]/backup.manifest
-----------------------------------------------------------------------------------------
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
[backrest]
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
backrest-format=5
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
backrest-version="[VERSION-1]"
[backup]
backup-label="[BACKUP-INCR-1]"
backup-prior="[BACKUP-FULL-2]"
backup-timestamp-copy-start=[TIMESTAMP]
backup-timestamp-start=[TIMESTAMP]
backup-timestamp-stop=[TIMESTAMP]
backup-type="incr"
[backup:db]
Database version refactoring. * Refactor db version constants into a separate module. * Update synthetic backup tests to PostgreSQL 9.4.
2016-08-12 04:35:24 +02:00
db-catalog-version=201409291
db-control-version=942
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
db-id=1
Automate generation of WAL and pg_control test files. The existing static files would not work with 32-bit or big-endian systems so create functions to generate these files dynamically rather than creating a bunch of new static files.
2017-11-19 03:02:54 +02:00
db-system-id=1000000000000000094
Database version refactoring. * Refactor db version constants into a separate module. * Update synthetic backup tests to PostgreSQL 9.4.
2016-08-12 04:35:24 +02:00
db-version="9.4"
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
[backup:option]
Set archive-check option in manifest correctly when offline. Archive check does not run when in offline backup mode but the option was set to true in the manifest. It's harmless since these options are informational only but it could cause confusion when debugging.
2019-11-28 15:27:21 +02:00
option-archive-check=false
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
option-archive-copy=true
Backup from a standby cluster. A connection to the primary cluster is still required to start/stop the backup and copy files that are not replicated, but the vast majority of files are copied from the standby in order to reduce load on the master.
2016-08-25 17:25:46 +02:00
option-backup-standby=false
Integration test improvements for disk and memory efficiency. Set log-level-file=off when more that one test will run. In this case is it impossible to see the logs anyway since they will be automatically cleaned up after the test. This improves performance pretty dramatically since trace-level logging is expensive. If a singe integration test is run then log-level-file is trace by default but can be changed with the --log-level-test-file option. Reduce buffer-size to 64k to save memory during testing and allow more processes to run in parallel. Update log replacement rules so that these options can change without affecting expect logs.
2019-12-17 22:23:07 +02:00
option-buffer-size=[BUFFER-SIZE]
Added the --checksum-page option. This option allows pgBackRest to validate page checksums in data files when checksums are enabled on PostgreSQL >= 9.3. Note that this functionality requires a C library which may not initially be available in OS packages. The option will automatically be enabled when the library is present and checksums are enabled on the cluster.
2016-12-13 01:54:07 +02:00
option-checksum-page=true
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
option-compress=false
Add additional options to backup.manifest for debugging purposes. Add the buffer-size, compress-level, compress-level-network, and process-max options to the backup:option section in backup.manifest to aid in debugging. It may also make sense to propagate these options up to backup.info so they can be displayed in the info command, but for now this is deemed sufficient. Contributed by blogh.
2019-03-10 11:03:52 +02:00
option-compress-level=3
option-compress-level-network=3
Add infrastructure for multiple compression type support. Add compress-type option and deprecate compress option. Since the compress option is boolean it won't work with multiple compression types. Add logic to cfgLoadUpdateOption() to update compress-type if it is not set directly. The compress option should no longer be referenced outside the cfgLoadUpdateOption() function. Add common/compress/helper module to contain interface functions that work with multiple compression types. Code outside this module should no longer call specific compression drivers, though it may be OK to reference a specific compression type using the new interface (e.g., saving backup history files in gz format). Unit tests only test compression using the gz format because other formats may not be available in all builds. It is the job of integration tests to exercise all compression types. Additional compression types will be added in future commits.
2020-03-06 21:41:03 +02:00
option-compress-type="none"
Add checksum delta for incremental backups. Use checksums rather than timestamps to determine if files have changed. This is useful in cases where the timestamps may not be trustworthy, e.g. when performing an incremental after failing over to a standby. If checksum delta is enabled then checksums will be used for verification of resumed backups, even if they are full. Resumes have always used checksums to verify the files in the repository, enabling delta performs checksums on the database files as well. Note that the user must manually enable this feature in cases were it would be useful or just keep in enabled all the time. A future commit will address automatically enabling the feature in cases where it seems likely to be useful. Contributed by Cynthia Shang.
2018-09-19 17:12:45 +02:00
option-delta=false
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
option-hardlink=false
Closed #167: Rename `--no-start-stop` option to `--no-online`.
2016-02-20 02:32:02 +02:00
option-online=false
Add additional options to backup.manifest for debugging purposes. Add the buffer-size, compress-level, compress-level-network, and process-max options to the backup:option section in backup.manifest to aid in debugging. It may also make sense to propagate these options up to backup.info so they can be displayed in the info command, but for now this is deemed sufficient. Contributed by blogh.
2019-03-10 11:03:52 +02:00
option-process-max=1
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
[backup:target]
Rename test paths for clarity. This was worked out as part of the test suite refactor [c8f806a] but not committed with it because of the large number of expect logs changes involved. Keeping them separate made it easier to audit the changes in the refactor.
2016-06-24 16:54:31 +02:00
pg_data={"path":"[TEST_PATH]/db-master/db/base","type":"path"}
pg_tblspc/1={"path":"[TEST_PATH]/db-master/db/tablespace/ts1","tablespace-id":"1","tablespace-name":"ts1","type":"link"}
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
[target:file]
Add checksum delta for incremental backups. Use checksums rather than timestamps to determine if files have changed. This is useful in cases where the timestamps may not be trustworthy, e.g. when performing an incremental after failing over to a standby. If checksum delta is enabled then checksums will be used for verification of resumed backups, even if they are full. Resumes have always used checksums to verify the files in the repository, enabling delta performs checksums on the database files as well. Note that the user must manually enable this feature in cases were it would be useful or just keep in enabled all the time. A future commit will address automatically enabling the feature in cases where it seems likely to be useful. Contributed by Cynthia Shang.
2018-09-19 17:12:45 +02:00
pg_data/PG_VERSION={"checksum":"184473f470864e067ee3a22e64b47b0a1c356f29","reference":"[BACKUP-FULL-2]","size":[SIZE],"timestamp":[TIMESTAMP-1]}
pg_data/badchecksum.txt={"checksum":"f927212cd08d11a42a666b2f04235398e9ceeb51","size":[SIZE],"timestamp":[TIMESTAMP-1]}
pg_data/base/1/12000={"checksum":"22c98d248ff548311eda88559e4a8405ed77c003","checksum-page":true,"master":false,"reference":"[BACKUP-FULL-2]","size":[SIZE],"timestamp":[TIMESTAMP-1]}
pg_data/base/1/PG_VERSION={"checksum":"184473f470864e067ee3a22e64b47b0a1c356f29","master":false,"mode":"0660","reference":"[BACKUP-FULL-2]","size":[SIZE],"timestamp":[TIMESTAMP-1]}
pg_data/base/16384/17000={"checksum":"e0101dd8ffb910c9c202ca35b5f828bcb9697bed","checksum-page":false,"checksum-page-error":[1],"master":false,"reference":"[BACKUP-FULL-2]","size":[SIZE],"timestamp":[TIMESTAMP-1]}
pg_data/base/16384/PG_VERSION={"checksum":"184473f470864e067ee3a22e64b47b0a1c356f29","master":false,"reference":"[BACKUP-FULL-2]","size":[SIZE],"timestamp":[TIMESTAMP-1]}
Fix incorrect offline upper bound for ignoring page checksum errors. For offline backups the upper bound was being set to 0x0000FFFF0000FFFF rather than UINT64_MAX. This meant that page checksum errors might be ignored for databases with a lot of past WAL in offline mode. Online mode is not affected since the upper bound is retrieved from pg_start_backup().
2019-07-11 15:13:56 +02:00
pg_data/base/32768/33000={"checksum":"7a16d165e4775f7c92e8cdf60c0af57313f0bf90","checksum-page":true,"master":false,"reference":"[BACKUP-FULL-2]","size":[SIZE],"timestamp":[TIMESTAMP-1]}
pg_data/base/32768/33000.32767={"checksum":"6e99b589e550e68e934fd235ccba59fe5b592a9e","checksum-page":true,"master":false,"reference":"[BACKUP-FULL-2]","size":[SIZE],"timestamp":[TIMESTAMP-1]}
Add checksum delta for incremental backups. Use checksums rather than timestamps to determine if files have changed. This is useful in cases where the timestamps may not be trustworthy, e.g. when performing an incremental after failing over to a standby. If checksum delta is enabled then checksums will be used for verification of resumed backups, even if they are full. Resumes have always used checksums to verify the files in the repository, enabling delta performs checksums on the database files as well. Note that the user must manually enable this feature in cases were it would be useful or just keep in enabled all the time. A future commit will address automatically enabling the feature in cases where it seems likely to be useful. Contributed by Cynthia Shang.
2018-09-19 17:12:45 +02:00
pg_data/base/32768/33001={"checksum":"6bf316f11d28c28914ea9be92c00de9bea6d9a6b","checksum-page":false,"checksum-page-error":[0,[3,5],7],"master":false,"reference":"[BACKUP-FULL-2]","size":[SIZE],"timestamp":[TIMESTAMP-1]}
Fix incorrect offline upper bound for ignoring page checksum errors. For offline backups the upper bound was being set to 0x0000FFFF0000FFFF rather than UINT64_MAX. This meant that page checksum errors might be ignored for databases with a lot of past WAL in offline mode. Online mode is not affected since the upper bound is retrieved from pg_start_backup().
2019-07-11 15:13:56 +02:00
pg_data/base/32768/44000_init={"checksum":"7a16d165e4775f7c92e8cdf60c0af57313f0bf90","checksum-page":true,"master":false,"reference":"[BACKUP-FULL-2]","size":[SIZE],"timestamp":[TIMESTAMP-1]}
Add checksum delta for incremental backups. Use checksums rather than timestamps to determine if files have changed. This is useful in cases where the timestamps may not be trustworthy, e.g. when performing an incremental after failing over to a standby. If checksum delta is enabled then checksums will be used for verification of resumed backups, even if they are full. Resumes have always used checksums to verify the files in the repository, enabling delta performs checksums on the database files as well. Note that the user must manually enable this feature in cases were it would be useful or just keep in enabled all the time. A future commit will address automatically enabling the feature in cases where it seems likely to be useful. Contributed by Cynthia Shang.
2018-09-19 17:12:45 +02:00
pg_data/base/32768/PG_VERSION={"checksum":"184473f470864e067ee3a22e64b47b0a1c356f29","master":false,"reference":"[BACKUP-FULL-2]","size":[SIZE],"timestamp":[TIMESTAMP-1]}
pg_data/changecontent.txt={"checksum":"238a131a3e8eb98d1fc5b27d882ca40b7618fd2a","reference":"[BACKUP-FULL-2]","size":[SIZE],"timestamp":[TIMESTAMP-1]}
pg_data/changesize.txt={"checksum":"88087292ed82e26f3eb824d0bffc05ccf7a30f8d","size":[SIZE],"timestamp":[TIMESTAMP-1]}
pg_data/changetime.txt={"checksum":"88087292ed82e26f3eb824d0bffc05ccf7a30f8d","reference":"[BACKUP-FULL-2]","size":[SIZE],"timestamp":[TIMESTAMP-1]}
Support configurable WAL segment size. PostgreSQL 11 introduces configurable WAL segment sizes, from 1MB to 1GB. There are two areas that needed to be updated to support this: building the archive-get queue and checking that WAL has been archived after a backup. Both operations require the WAL segment size to properly build a list. Checking the archive after a backup is still implemented in Perl and has an active database connection, so just get the WAL segment size from the database. The archive-get command does not have a connection to the database, so get the WAL segment size from pg_control instead. This requires a deeper inspection of pg_control than has been done in the past, so it seemed best to copy the relevant data structures from each version of PostgreSQL and build a generic interface layer to address them. While this approach is a bit verbose, it has the advantage of being relatively simple, and can easily be updated for new versions of PostgreSQL. Since the integration tests generate pg_control files for testing, teach Perl how to generate files with the correct offsets for both 32-bit and 64-bit architectures.
2018-09-25 11:24:42 +02:00
pg_data/global/pg_control={"checksum":"4c77c900f7af0d9ab13fa9982051a42e0b637f6c","reference":"[BACKUP-FULL-2]","size":[SIZE],"timestamp":[TIMESTAMP-2]}
Add checksum delta for incremental backups. Use checksums rather than timestamps to determine if files have changed. This is useful in cases where the timestamps may not be trustworthy, e.g. when performing an incremental after failing over to a standby. If checksum delta is enabled then checksums will be used for verification of resumed backups, even if they are full. Resumes have always used checksums to verify the files in the repository, enabling delta performs checksums on the database files as well. Note that the user must manually enable this feature in cases were it would be useful or just keep in enabled all the time. A future commit will address automatically enabling the feature in cases where it seems likely to be useful. Contributed by Cynthia Shang.
2018-09-19 17:12:45 +02:00
pg_data/pg_stat/global.stat={"checksum":"e350d5ce0153f3e22d5db21cf2a4eff00f3ee877","reference":"[BACKUP-FULL-2]","size":[SIZE],"timestamp":[TIMESTAMP-2]}
pg_data/postgresql.conf={"checksum":"6721d92c9fcdf4248acff1f9a1377127d9064807","reference":"[BACKUP-FULL-2]","size":[SIZE],"timestamp":[TIMESTAMP-2]}
Use minio for integration testing. ScalityS3 has not received any maintenance in years and is slow to start which is bad for testing. Replace it with minio which starts quickly and ships as a single executable or a tiny container. Minio has stricter limits on allowable characters but should still provide enough coverage to show that our encoding is working correctly. This commit also includes the upgrade to openssl 1.1.1 in the Ubuntu 18.04 container.
2019-07-03 04:20:35 +02:00
pg_data/special-!_.*'()&!@;:+,?={"reference":"[BACKUP-FULL-2]","size":[SIZE],"timestamp":[TIMESTAMP-1]}
Add checksum delta for incremental backups. Use checksums rather than timestamps to determine if files have changed. This is useful in cases where the timestamps may not be trustworthy, e.g. when performing an incremental after failing over to a standby. If checksum delta is enabled then checksums will be used for verification of resumed backups, even if they are full. Resumes have always used checksums to verify the files in the repository, enabling delta performs checksums on the database files as well. Note that the user must manually enable this feature in cases were it would be useful or just keep in enabled all the time. A future commit will address automatically enabling the feature in cases where it seems likely to be useful. Contributed by Cynthia Shang.
2018-09-19 17:12:45 +02:00
pg_data/zero_from_start={"reference":"[BACKUP-FULL-2]","size":[SIZE],"timestamp":[TIMESTAMP-1]}
pg_data/zerosize.txt={"size":[SIZE],"timestamp":[TIMESTAMP-2]}
pg_tblspc/1/[TS_PATH-1]/16384/tablespace1.txt={"checksum":"14c44cef6287269b08d41de489fd492bb9fc795d","checksum-page":false,"master":false,"size":[SIZE],"timestamp":[TIMESTAMP-2]}
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
[target:file:default]
Continuous integration using travis-ci.
2017-02-21 15:59:23 +02:00
group="[GROUP-1]"
Add checksum delta for incremental backups. Use checksums rather than timestamps to determine if files have changed. This is useful in cases where the timestamps may not be trustworthy, e.g. when performing an incremental after failing over to a standby. If checksum delta is enabled then checksums will be used for verification of resumed backups, even if they are full. Resumes have always used checksums to verify the files in the repository, enabling delta performs checksums on the database files as well. Note that the user must manually enable this feature in cases were it would be useful or just keep in enabled all the time. A future commit will address automatically enabling the feature in cases where it seems likely to be useful. Contributed by Cynthia Shang.
2018-09-19 17:12:45 +02:00
master=true
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
mode="0600"
Regression tests can now be run as any properly-configured user, not just vagrant.
2016-10-01 19:39:44 +02:00
user="[USER-1]"
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
[target:link]
Rename test paths for clarity. This was worked out as part of the test suite refactor [c8f806a] but not committed with it because of the large number of expect logs changes involved. Keeping them separate made it easier to audit the changes in the refactor.
2016-06-24 16:54:31 +02:00
pg_data/pg_tblspc/1={"destination":"[TEST_PATH]/db-master/db/tablespace/ts1"}
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
[target:link:default]
Continuous integration using travis-ci.
2017-02-21 15:59:23 +02:00
group="[GROUP-1]"
Regression tests can now be run as any properly-configured user, not just vagrant.
2016-10-01 19:39:44 +02:00
user="[USER-1]"
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
[target:path]
Fixed an issue with invalid backup groups being set correctly on restore. If the backup cannot map a group to a name it stores the group in the manifest as false then uses either the owner of $PGDATA to set the group during restore or failing that the group of the current user. This logic was not working correctly because the selected group was overwriting the user on restore leaving the group undefined and the user incorrectly set to the group. (Reported by Jeff McCormick.)
2017-11-20 19:46:35 +02:00
pg_data={"mode":"0777"}
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
pg_data/base={}
Allow selective restore of databases from a cluster backup. This feature can result in major space and time savings when only specific databases are restored. Unrestored databases will not be accessible but must be manually dropped before they will be removed from the shared catalogue.
2016-05-11 15:21:39 +02:00
pg_data/base/1={}
pg_data/base/16384={}
pg_data/base/32768={}
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
pg_data/global={}
pg_data/pg_clog={}
Exclude directories during backup that are cleaned, recreated, or zeroed by PostgreSQL at startup. These include (depending on the version where they were introduced): pgsql_tmp, pg_dynshmem, pg_notify, pg_replslot, pg_serial, pg_snapshots, pg_stat_tmp, pg_subtrans. The postgresql.auto.conf.tmp file is now excluded in addition to files that were already excluded: backup_label.old, postmaster.opts, postmaster.pid, recovery.conf, recovery.done.
2016-08-16 15:35:16 +02:00
pg_data/pg_dynshmem={}
Allow arbitrary directories and/or files to be excluded from a backup. Misuse of this feature can lead to inconsistent backups so read the --exclude documentation carefully before using.
2018-08-27 21:51:05 +02:00
pg_data/pg_log={}
Exclude directories during backup that are cleaned, recreated, or zeroed by PostgreSQL at startup. These include (depending on the version where they were introduced): pgsql_tmp, pg_dynshmem, pg_notify, pg_replslot, pg_serial, pg_snapshots, pg_stat_tmp, pg_subtrans. The postgresql.auto.conf.tmp file is now excluded in addition to files that were already excluded: backup_label.old, postmaster.opts, postmaster.pid, recovery.conf, recovery.done.
2016-08-16 15:35:16 +02:00
pg_data/pg_notify={}
pg_data/pg_replslot={}
pg_data/pg_serial={}
pg_data/pg_snapshots={}
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
pg_data/pg_stat={}
Exclude directories during backup that are cleaned, recreated, or zeroed by PostgreSQL at startup. These include (depending on the version where they were introduced): pgsql_tmp, pg_dynshmem, pg_notify, pg_replslot, pg_serial, pg_snapshots, pg_stat_tmp, pg_subtrans. The postgresql.auto.conf.tmp file is now excluded in addition to files that were already excluded: backup_label.old, postmaster.opts, postmaster.pid, recovery.conf, recovery.done.
2016-08-16 15:35:16 +02:00
pg_data/pg_stat_tmp={}
pg_data/pg_subtrans={}
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
pg_data/pg_tblspc={}
Fixed an issue with invalid backup groups being set correctly on restore. If the backup cannot map a group to a name it stores the group in the manifest as false then uses either the owner of $PGDATA to set the group during restore or failing that the group of the current user. This logic was not working correctly because the selected group was overwriting the user on restore leaving the group undefined and the user incorrectly set to the group. (Reported by Jeff McCormick.)
2017-11-20 19:46:35 +02:00
pg_tblspc={"mode":"0777"}
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
pg_tblspc/1={}
pg_tblspc/1/[TS_PATH-1]={}
Allow selective restore of databases from a cluster backup. This feature can result in major space and time savings when only specific databases are restored. Unrestored databases will not be accessible but must be manually dropped before they will be removed from the shared catalogue.
2016-05-11 15:21:39 +02:00
pg_tblspc/1/[TS_PATH-1]/16384={}
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
[target:path:default]
Continuous integration using travis-ci.
2017-02-21 15:59:23 +02:00
group="[GROUP-1]"
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
mode="0700"
Regression tests can now be run as any properly-configured user, not just vagrant.
2016-10-01 19:39:44 +02:00
user="[USER-1]"
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
Move info file checksum to the end of the file. Putting the checksum at the beginning of the file made it impossible to stream the file out when saving. The entire file had to be held in memory while it was checksummed so the checksum could be written at the beginning. Instead place the checksum at the end. This does not break the existing Perl or C code since the read is not order dependent. There are no plans to improve the Perl code to take advantage of this change, but it will make the C implementation more efficient. Reviewed by Cynthia Shang.
2019-08-22 01:45:48 +02:00
[backrest]
backrest-checksum="[CHECKSUM]"
Rename test paths for clarity. This was worked out as part of the test suite refactor [c8f806a] but not committed with it because of the large number of expect logs changes involved. Keeping them separate made it easier to audit the changes in the refactor.
2016-06-24 16:54:31 +02:00
+ supplemental file: [TEST_PATH]/db-master/repo/backup/db/backup.info
---------------------------------------------------------------------
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
[backrest]
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
backrest-format=5
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
backrest-version="[VERSION-1]"
[backup:current]
Set archive-check option in manifest correctly when offline. Archive check does not run when in offline backup mode but the option was set to true in the manifest. It's harmless since these options are informational only but it could cause confusion when debugging.
2019-11-28 15:27:21 +02:00
[BACKUP-FULL-2]={"backrest-format":5,"backrest-version":"[VERSION-1]","backup-archive-start":null,"backup-archive-stop":null,"backup-info-repo-size":[SIZE],"backup-info-repo-size-delta":[DELTA],"backup-info-size":[SIZE],"backup-info-size-delta":[DELTA],"backup-timestamp-start":[TIMESTAMP],"backup-timestamp-stop":[TIMESTAMP],"backup-type":"full","db-id":1,"option-archive-check":false,"option-archive-copy":true,"option-backup-standby":false,"option-checksum-page":true,"option-compress":false,"option-hardlink":false,"option-online":false}
[BACKUP-INCR-1]={"backrest-format":5,"backrest-version":"[VERSION-1]","backup-archive-start":null,"backup-archive-stop":null,"backup-info-repo-size":[SIZE],"backup-info-repo-size-delta":[DELTA],"backup-info-size":[SIZE],"backup-info-size-delta":[DELTA],"backup-prior":"[BACKUP-FULL-2]","backup-reference":["[BACKUP-FULL-2]"],"backup-timestamp-start":[TIMESTAMP],"backup-timestamp-stop":[TIMESTAMP],"backup-type":"incr","db-id":1,"option-archive-check":false,"option-archive-copy":true,"option-backup-standby":false,"option-checksum-page":true,"option-compress":false,"option-hardlink":false,"option-online":false}
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
[db]
Database version refactoring. * Refactor db version constants into a separate module. * Update synthetic backup tests to PostgreSQL 9.4.
2016-08-12 04:35:24 +02:00
db-catalog-version=201409291
db-control-version=942
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
db-id=1
Automate generation of WAL and pg_control test files. The existing static files would not work with 32-bit or big-endian systems so create functions to generate these files dynamically rather than creating a bunch of new static files.
2017-11-19 03:02:54 +02:00
db-system-id=1000000000000000094
Database version refactoring. * Refactor db version constants into a separate module. * Update synthetic backup tests to PostgreSQL 9.4.
2016-08-12 04:35:24 +02:00
db-version="9.4"
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
[db:history]
Automate generation of WAL and pg_control test files. The existing static files would not work with 32-bit or big-endian systems so create functions to generate these files dynamically rather than creating a bunch of new static files.
2017-11-19 03:02:54 +02:00
1={"db-catalog-version":201409291,"db-control-version":942,"db-system-id":1000000000000000094,"db-version":"9.4"}
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
Improve performance of info file load/save. Info files required three copies in memory to be loaded (the original string, an ini representation, and the final info object). Not only was this memory inefficient but the Ini object does sequential scans when searching for keys making large files very slow to load. This has not been an issue since archive.info and backup.info are very small, but it becomes a big deal when loading manifests with hundreds of thousands of files. Instead of holding copies of the data in memory, use a callback to deliver the ini data directly to the object when loading. Use a similar method for save to avoid having an intermediate copy. Save is a bit complex because sections/keys must be written in alpha order or older versions of pgBackRest will not calculate the correct checksum. Also move the load retry logic to helper functions rather than embedding it in the Info object. This allows for more flexibility in loading and ensures that stack traces will be available when developing unit tests. Reviewed by Cynthia Shang.
2019-09-06 19:48:28 +02:00
[backrest]
backrest-checksum="[CHECKSUM]"
Major refactor of the test suite. * Make the code more modular and object-oriented. * Multiple Docker containers can now be created for a single test to simulate more realistic environments.
2016-06-24 14:12:58 +02:00
incr backup - resume and add tablespace 2 (db-master host)
Remove obsolete integration tests from mock/all. The protocol timeout tests have been superceded by unit tests. The TEST_BACKUP_RESUME test point was incorrectly included into a number of tests, probably a copy pasto. It didn't hurt anything but it did add 200ms to each test where it appeared. Catalog and control version tests were redundant. The database version and system id tests covered the important code paths and the C code gets these values from a lookup table. Finally, fix an incomplete update to the backup.info file while munging for tests.
2019-11-21 23:06:27 +02:00
> [CONTAINER-EXEC] db-master [BACKREST-BIN] --config=[TEST_PATH]/db-master/pgbackrest.conf --no-online --process-max=1 --delta --stanza=db backup
Log testing can now be enabled for certain deterministic tests. This works by comparing the generated logs against a previous copy. Currently only enabled for the backup/synthetic tests.
2015-05-05 19:08:48 +02:00
------------------------------------------------------------------------------------------------------------------------------------
Add infrastructure for multiple compression type support. Add compress-type option and deprecate compress option. Since the compress option is boolean it won't work with multiple compression types. Add logic to cfgLoadUpdateOption() to update compress-type if it is not set directly. The compress option should no longer be referenced outside the cfgLoadUpdateOption() function. Add common/compress/helper module to contain interface functions that work with multiple compression types. Code outside this module should no longer call specific compression drivers, though it may be OK to reference a specific compression type using the new interface (e.g., saving backup history files in gz format). Unit tests only test compression using the gz format because other formats may not be available in all builds. It is the job of integration tests to exercise all compression types. Additional compression types will be added in future commits.
2020-03-06 21:41:03 +02:00
P00 INFO: backup command begin [BACKREST-VERSION]: --buffer-size=[BUFFER-SIZE] --compress-level=3 --compress-type=none --config=[TEST_PATH]/db-master/pgbackrest.conf --db-timeout=45 --delta --lock-path=[TEST_PATH]/db-master/lock --log-level-console=detail --log-level-file=[LOG-LEVEL-FILE] --log-level-stderr=off --log-path=[TEST_PATH]/db-master/log[] --no-log-timestamp --no-online --pg1-path=[TEST_PATH]/db-master/db/base --process-max=1 --protocol-timeout=60 --repo1-path=[TEST_PATH]/db-master/repo --stanza=db --start-fast
Add time-based retention for full backups. The --repo-retention-full-type option allows retention of full backups based on a time period, specified in days. The new option will default to 'count' and therefore will not affect current installations. Setting repo-retention-full-type to 'time' will allow the user to use a time period, in days, to indicate full backup retention. Using this method, a full backup can be expired only if the time the backup completed is older than the number of days set with repo-retention-full (calculated from the moment the 'expire' command is run) and at least one full backup meets the retention period. If archive retention has not been configured, then the default settings will expire archives that are prior to the oldest retained full backup. For example, if there are three full backups ending in times that are 25 days old (F1), 20 days old (F2) and 10 days old (F3), then if the full retention period is 15 days, then only F1 will be expired; F2 will be retained because F1 is not at least 15 days old.
2020-05-08 21:25:03 +02:00
P00 WARN: option 'repo1-retention-full' is not set for 'repo1-retention-full-type=count', the repository may run out of space
Rename retention-* options to repo-retention-*.
2018-02-19 22:07:24 +02:00
HINT: to retain full backups indefinitely (without warning), set option 'repo1-retention-full' to the maximum.
Update missing manifest warning in BackupInfo. This brings the Perl message in line with C to reduce expect log churn.
2019-11-25 15:51:28 +02:00
P00 WARN: backup '[BACKUP-INCR-1]' missing manifest removed from backup.info
Improvements and additional tests for stanza-upgrade. Contributed by Cynthia Shang.
2017-04-07 01:36:37 +02:00
P00 INFO: last backup label = [BACKUP-FULL-2], version = [VERSION-1]
Added the --checksum-page option. This option allows pgBackRest to validate page checksums in data files when checksums are enabled on PostgreSQL >= 9.3. Note that this functionality requires a C library which may not initially be available in OS packages. The option will automatically be enabled when the library is present and checksums are enabled on the cluster.
2016-12-13 01:54:07 +02:00
P00 WARN: incr backup cannot alter 'checksum-page' option to 'false', reset to 'true' from [BACKUP-FULL-2]
The backup command is implemented entirely in C. For the most part this is a direct migration of the Perl code into C except as noted below. A backup can now be initiated from a linked directory. The link will not be stored in the manifest or recreated on restore. If a link or directory does not already exist in the restore location then a directory will be created. The logic for creating backup labels has been improved and it should no longer be possible to get a backup label earlier than the latest backup even with timezone changes or clock skew. This has never been an issue in the field that we know of, but we found it in testing. For online backups all times are fetched from the PostgreSQL primary host (before only copy start was). This doesn't affect backup integrity but it does prevent clock skew between hosts affecting backup duration reporting. Archive copy now works as expected when the archive and backup have different compression settings, i.e. when one is compressed and the other is not. This was a long-standing bug in the Perl code. Resume will now work even if hardlink settings have been changed. Reviewed by Cynthia Shang.
2019-12-14 00:14:26 +02:00
P00 WARN: resumable backup [BACKUP-INCR-2] of same type exists -- remove invalid files and resume
P00 DETAIL: remove file '[TEST_PATH]/db-master/repo/backup/db/[BACKUP-INCR-2]/pg_data/changesize.txt' from resumed backup (mismatched size)
P00 DETAIL: remove file '[TEST_PATH]/db-master/repo/backup/db/[BACKUP-INCR-2]/pg_data/zerosize.txt' from resumed backup (zero size)
Add checksum delta for incremental backups. Use checksums rather than timestamps to determine if files have changed. This is useful in cases where the timestamps may not be trustworthy, e.g. when performing an incremental after failing over to a standby. If checksum delta is enabled then checksums will be used for verification of resumed backups, even if they are full. Resumes have always used checksums to verify the files in the repository, enabling delta performs checksums on the database files as well. Note that the user must manually enable this feature in cases were it would be useful or just keep in enabled all the time. A future commit will address automatically enabling the feature in cases where it seems likely to be useful. Contributed by Cynthia Shang.
2018-09-19 17:12:45 +02:00
P01 DETAIL: match file from prior backup [TEST_PATH]/db-master/db/base/base/32768/33001 (64KB, 33%) checksum 6bf316f11d28c28914ea9be92c00de9bea6d9a6b
Fix incorrect offline upper bound for ignoring page checksum errors. For offline backups the upper bound was being set to 0x0000FFFF0000FFFF rather than UINT64_MAX. This meant that page checksum errors might be ignored for databases with a lot of past WAL in offline mode. Online mode is not affected since the upper bound is retrieved from pg_start_backup().
2019-07-11 15:13:56 +02:00
P01 DETAIL: match file from prior backup [TEST_PATH]/db-master/db/base/base/32768/44000_init (32KB, 49%) checksum 7a16d165e4775f7c92e8cdf60c0af57313f0bf90
P01 DETAIL: match file from prior backup [TEST_PATH]/db-master/db/base/base/32768/33000.32767 (32KB, 66%) checksum 6e99b589e550e68e934fd235ccba59fe5b592a9e
P01 DETAIL: match file from prior backup [TEST_PATH]/db-master/db/base/base/32768/33000 (32KB, 83%) checksum 7a16d165e4775f7c92e8cdf60c0af57313f0bf90
Add checksum delta for incremental backups. Use checksums rather than timestamps to determine if files have changed. This is useful in cases where the timestamps may not be trustworthy, e.g. when performing an incremental after failing over to a standby. If checksum delta is enabled then checksums will be used for verification of resumed backups, even if they are full. Resumes have always used checksums to verify the files in the repository, enabling delta performs checksums on the database files as well. Note that the user must manually enable this feature in cases were it would be useful or just keep in enabled all the time. A future commit will address automatically enabling the feature in cases where it seems likely to be useful. Contributed by Cynthia Shang.
2018-09-19 17:12:45 +02:00
P01 DETAIL: match file from prior backup [TEST_PATH]/db-master/db/base/base/16384/17000 (16KB, 91%) checksum e0101dd8ffb910c9c202ca35b5f828bcb9697bed
Support configurable WAL segment size. PostgreSQL 11 introduces configurable WAL segment sizes, from 1MB to 1GB. There are two areas that needed to be updated to support this: building the archive-get queue and checking that WAL has been archived after a backup. Both operations require the WAL segment size to properly build a list. Checking the archive after a backup is still implemented in Perl and has an active database connection, so just get the WAL segment size from the database. The archive-get command does not have a connection to the database, so get the WAL segment size from pg_control instead. This requires a deeper inspection of pg_control than has been done in the past, so it seemed best to copy the relevant data structures from each version of PostgreSQL and build a generic interface layer to address them. While this approach is a bit verbose, it has the advantage of being relatively simple, and can easily be updated for new versions of PostgreSQL. Since the integration tests generate pg_control files for testing, teach Perl how to generate files with the correct offsets for both 32-bit and 64-bit architectures.
2018-09-25 11:24:42 +02:00
P01 DETAIL: match file from prior backup [TEST_PATH]/db-master/db/base/global/pg_control (8KB, 95%) checksum 4c77c900f7af0d9ab13fa9982051a42e0b637f6c
Add checksum delta for incremental backups. Use checksums rather than timestamps to determine if files have changed. This is useful in cases where the timestamps may not be trustworthy, e.g. when performing an incremental after failing over to a standby. If checksum delta is enabled then checksums will be used for verification of resumed backups, even if they are full. Resumes have always used checksums to verify the files in the repository, enabling delta performs checksums on the database files as well. Note that the user must manually enable this feature in cases were it would be useful or just keep in enabled all the time. A future commit will address automatically enabling the feature in cases where it seems likely to be useful. Contributed by Cynthia Shang.
2018-09-19 17:12:45 +02:00
P01 DETAIL: match file from prior backup [TEST_PATH]/db-master/db/base/base/1/12000 (8KB, 99%) checksum 22c98d248ff548311eda88559e4a8405ed77c003
P01 DETAIL: match file from prior backup [TEST_PATH]/db-master/db/base/postgresql.conf (21B, 99%) checksum 6721d92c9fcdf4248acff1f9a1377127d9064807
Fix critical bug in resume that resulted in inconsistent backups. A regression in v0.82 removed the timestamp comparison when deciding which files from the aborted backup to keep on resume. All resumed backups should be considered inconsistent. A resumed backup can be identified by checking the log for the message "aborted backup of same type exists, will be cleaned to remove invalid files and resumed". Reported by David Youatt, Yogesh Sharma, Stephen Frost.
2018-07-03 20:01:57 +02:00
P00 WARN: resumed backup file pg_data/badchecksum.txt does not have expected checksum f927212cd08d11a42a666b2f04235398e9ceeb51. The file will be recopied and backup will continue but this may be an issue unless the resumed backup path in the repository is known to be corrupted.
NOTE: this does not indicate a problem with the PostgreSQL page checksums.
Add checksum delta for incremental backups. Use checksums rather than timestamps to determine if files have changed. This is useful in cases where the timestamps may not be trustworthy, e.g. when performing an incremental after failing over to a standby. If checksum delta is enabled then checksums will be used for verification of resumed backups, even if they are full. Resumes have always used checksums to verify the files in the repository, enabling delta performs checksums on the database files as well. Note that the user must manually enable this feature in cases were it would be useful or just keep in enabled all the time. A future commit will address automatically enabling the feature in cases where it seems likely to be useful. Contributed by Cynthia Shang.
2018-09-19 17:12:45 +02:00
P01 INFO: backup file [TEST_PATH]/db-master/db/base/badchecksum.txt (11B, 99%) checksum f927212cd08d11a42a666b2f04235398e9ceeb51
P01 INFO: backup file [TEST_PATH]/db-master/db/base/changesize.txt (9B, 99%) checksum 3905d5be2ec8d67f41435dab5e0dcda3ae47455d
P01 INFO: backup file [TEST_PATH]/db-master/db/base/changecontent.txt (7B, 99%) checksum a094d94583e209556d03c3c5da33131a065f1689
P01 DETAIL: match file from prior backup [TEST_PATH]/db-master/db/base/pg_stat/global.stat (5B, 99%) checksum e350d5ce0153f3e22d5db21cf2a4eff00f3ee877
P01 DETAIL: match file from prior backup [TEST_PATH]/db-master/db/base/changetime.txt (4B, 99%) checksum 88087292ed82e26f3eb824d0bffc05ccf7a30f8d
P01 DETAIL: match file from prior backup [TEST_PATH]/db-master/db/base/base/32768/PG_VERSION (3B, 99%) checksum 184473f470864e067ee3a22e64b47b0a1c356f29
P01 DETAIL: match file from prior backup [TEST_PATH]/db-master/db/base/base/16384/PG_VERSION (3B, 99%) checksum 184473f470864e067ee3a22e64b47b0a1c356f29
P01 DETAIL: match file from prior backup [TEST_PATH]/db-master/db/base/base/1/PG_VERSION (3B, 99%) checksum 184473f470864e067ee3a22e64b47b0a1c356f29
P01 DETAIL: match file from prior backup [TEST_PATH]/db-master/db/base/PG_VERSION (3B, 99%) checksum 184473f470864e067ee3a22e64b47b0a1c356f29
P01 INFO: backup file [TEST_PATH]/db-master/db/base/zerosize.txt (0B, 99%)
The backup command is implemented entirely in C. For the most part this is a direct migration of the Perl code into C except as noted below. A backup can now be initiated from a linked directory. The link will not be stored in the manifest or recreated on restore. If a link or directory does not already exist in the restore location then a directory will be created. The logic for creating backup labels has been improved and it should no longer be possible to get a backup label earlier than the latest backup even with timezone changes or clock skew. This has never been an issue in the field that we know of, but we found it in testing. For online backups all times are fetched from the PostgreSQL primary host (before only copy start was). This doesn't affect backup integrity but it does prevent clock skew between hosts affecting backup duration reporting. Archive copy now works as expected when the archive and backup have different compression settings, i.e. when one is compressed and the other is not. This was a long-standing bug in the Perl code. Resume will now work even if hardlink settings have been changed. Reviewed by Cynthia Shang.
2019-12-14 00:14:26 +02:00
P01 INFO: backup file [TEST_PATH]/db-master/db/base/pg_tblspc/1/[TS_PATH-1]/16384/tablespace1.txt (7B, 99%) checksum d85de07d6421d90aa9191c11c889bfde43680f0f
Fix critical bug in resume that resulted in inconsistent backups. A regression in v0.82 removed the timestamp comparison when deciding which files from the aborted backup to keep on resume. All resumed backups should be considered inconsistent. A resumed backup can be identified by checking the log for the message "aborted backup of same type exists, will be cleaned to remove invalid files and resumed". Reported by David Youatt, Yogesh Sharma, Stephen Frost.
2018-07-03 20:01:57 +02:00
P00 WARN: page misalignment in file [TEST_PATH]/db-master/db/base/pg_tblspc/1/[TS_PATH-1]/16384/tablespace1.txt: file size 7 is not divisible by page size 8192
The backup command is implemented entirely in C. For the most part this is a direct migration of the Perl code into C except as noted below. A backup can now be initiated from a linked directory. The link will not be stored in the manifest or recreated on restore. If a link or directory does not already exist in the restore location then a directory will be created. The logic for creating backup labels has been improved and it should no longer be possible to get a backup label earlier than the latest backup even with timezone changes or clock skew. This has never been an issue in the field that we know of, but we found it in testing. For online backups all times are fetched from the PostgreSQL primary host (before only copy start was). This doesn't affect backup integrity but it does prevent clock skew between hosts affecting backup duration reporting. Archive copy now works as expected when the archive and backup have different compression settings, i.e. when one is compressed and the other is not. This was a long-standing bug in the Perl code. Resume will now work even if hardlink settings have been changed. Reviewed by Cynthia Shang.
2019-12-14 00:14:26 +02:00
P01 INFO: backup file [TEST_PATH]/db-master/db/base/pg_tblspc/2/[TS_PATH-1]/32768/tablespace2.txt (7B, 100%) checksum dc7f76e43c46101b47acc55ae4d593a9e6983578
P00 WARN: page misalignment in file [TEST_PATH]/db-master/db/base/pg_tblspc/2/[TS_PATH-1]/32768/tablespace2.txt: file size 7 is not divisible by page size 8192
P00 DETAIL: reference pg_data/PG_VERSION to [BACKUP-FULL-2]
P00 DETAIL: reference pg_data/base/1/12000 to [BACKUP-FULL-2]
P00 DETAIL: reference pg_data/base/1/PG_VERSION to [BACKUP-FULL-2]
P00 DETAIL: reference pg_data/base/16384/17000 to [BACKUP-FULL-2]
P00 DETAIL: reference pg_data/base/16384/PG_VERSION to [BACKUP-FULL-2]
P00 DETAIL: reference pg_data/base/32768/33000 to [BACKUP-FULL-2]
P00 DETAIL: reference pg_data/base/32768/33000.32767 to [BACKUP-FULL-2]
P00 DETAIL: reference pg_data/base/32768/33001 to [BACKUP-FULL-2]
P00 DETAIL: reference pg_data/base/32768/44000_init to [BACKUP-FULL-2]
P00 DETAIL: reference pg_data/base/32768/PG_VERSION to [BACKUP-FULL-2]
P00 DETAIL: reference pg_data/changetime.txt to [BACKUP-FULL-2]
P00 DETAIL: reference pg_data/global/pg_control to [BACKUP-FULL-2]
P00 DETAIL: reference pg_data/pg_stat/global.stat to [BACKUP-FULL-2]
P00 DETAIL: reference pg_data/postgresql.conf to [BACKUP-FULL-2]
P00 DETAIL: reference pg_data/special-!_.*'()&!@;:+,? to [BACKUP-FULL-2]
P00 DETAIL: reference pg_data/zero_from_start to [BACKUP-FULL-2]
Add checksum delta for incremental backups. Use checksums rather than timestamps to determine if files have changed. This is useful in cases where the timestamps may not be trustworthy, e.g. when performing an incremental after failing over to a standby. If checksum delta is enabled then checksums will be used for verification of resumed backups, even if they are full. Resumes have always used checksums to verify the files in the repository, enabling delta performs checksums on the database files as well. Note that the user must manually enable this feature in cases were it would be useful or just keep in enabled all the time. A future commit will address automatically enabling the feature in cases where it seems likely to be useful. Contributed by Cynthia Shang.
2018-09-19 17:12:45 +02:00
P00 INFO: incr backup size = 192KB
Show Process ID in log output instead of filtering it out with the timestamp.
2016-09-14 23:30:57 +02:00
P00 INFO: new backup label = [BACKUP-INCR-2]
Added a predictable log entry to signal that a command has completed successfully. For example a backup ends successfully with: INFO: backup command end: completed successfully.
2016-12-04 21:44:53 +02:00
P00 INFO: backup command end: completed successfully
Integration test improvements for disk and memory efficiency. Set log-level-file=off when more that one test will run. In this case is it impossible to see the logs anyway since they will be automatically cleaned up after the test. This improves performance pretty dramatically since trace-level logging is expensive. If a singe integration test is run then log-level-file is trace by default but can be changed with the --log-level-test-file option. Reduce buffer-size to 64k to save memory during testing and allow more processes to run in parallel. Update log replacement rules so that these options can change without affecting expect logs.
2019-12-17 22:23:07 +02:00
P00 INFO: expire command begin [BACKREST-VERSION]: --buffer-size=[BUFFER-SIZE] --config=[TEST_PATH]/db-master/pgbackrest.conf --lock-path=[TEST_PATH]/db-master/lock --log-level-console=detail --log-level-file=[LOG-LEVEL-FILE] --log-level-stderr=off --log-path=[TEST_PATH]/db-master/log[] --no-log-timestamp --repo1-path=[TEST_PATH]/db-master/repo --stanza=db
Rename retention-* options to repo-retention-*.
2018-02-19 22:07:24 +02:00
P00 INFO: option 'repo1-retention-archive' is not set - archive logs will not be expired
Added a predictable log entry to signal that a command has completed successfully. For example a backup ends successfully with: INFO: backup command end: completed successfully.
2016-12-04 21:44:53 +02:00
P00 INFO: expire command end: completed successfully
Log testing can now be enabled for certain deterministic tests. This works by comparing the generated logs against a previous copy. Currently only enabled for the backup/synthetic tests.
2015-05-05 19:08:48 +02:00
Rename test paths for clarity. This was worked out as part of the test suite refactor [c8f806a] but not committed with it because of the large number of expect logs changes involved. Keeping them separate made it easier to audit the changes in the refactor.
2016-06-24 16:54:31 +02:00
+ supplemental file: [TEST_PATH]/db-master/pgbackrest.conf
----------------------------------------------------------
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
[db]
Rename db-* options to pg-* and backup-* options to repo-* to improve consistency. * repo-* options are now indexed although only one is allowed. * List deprecated option names in documentation and command-line help.
2018-02-04 01:27:38 +02:00
pg1-path=[TEST_PATH]/db-master/db/base
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
New simpler configuration and consistent project/exe/path naming. * The repo-path option now always refers to the repository where backups and archive are stored, whether local or remote, so the repo-remote-path option has been removed. The new spool-path option can be used to define a location for queueing WAL segments when archiving asynchronously. Otherwise, a local repository is no longer required. * Implemented a new config format which should be far simpler to use. See the User Guide and Configuration Reference for details but for a simple configuration all options can now be placed in the stanza section. Options that are shared between stanzas can be placed in the [global] section. More complex configurations can still make use of command sections though this should be a rare use case. * The default configuration filename is now pgbackrest.conf instead of pg_backrest.conf. This was done for consistency with other naming changes but also to prevent old config files from being loaded accidentally. * The default repository name was changed from /var/lib/backup to /var/lib/pgbackrest. * Lock files are now stored in /tmp/pgbackrest by default. These days /run/pgbackrest would be the preferred location but that would require init scripts which are not part of this release. The lock-path option can be used to configure the lock directory. * Log files are now stored in /var/log/pgbackrest by default and no longer have the date appended so they can be managed with logrotate. The log-path option can be used to configure the lock directory. * Executable filename changed from pg_backrest to pgbackrest.
2016-04-14 15:30:54 +02:00
[global]
Integration test improvements for disk and memory efficiency. Set log-level-file=off when more that one test will run. In this case is it impossible to see the logs anyway since they will be automatically cleaned up after the test. This improves performance pretty dramatically since trace-level logging is expensive. If a singe integration test is run then log-level-file is trace by default but can be changed with the --log-level-test-file option. Reduce buffer-size to 64k to save memory during testing and allow more processes to run in parallel. Update log replacement rules so that these options can change without affecting expect logs.
2019-12-17 22:23:07 +02:00
buffer-size=[BUFFER-SIZE]
Fix custom settings for compress-level option being ignored. Reported by Jens Wilke.
2017-10-23 00:05:46 +02:00
compress-level=3
Add infrastructure for multiple compression type support. Add compress-type option and deprecate compress option. Since the compress option is boolean it won't work with multiple compression types. Add logic to cfgLoadUpdateOption() to update compress-type if it is not set directly. The compress option should no longer be referenced outside the cfgLoadUpdateOption() function. Add common/compress/helper module to contain interface functions that work with multiple compression types. Code outside this module should no longer call specific compression drivers, though it may be OK to reference a specific compression type using the new interface (e.g., saving backup history files in gz format). Unit tests only test compression using the gz format because other formats may not be available in all builds. It is the job of integration tests to exercise all compression types. Additional compression types will be added in future commits.
2020-03-06 21:41:03 +02:00
compress-type=none
Storage and IO layer refactor: Refactor storage layer to allow for new repository filesystems using drivers. (Reviewed by Cynthia Shang.) Refactor IO layer to allow for new compression formats, checksum types, and other capabilities using filters. (Reviewed by Cynthia Shang.)
2017-06-09 23:51:41 +02:00
db-timeout=45
lock-path=[TEST_PATH]/db-master/lock
Reduce log level for all expect tests to detail. The C code is designed to be efficient rather than deterministic at the debug log level. As we move more testing from integration to unit tests it makes less sense to try and maintain the expect logs at this log level. Most of the expect logs have already been moved to detail level but mock/all still had tests at debug level. Change the logging defaults in the config file and remove as many references to log-level-console as possible.
2019-05-23 00:23:44 +02:00
log-level-console=detail
Integration test improvements for disk and memory efficiency. Set log-level-file=off when more that one test will run. In this case is it impossible to see the logs anyway since they will be automatically cleaned up after the test. This improves performance pretty dramatically since trace-level logging is expensive. If a singe integration test is run then log-level-file is trace by default but can be changed with the --log-level-test-file option. Reduce buffer-size to 64k to save memory during testing and allow more processes to run in parallel. Update log replacement rules so that these options can change without affecting expect logs.
2019-12-17 22:23:07 +02:00
log-level-file=[LOG-LEVEL-FILE]
Added the log-level-stderr option. Controls whether console log messages are sent to stderr or stdout. By default this is set to warn which represents a change in behavior from previous versions, even though it may be more intuitive. Setting log-level-stderr=off will preserve the old behavior. Suggested by Sascha Biberhofer.
2016-10-05 15:09:30 +02:00
log-level-stderr=off
Storage and IO layer refactor: Refactor storage layer to allow for new repository filesystems using drivers. (Reviewed by Cynthia Shang.) Refactor IO layer to allow for new compression formats, checksum types, and other capabilities using filters. (Reviewed by Cynthia Shang.)
2017-06-09 23:51:41 +02:00
log-path=[TEST_PATH]/db-master/log
Integration test improvements for disk and memory efficiency. Set log-level-file=off when more that one test will run. In this case is it impossible to see the logs anyway since they will be automatically cleaned up after the test. This improves performance pretty dramatically since trace-level logging is expensive. If a singe integration test is run then log-level-file is trace by default but can be changed with the --log-level-test-file option. Reduce buffer-size to 64k to save memory during testing and allow more processes to run in parallel. Update log replacement rules so that these options can change without affecting expect logs.
2019-12-17 22:23:07 +02:00
log-subprocess=[LOG-SUBPROCESS]
Set log-timestamp=n for integration tests. This means less filtering of logs needs to be done and new timestamps can be added without adding new filters.
2018-05-11 17:24:38 +02:00
log-timestamp=n
Storage and IO layer refactor: Refactor storage layer to allow for new repository filesystems using drivers. (Reviewed by Cynthia Shang.) Refactor IO layer to allow for new compression formats, checksum types, and other capabilities using filters. (Reviewed by Cynthia Shang.)
2017-06-09 23:51:41 +02:00
protocol-timeout=60
Rename db-* options to pg-* and backup-* options to repo-* to improve consistency. * repo-* options are now indexed although only one is allowed. * List deprecated option names in documentation and command-line help.
2018-02-04 01:27:38 +02:00
repo1-path=[TEST_PATH]/db-master/repo
Consolidate archive-push tests into new archive test.
2017-06-21 14:38:53 +02:00
spool-path=[TEST_PATH]/db-master/spool
New simpler configuration and consistent project/exe/path naming. * The repo-path option now always refers to the repository where backups and archive are stored, whether local or remote, so the repo-remote-path option has been removed. The new spool-path option can be used to define a location for queueing WAL segments when archiving asynchronously. Otherwise, a local repository is no longer required. * Implemented a new config format which should be far simpler to use. See the User Guide and Configuration Reference for details but for a simple configuration all options can now be placed in the stanza section. Options that are shared between stanzas can be placed in the [global] section. More complex configurations can still make use of command sections though this should be a rare use case. * The default configuration filename is now pgbackrest.conf instead of pg_backrest.conf. This was done for consistency with other naming changes but also to prevent old config files from being loaded accidentally. * The default repository name was changed from /var/lib/backup to /var/lib/pgbackrest. * Lock files are now stored in /tmp/pgbackrest by default. These days /run/pgbackrest would be the preferred location but that would require init scripts which are not part of this release. The lock-path option can be used to configure the lock directory. * Log files are now stored in /var/log/pgbackrest by default and no longer have the date appended so they can be managed with logrotate. The log-path option can be used to configure the lock directory. * Executable filename changed from pg_backrest to pgbackrest.
2016-04-14 15:30:54 +02:00
[global:backup]
archive-copy=y
start-fast=y
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
Rename test paths for clarity. This was worked out as part of the test suite refactor [c8f806a] but not committed with it because of the large number of expect logs changes involved. Keeping them separate made it easier to audit the changes in the refactor.
2016-06-24 16:54:31 +02:00
+ supplemental file: [TEST_PATH]/db-master/repo/backup/db/[BACKUP-INCR-2]/backup.manifest
-----------------------------------------------------------------------------------------
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
[backrest]
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
backrest-format=5
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
backrest-version="[VERSION-1]"
[backup]
backup-label="[BACKUP-INCR-2]"
backup-prior="[BACKUP-FULL-2]"
backup-timestamp-copy-start=[TIMESTAMP]
backup-timestamp-start=[TIMESTAMP]
backup-timestamp-stop=[TIMESTAMP]
backup-type="incr"
[backup:db]
Database version refactoring. * Refactor db version constants into a separate module. * Update synthetic backup tests to PostgreSQL 9.4.
2016-08-12 04:35:24 +02:00
db-catalog-version=201409291
db-control-version=942
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
db-id=1
Automate generation of WAL and pg_control test files. The existing static files would not work with 32-bit or big-endian systems so create functions to generate these files dynamically rather than creating a bunch of new static files.
2017-11-19 03:02:54 +02:00
db-system-id=1000000000000000094
Database version refactoring. * Refactor db version constants into a separate module. * Update synthetic backup tests to PostgreSQL 9.4.
2016-08-12 04:35:24 +02:00
db-version="9.4"
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
[backup:option]
Set archive-check option in manifest correctly when offline. Archive check does not run when in offline backup mode but the option was set to true in the manifest. It's harmless since these options are informational only but it could cause confusion when debugging.
2019-11-28 15:27:21 +02:00
option-archive-check=false
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
option-archive-copy=true
Backup from a standby cluster. A connection to the primary cluster is still required to start/stop the backup and copy files that are not replicated, but the vast majority of files are copied from the standby in order to reduce load on the master.
2016-08-25 17:25:46 +02:00
option-backup-standby=false
Integration test improvements for disk and memory efficiency. Set log-level-file=off when more that one test will run. In this case is it impossible to see the logs anyway since they will be automatically cleaned up after the test. This improves performance pretty dramatically since trace-level logging is expensive. If a singe integration test is run then log-level-file is trace by default but can be changed with the --log-level-test-file option. Reduce buffer-size to 64k to save memory during testing and allow more processes to run in parallel. Update log replacement rules so that these options can change without affecting expect logs.
2019-12-17 22:23:07 +02:00
option-buffer-size=[BUFFER-SIZE]
Added the --checksum-page option. This option allows pgBackRest to validate page checksums in data files when checksums are enabled on PostgreSQL >= 9.3. Note that this functionality requires a C library which may not initially be available in OS packages. The option will automatically be enabled when the library is present and checksums are enabled on the cluster.
2016-12-13 01:54:07 +02:00
option-checksum-page=true
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
option-compress=false
Add additional options to backup.manifest for debugging purposes. Add the buffer-size, compress-level, compress-level-network, and process-max options to the backup:option section in backup.manifest to aid in debugging. It may also make sense to propagate these options up to backup.info so they can be displayed in the info command, but for now this is deemed sufficient. Contributed by blogh.
2019-03-10 11:03:52 +02:00
option-compress-level=3
option-compress-level-network=3
Add infrastructure for multiple compression type support. Add compress-type option and deprecate compress option. Since the compress option is boolean it won't work with multiple compression types. Add logic to cfgLoadUpdateOption() to update compress-type if it is not set directly. The compress option should no longer be referenced outside the cfgLoadUpdateOption() function. Add common/compress/helper module to contain interface functions that work with multiple compression types. Code outside this module should no longer call specific compression drivers, though it may be OK to reference a specific compression type using the new interface (e.g., saving backup history files in gz format). Unit tests only test compression using the gz format because other formats may not be available in all builds. It is the job of integration tests to exercise all compression types. Additional compression types will be added in future commits.
2020-03-06 21:41:03 +02:00
option-compress-type="none"
Add checksum delta for incremental backups. Use checksums rather than timestamps to determine if files have changed. This is useful in cases where the timestamps may not be trustworthy, e.g. when performing an incremental after failing over to a standby. If checksum delta is enabled then checksums will be used for verification of resumed backups, even if they are full. Resumes have always used checksums to verify the files in the repository, enabling delta performs checksums on the database files as well. Note that the user must manually enable this feature in cases were it would be useful or just keep in enabled all the time. A future commit will address automatically enabling the feature in cases where it seems likely to be useful. Contributed by Cynthia Shang.
2018-09-19 17:12:45 +02:00
option-delta=true
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
option-hardlink=false
Closed #167: Rename `--no-start-stop` option to `--no-online`.
2016-02-20 02:32:02 +02:00
option-online=false
Add additional options to backup.manifest for debugging purposes. Add the buffer-size, compress-level, compress-level-network, and process-max options to the backup:option section in backup.manifest to aid in debugging. It may also make sense to propagate these options up to backup.info so they can be displayed in the info command, but for now this is deemed sufficient. Contributed by blogh.
2019-03-10 11:03:52 +02:00
option-process-max=1
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
[backup:target]
Rename test paths for clarity. This was worked out as part of the test suite refactor [c8f806a] but not committed with it because of the large number of expect logs changes involved. Keeping them separate made it easier to audit the changes in the refactor.
2016-06-24 16:54:31 +02:00
pg_data={"path":"[TEST_PATH]/db-master/db/base","type":"path"}
pg_tblspc/1={"path":"[TEST_PATH]/db-master/db/tablespace/ts1","tablespace-id":"1","tablespace-name":"ts1","type":"link"}
Fixed an issue where tablespace paths with the same prefix would cause an invalid link error. Reported by Nikhilchandra Kulkarni.
2016-11-07 16:37:16 +02:00
pg_tblspc/11={"path":"[TEST_PATH]/db-master/db/tablespace/ts11","tablespace-id":"11","tablespace-name":"ts11","type":"link"}
Rename test paths for clarity. This was worked out as part of the test suite refactor [c8f806a] but not committed with it because of the large number of expect logs changes involved. Keeping them separate made it easier to audit the changes in the refactor.
2016-06-24 16:54:31 +02:00
pg_tblspc/2={"path":"[TEST_PATH]/db-master/db/tablespace/ts2","tablespace-id":"2","tablespace-name":"ts2","type":"link"}
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
[target:file]
Fixed a issue where tablespaces were copied from the master during standby backup.
2016-09-04 15:19:44 +02:00
pg_data/PG_VERSION={"checksum":"184473f470864e067ee3a22e64b47b0a1c356f29","master":true,"reference":"[BACKUP-FULL-2]","size":[SIZE],"timestamp":[TIMESTAMP-1]}
pg_data/badchecksum.txt={"checksum":"f927212cd08d11a42a666b2f04235398e9ceeb51","master":true,"size":[SIZE],"timestamp":[TIMESTAMP-1]}
Added the --checksum-page option. This option allows pgBackRest to validate page checksums in data files when checksums are enabled on PostgreSQL >= 9.3. Note that this functionality requires a C library which may not initially be available in OS packages. The option will automatically be enabled when the library is present and checksums are enabled on the cluster.
2016-12-13 01:54:07 +02:00
pg_data/base/1/12000={"checksum":"22c98d248ff548311eda88559e4a8405ed77c003","checksum-page":true,"reference":"[BACKUP-FULL-2]","size":[SIZE],"timestamp":[TIMESTAMP-1]}
Database version refactoring. * Refactor db version constants into a separate module. * Update synthetic backup tests to PostgreSQL 9.4.
2016-08-12 04:35:24 +02:00
pg_data/base/1/PG_VERSION={"checksum":"184473f470864e067ee3a22e64b47b0a1c356f29","mode":"0660","reference":"[BACKUP-FULL-2]","size":[SIZE],"timestamp":[TIMESTAMP-1]}
Added the --checksum-page option. This option allows pgBackRest to validate page checksums in data files when checksums are enabled on PostgreSQL >= 9.3. Note that this functionality requires a C library which may not initially be available in OS packages. The option will automatically be enabled when the library is present and checksums are enabled on the cluster.
2016-12-13 01:54:07 +02:00
pg_data/base/16384/17000={"checksum":"e0101dd8ffb910c9c202ca35b5f828bcb9697bed","checksum-page":false,"checksum-page-error":[1],"reference":"[BACKUP-FULL-2]","size":[SIZE],"timestamp":[TIMESTAMP-1]}
Database version refactoring. * Refactor db version constants into a separate module. * Update synthetic backup tests to PostgreSQL 9.4.
2016-08-12 04:35:24 +02:00
pg_data/base/16384/PG_VERSION={"checksum":"184473f470864e067ee3a22e64b47b0a1c356f29","reference":"[BACKUP-FULL-2]","size":[SIZE],"timestamp":[TIMESTAMP-1]}
Fix incorrect offline upper bound for ignoring page checksum errors. For offline backups the upper bound was being set to 0x0000FFFF0000FFFF rather than UINT64_MAX. This meant that page checksum errors might be ignored for databases with a lot of past WAL in offline mode. Online mode is not affected since the upper bound is retrieved from pg_start_backup().
2019-07-11 15:13:56 +02:00
pg_data/base/32768/33000={"checksum":"7a16d165e4775f7c92e8cdf60c0af57313f0bf90","checksum-page":true,"reference":"[BACKUP-FULL-2]","size":[SIZE],"timestamp":[TIMESTAMP-1]}
pg_data/base/32768/33000.32767={"checksum":"6e99b589e550e68e934fd235ccba59fe5b592a9e","checksum-page":true,"reference":"[BACKUP-FULL-2]","size":[SIZE],"timestamp":[TIMESTAMP-1]}
Added the --checksum-page option. This option allows pgBackRest to validate page checksums in data files when checksums are enabled on PostgreSQL >= 9.3. Note that this functionality requires a C library which may not initially be available in OS packages. The option will automatically be enabled when the library is present and checksums are enabled on the cluster.
2016-12-13 01:54:07 +02:00
pg_data/base/32768/33001={"checksum":"6bf316f11d28c28914ea9be92c00de9bea6d9a6b","checksum-page":false,"checksum-page-error":[0,[3,5],7],"reference":"[BACKUP-FULL-2]","size":[SIZE],"timestamp":[TIMESTAMP-1]}
Fix incorrect offline upper bound for ignoring page checksum errors. For offline backups the upper bound was being set to 0x0000FFFF0000FFFF rather than UINT64_MAX. This meant that page checksum errors might be ignored for databases with a lot of past WAL in offline mode. Online mode is not affected since the upper bound is retrieved from pg_start_backup().
2019-07-11 15:13:56 +02:00
pg_data/base/32768/44000_init={"checksum":"7a16d165e4775f7c92e8cdf60c0af57313f0bf90","checksum-page":true,"reference":"[BACKUP-FULL-2]","size":[SIZE],"timestamp":[TIMESTAMP-1]}
Database version refactoring. * Refactor db version constants into a separate module. * Update synthetic backup tests to PostgreSQL 9.4.
2016-08-12 04:35:24 +02:00
pg_data/base/32768/PG_VERSION={"checksum":"184473f470864e067ee3a22e64b47b0a1c356f29","reference":"[BACKUP-FULL-2]","size":[SIZE],"timestamp":[TIMESTAMP-1]}
Add checksum delta for incremental backups. Use checksums rather than timestamps to determine if files have changed. This is useful in cases where the timestamps may not be trustworthy, e.g. when performing an incremental after failing over to a standby. If checksum delta is enabled then checksums will be used for verification of resumed backups, even if they are full. Resumes have always used checksums to verify the files in the repository, enabling delta performs checksums on the database files as well. Note that the user must manually enable this feature in cases were it would be useful or just keep in enabled all the time. A future commit will address automatically enabling the feature in cases where it seems likely to be useful. Contributed by Cynthia Shang.
2018-09-19 17:12:45 +02:00
pg_data/changecontent.txt={"checksum":"a094d94583e209556d03c3c5da33131a065f1689","master":true,"size":[SIZE],"timestamp":[TIMESTAMP-1]}
Fix critical bug in resume that resulted in inconsistent backups. A regression in v0.82 removed the timestamp comparison when deciding which files from the aborted backup to keep on resume. All resumed backups should be considered inconsistent. A resumed backup can be identified by checking the log for the message "aborted backup of same type exists, will be cleaned to remove invalid files and resumed". Reported by David Youatt, Yogesh Sharma, Stephen Frost.
2018-07-03 20:01:57 +02:00
pg_data/changesize.txt={"checksum":"3905d5be2ec8d67f41435dab5e0dcda3ae47455d","master":true,"size":[SIZE],"timestamp":[TIMESTAMP-1]}
Add checksum delta for incremental backups. Use checksums rather than timestamps to determine if files have changed. This is useful in cases where the timestamps may not be trustworthy, e.g. when performing an incremental after failing over to a standby. If checksum delta is enabled then checksums will be used for verification of resumed backups, even if they are full. Resumes have always used checksums to verify the files in the repository, enabling delta performs checksums on the database files as well. Note that the user must manually enable this feature in cases were it would be useful or just keep in enabled all the time. A future commit will address automatically enabling the feature in cases where it seems likely to be useful. Contributed by Cynthia Shang.
2018-09-19 17:12:45 +02:00
pg_data/changetime.txt={"checksum":"88087292ed82e26f3eb824d0bffc05ccf7a30f8d","master":true,"reference":"[BACKUP-FULL-2]","size":[SIZE],"timestamp":[TIMESTAMP-2]}
Support configurable WAL segment size. PostgreSQL 11 introduces configurable WAL segment sizes, from 1MB to 1GB. There are two areas that needed to be updated to support this: building the archive-get queue and checking that WAL has been archived after a backup. Both operations require the WAL segment size to properly build a list. Checking the archive after a backup is still implemented in Perl and has an active database connection, so just get the WAL segment size from the database. The archive-get command does not have a connection to the database, so get the WAL segment size from pg_control instead. This requires a deeper inspection of pg_control than has been done in the past, so it seemed best to copy the relevant data structures from each version of PostgreSQL and build a generic interface layer to address them. While this approach is a bit verbose, it has the advantage of being relatively simple, and can easily be updated for new versions of PostgreSQL. Since the integration tests generate pg_control files for testing, teach Perl how to generate files with the correct offsets for both 32-bit and 64-bit architectures.
2018-09-25 11:24:42 +02:00
pg_data/global/pg_control={"checksum":"4c77c900f7af0d9ab13fa9982051a42e0b637f6c","master":true,"reference":"[BACKUP-FULL-2]","size":[SIZE],"timestamp":[TIMESTAMP-2]}
Fixed a issue where tablespaces were copied from the master during standby backup.
2016-09-04 15:19:44 +02:00
pg_data/pg_stat/global.stat={"checksum":"e350d5ce0153f3e22d5db21cf2a4eff00f3ee877","master":true,"reference":"[BACKUP-FULL-2]","size":[SIZE],"timestamp":[TIMESTAMP-2]}
pg_data/postgresql.conf={"checksum":"6721d92c9fcdf4248acff1f9a1377127d9064807","master":true,"reference":"[BACKUP-FULL-2]","size":[SIZE],"timestamp":[TIMESTAMP-2]}
Use minio for integration testing. ScalityS3 has not received any maintenance in years and is slow to start which is bad for testing. Replace it with minio which starts quickly and ships as a single executable or a tiny container. Minio has stricter limits on allowable characters but should still provide enough coverage to show that our encoding is working correctly. This commit also includes the upgrade to openssl 1.1.1 in the Ubuntu 18.04 container.
2019-07-03 04:20:35 +02:00
pg_data/special-!_.*'()&!@;:+,?={"master":true,"reference":"[BACKUP-FULL-2]","size":[SIZE],"timestamp":[TIMESTAMP-1]}
Allow zero-size files in backup manifest to reference a prior manifest regardless of timestamp delta. Contributed by Cynthia Shang.
2018-08-24 22:50:33 +02:00
pg_data/zero_from_start={"master":true,"reference":"[BACKUP-FULL-2]","size":[SIZE],"timestamp":[TIMESTAMP-1]}
Add zero-length file to mock/all test.
2018-07-05 21:40:50 +02:00
pg_data/zerosize.txt={"master":true,"size":[SIZE],"timestamp":[TIMESTAMP-2]}
Added the --checksum-page option. This option allows pgBackRest to validate page checksums in data files when checksums are enabled on PostgreSQL >= 9.3. Note that this functionality requires a C library which may not initially be available in OS packages. The option will automatically be enabled when the library is present and checksums are enabled on the cluster.
2016-12-13 01:54:07 +02:00
pg_tblspc/1/[TS_PATH-1]/16384/tablespace1.txt={"checksum":"d85de07d6421d90aa9191c11c889bfde43680f0f","checksum-page":false,"size":[SIZE],"timestamp":[TIMESTAMP-1]}
pg_tblspc/2/[TS_PATH-1]/32768/tablespace2.txt={"checksum":"dc7f76e43c46101b47acc55ae4d593a9e6983578","checksum-page":false,"size":[SIZE],"timestamp":[TIMESTAMP-1]}
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
[target:file:default]
Continuous integration using travis-ci.
2017-02-21 15:59:23 +02:00
group="[GROUP-1]"
Fixed a issue where tablespaces were copied from the master during standby backup.
2016-09-04 15:19:44 +02:00
master=false
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
mode="0600"
Regression tests can now be run as any properly-configured user, not just vagrant.
2016-10-01 19:39:44 +02:00
user="[USER-1]"
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
[target:link]
Rename test paths for clarity. This was worked out as part of the test suite refactor [c8f806a] but not committed with it because of the large number of expect logs changes involved. Keeping them separate made it easier to audit the changes in the refactor.
2016-06-24 16:54:31 +02:00
pg_data/pg_tblspc/1={"destination":"[TEST_PATH]/db-master/db/tablespace/ts1"}
Fixed an issue where tablespace paths with the same prefix would cause an invalid link error. Reported by Nikhilchandra Kulkarni.
2016-11-07 16:37:16 +02:00
pg_data/pg_tblspc/11={"destination":"[TEST_PATH]/db-master/db/tablespace/ts11"}
Rename test paths for clarity. This was worked out as part of the test suite refactor [c8f806a] but not committed with it because of the large number of expect logs changes involved. Keeping them separate made it easier to audit the changes in the refactor.
2016-06-24 16:54:31 +02:00
pg_data/pg_tblspc/2={"destination":"[TEST_PATH]/db-master/db/tablespace/ts2"}
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
[target:link:default]
Continuous integration using travis-ci.
2017-02-21 15:59:23 +02:00
group="[GROUP-1]"
Regression tests can now be run as any properly-configured user, not just vagrant.
2016-10-01 19:39:44 +02:00
user="[USER-1]"
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
[target:path]
Fixed an issue with invalid backup groups being set correctly on restore. If the backup cannot map a group to a name it stores the group in the manifest as false then uses either the owner of $PGDATA to set the group during restore or failing that the group of the current user. This logic was not working correctly because the selected group was overwriting the user on restore leaving the group undefined and the user incorrectly set to the group. (Reported by Jeff McCormick.)
2017-11-20 19:46:35 +02:00
pg_data={"mode":"0777"}
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
pg_data/base={}
Allow selective restore of databases from a cluster backup. This feature can result in major space and time savings when only specific databases are restored. Unrestored databases will not be accessible but must be manually dropped before they will be removed from the shared catalogue.
2016-05-11 15:21:39 +02:00
pg_data/base/1={}
pg_data/base/16384={}
pg_data/base/32768={}
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
pg_data/global={}
pg_data/pg_clog={}
Exclude directories during backup that are cleaned, recreated, or zeroed by PostgreSQL at startup. These include (depending on the version where they were introduced): pgsql_tmp, pg_dynshmem, pg_notify, pg_replslot, pg_serial, pg_snapshots, pg_stat_tmp, pg_subtrans. The postgresql.auto.conf.tmp file is now excluded in addition to files that were already excluded: backup_label.old, postmaster.opts, postmaster.pid, recovery.conf, recovery.done.
2016-08-16 15:35:16 +02:00
pg_data/pg_dynshmem={}
Allow arbitrary directories and/or files to be excluded from a backup. Misuse of this feature can lead to inconsistent backups so read the --exclude documentation carefully before using.
2018-08-27 21:51:05 +02:00
pg_data/pg_log={}
Exclude directories during backup that are cleaned, recreated, or zeroed by PostgreSQL at startup. These include (depending on the version where they were introduced): pgsql_tmp, pg_dynshmem, pg_notify, pg_replslot, pg_serial, pg_snapshots, pg_stat_tmp, pg_subtrans. The postgresql.auto.conf.tmp file is now excluded in addition to files that were already excluded: backup_label.old, postmaster.opts, postmaster.pid, recovery.conf, recovery.done.
2016-08-16 15:35:16 +02:00
pg_data/pg_notify={}
pg_data/pg_replslot={}
pg_data/pg_serial={}
pg_data/pg_snapshots={}
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
pg_data/pg_stat={}
Exclude directories during backup that are cleaned, recreated, or zeroed by PostgreSQL at startup. These include (depending on the version where they were introduced): pgsql_tmp, pg_dynshmem, pg_notify, pg_replslot, pg_serial, pg_snapshots, pg_stat_tmp, pg_subtrans. The postgresql.auto.conf.tmp file is now excluded in addition to files that were already excluded: backup_label.old, postmaster.opts, postmaster.pid, recovery.conf, recovery.done.
2016-08-16 15:35:16 +02:00
pg_data/pg_stat_tmp={}
pg_data/pg_subtrans={}
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
pg_data/pg_tblspc={}
Fixed an issue with invalid backup groups being set correctly on restore. If the backup cannot map a group to a name it stores the group in the manifest as false then uses either the owner of $PGDATA to set the group during restore or failing that the group of the current user. This logic was not working correctly because the selected group was overwriting the user on restore leaving the group undefined and the user incorrectly set to the group. (Reported by Jeff McCormick.)
2017-11-20 19:46:35 +02:00
pg_tblspc={"mode":"0777"}
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
pg_tblspc/1={}
pg_tblspc/1/[TS_PATH-1]={}
Allow selective restore of databases from a cluster backup. This feature can result in major space and time savings when only specific databases are restored. Unrestored databases will not be accessible but must be manually dropped before they will be removed from the shared catalogue.
2016-05-11 15:21:39 +02:00
pg_tblspc/1/[TS_PATH-1]/16384={}
Remove integration tests that are now covered in the unit tests. Most of these tests are just checking that errors are thrown when required. These are well covered in various unit tests. The "cannot resume" tests are also well covered in the backup unit tests. Finally, config warnings are well covered in the config unit tests. There is more to be done here, but this accounts for the low-hanging fruit.
2019-12-18 03:14:45 +02:00
pg_tblspc/11={}
pg_tblspc/11/[TS_PATH-1]={}
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
pg_tblspc/2={}
pg_tblspc/2/[TS_PATH-1]={}
Allow selective restore of databases from a cluster backup. This feature can result in major space and time savings when only specific databases are restored. Unrestored databases will not be accessible but must be manually dropped before they will be removed from the shared catalogue.
2016-05-11 15:21:39 +02:00
pg_tblspc/2/[TS_PATH-1]/32768={}
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
[target:path:default]
Continuous integration using travis-ci.
2017-02-21 15:59:23 +02:00
group="[GROUP-1]"
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
mode="0700"
Regression tests can now be run as any properly-configured user, not just vagrant.
2016-10-01 19:39:44 +02:00
user="[USER-1]"
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
Move info file checksum to the end of the file. Putting the checksum at the beginning of the file made it impossible to stream the file out when saving. The entire file had to be held in memory while it was checksummed so the checksum could be written at the beginning. Instead place the checksum at the end. This does not break the existing Perl or C code since the read is not order dependent. There are no plans to improve the Perl code to take advantage of this change, but it will make the C implementation more efficient. Reviewed by Cynthia Shang.
2019-08-22 01:45:48 +02:00
[backrest]
backrest-checksum="[CHECKSUM]"
Rename test paths for clarity. This was worked out as part of the test suite refactor [c8f806a] but not committed with it because of the large number of expect logs changes involved. Keeping them separate made it easier to audit the changes in the refactor.
2016-06-24 16:54:31 +02:00
+ supplemental file: [TEST_PATH]/db-master/repo/backup/db/backup.info
---------------------------------------------------------------------
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
[backrest]
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
backrest-format=5
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
backrest-version="[VERSION-1]"
[backup:current]
Set archive-check option in manifest correctly when offline. Archive check does not run when in offline backup mode but the option was set to true in the manifest. It's harmless since these options are informational only but it could cause confusion when debugging.
2019-11-28 15:27:21 +02:00
[BACKUP-FULL-2]={"backrest-format":5,"backrest-version":"[VERSION-1]","backup-archive-start":null,"backup-archive-stop":null,"backup-info-repo-size":[SIZE],"backup-info-repo-size-delta":[DELTA],"backup-info-size":[SIZE],"backup-info-size-delta":[DELTA],"backup-timestamp-start":[TIMESTAMP],"backup-timestamp-stop":[TIMESTAMP],"backup-type":"full","db-id":1,"option-archive-check":false,"option-archive-copy":true,"option-backup-standby":false,"option-checksum-page":true,"option-compress":false,"option-hardlink":false,"option-online":false}
Remove integration tests that are now covered in the unit tests. Most of these tests are just checking that errors are thrown when required. These are well covered in various unit tests. The "cannot resume" tests are also well covered in the backup unit tests. Finally, config warnings are well covered in the config unit tests. There is more to be done here, but this accounts for the low-hanging fruit.
2019-12-18 03:14:45 +02:00
[BACKUP-INCR-2]={"backrest-format":5,"backrest-version":"[VERSION-1]","backup-archive-start":null,"backup-archive-stop":null,"backup-info-repo-size":[SIZE],"backup-info-repo-size-delta":[DELTA],"backup-info-size":[SIZE],"backup-info-size-delta":[DELTA],"backup-prior":"[BACKUP-FULL-2]","backup-reference":["[BACKUP-FULL-2]"],"backup-timestamp-start":[TIMESTAMP],"backup-timestamp-stop":[TIMESTAMP],"backup-type":"incr","db-id":1,"option-archive-check":false,"option-archive-copy":true,"option-backup-standby":false,"option-checksum-page":true,"option-compress":false,"option-hardlink":false,"option-online":false}
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
[db]
Database version refactoring. * Refactor db version constants into a separate module. * Update synthetic backup tests to PostgreSQL 9.4.
2016-08-12 04:35:24 +02:00
db-catalog-version=201409291
db-control-version=942
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
db-id=1
Automate generation of WAL and pg_control test files. The existing static files would not work with 32-bit or big-endian systems so create functions to generate these files dynamically rather than creating a bunch of new static files.
2017-11-19 03:02:54 +02:00
db-system-id=1000000000000000094
Database version refactoring. * Refactor db version constants into a separate module. * Update synthetic backup tests to PostgreSQL 9.4.
2016-08-12 04:35:24 +02:00
db-version="9.4"
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
[db:history]
Automate generation of WAL and pg_control test files. The existing static files would not work with 32-bit or big-endian systems so create functions to generate these files dynamically rather than creating a bunch of new static files.
2017-11-19 03:02:54 +02:00
1={"db-catalog-version":201409291,"db-control-version":942,"db-system-id":1000000000000000094,"db-version":"9.4"}
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
Improve performance of info file load/save. Info files required three copies in memory to be loaded (the original string, an ini representation, and the final info object). Not only was this memory inefficient but the Ini object does sequential scans when searching for keys making large files very slow to load. This has not been an issue since archive.info and backup.info are very small, but it becomes a big deal when loading manifests with hundreds of thousands of files. Instead of holding copies of the data in memory, use a callback to deliver the ini data directly to the object when loading. Use a similar method for save to avoid having an intermediate copy. Save is a bit complex because sections/keys must be written in alpha order or older versions of pgBackRest will not calculate the correct checksum. Also move the load retry logic to helper functions rather than embedding it in the Info object. This allows for more flexibility in loading and ensures that stack traces will be available when developing unit tests. Reviewed by Cynthia Shang.
2019-09-06 19:48:28 +02:00
[backrest]
backrest-checksum="[CHECKSUM]"
Remove integration tests that are now covered in the unit tests. Most of these tests are just checking that errors are thrown when required. These are well covered in various unit tests. The "cannot resume" tests are also well covered in the backup unit tests. Finally, config warnings are well covered in the config unit tests. There is more to be done here, but this accounts for the low-hanging fruit.
2019-12-18 03:14:45 +02:00
diff backup - drop tablespace 11 (db-master host)
> [CONTAINER-EXEC] db-master [BACKREST-BIN] --config=[TEST_PATH]/db-master/pgbackrest.conf --no-online --process-max=1 --delta --type=diff --stanza=db backup
Added resume setting to allow disabling resume feature.
2015-05-07 18:29:30 +02:00
------------------------------------------------------------------------------------------------------------------------------------
Add infrastructure for multiple compression type support. Add compress-type option and deprecate compress option. Since the compress option is boolean it won't work with multiple compression types. Add logic to cfgLoadUpdateOption() to update compress-type if it is not set directly. The compress option should no longer be referenced outside the cfgLoadUpdateOption() function. Add common/compress/helper module to contain interface functions that work with multiple compression types. Code outside this module should no longer call specific compression drivers, though it may be OK to reference a specific compression type using the new interface (e.g., saving backup history files in gz format). Unit tests only test compression using the gz format because other formats may not be available in all builds. It is the job of integration tests to exercise all compression types. Additional compression types will be added in future commits.
2020-03-06 21:41:03 +02:00
P00 INFO: backup command begin [BACKREST-VERSION]: --buffer-size=[BUFFER-SIZE] --compress-level=3 --compress-type=none --config=[TEST_PATH]/db-master/pgbackrest.conf --db-timeout=45 --delta --lock-path=[TEST_PATH]/db-master/lock --log-level-console=detail --log-level-file=[LOG-LEVEL-FILE] --log-level-stderr=off --log-path=[TEST_PATH]/db-master/log[] --no-log-timestamp --no-online --pg1-path=[TEST_PATH]/db-master/db/base --process-max=1 --protocol-timeout=60 --repo1-path=[TEST_PATH]/db-master/repo --stanza=db --start-fast --type=diff
Add time-based retention for full backups. The --repo-retention-full-type option allows retention of full backups based on a time period, specified in days. The new option will default to 'count' and therefore will not affect current installations. Setting repo-retention-full-type to 'time' will allow the user to use a time period, in days, to indicate full backup retention. Using this method, a full backup can be expired only if the time the backup completed is older than the number of days set with repo-retention-full (calculated from the moment the 'expire' command is run) and at least one full backup meets the retention period. If archive retention has not been configured, then the default settings will expire archives that are prior to the oldest retained full backup. For example, if there are three full backups ending in times that are 25 days old (F1), 20 days old (F2) and 10 days old (F3), then if the full retention period is 15 days, then only F1 will be expired; F2 will be retained because F1 is not at least 15 days old.
2020-05-08 21:25:03 +02:00
P00 WARN: option 'repo1-retention-full' is not set for 'repo1-retention-full-type=count', the repository may run out of space
Rename retention-* options to repo-retention-*.
2018-02-19 22:07:24 +02:00
HINT: to retain full backups indefinitely (without warning), set option 'repo1-retention-full' to the maximum.
Show Process ID in log output instead of filtering it out with the timestamp.
2016-09-14 23:30:57 +02:00
P00 INFO: last backup label = [BACKUP-FULL-2], version = [VERSION-1]
Storage and IO layer refactor: Refactor storage layer to allow for new repository filesystems using drivers. (Reviewed by Cynthia Shang.) Refactor IO layer to allow for new compression formats, checksum types, and other capabilities using filters. (Reviewed by Cynthia Shang.)
2017-06-09 23:51:41 +02:00
P00 WARN: diff backup cannot alter 'checksum-page' option to 'false', reset to 'true' from [BACKUP-FULL-2]
Add checksum delta for incremental backups. Use checksums rather than timestamps to determine if files have changed. This is useful in cases where the timestamps may not be trustworthy, e.g. when performing an incremental after failing over to a standby. If checksum delta is enabled then checksums will be used for verification of resumed backups, even if they are full. Resumes have always used checksums to verify the files in the repository, enabling delta performs checksums on the database files as well. Note that the user must manually enable this feature in cases were it would be useful or just keep in enabled all the time. A future commit will address automatically enabling the feature in cases where it seems likely to be useful. Contributed by Cynthia Shang.
2018-09-19 17:12:45 +02:00
P01 DETAIL: match file from prior backup [TEST_PATH]/db-master/db/base/base/32768/33001 (64KB, 33%) checksum 6bf316f11d28c28914ea9be92c00de9bea6d9a6b
Fix incorrect offline upper bound for ignoring page checksum errors. For offline backups the upper bound was being set to 0x0000FFFF0000FFFF rather than UINT64_MAX. This meant that page checksum errors might be ignored for databases with a lot of past WAL in offline mode. Online mode is not affected since the upper bound is retrieved from pg_start_backup().
2019-07-11 15:13:56 +02:00
P01 DETAIL: match file from prior backup [TEST_PATH]/db-master/db/base/base/32768/44000_init (32KB, 49%) checksum 7a16d165e4775f7c92e8cdf60c0af57313f0bf90
P01 DETAIL: match file from prior backup [TEST_PATH]/db-master/db/base/base/32768/33000.32767 (32KB, 66%) checksum 6e99b589e550e68e934fd235ccba59fe5b592a9e
P01 DETAIL: match file from prior backup [TEST_PATH]/db-master/db/base/base/32768/33000 (32KB, 83%) checksum 7a16d165e4775f7c92e8cdf60c0af57313f0bf90
Add checksum delta for incremental backups. Use checksums rather than timestamps to determine if files have changed. This is useful in cases where the timestamps may not be trustworthy, e.g. when performing an incremental after failing over to a standby. If checksum delta is enabled then checksums will be used for verification of resumed backups, even if they are full. Resumes have always used checksums to verify the files in the repository, enabling delta performs checksums on the database files as well. Note that the user must manually enable this feature in cases were it would be useful or just keep in enabled all the time. A future commit will address automatically enabling the feature in cases where it seems likely to be useful. Contributed by Cynthia Shang.
2018-09-19 17:12:45 +02:00
P01 DETAIL: match file from prior backup [TEST_PATH]/db-master/db/base/base/16384/17000 (16KB, 91%) checksum e0101dd8ffb910c9c202ca35b5f828bcb9697bed
Support configurable WAL segment size. PostgreSQL 11 introduces configurable WAL segment sizes, from 1MB to 1GB. There are two areas that needed to be updated to support this: building the archive-get queue and checking that WAL has been archived after a backup. Both operations require the WAL segment size to properly build a list. Checking the archive after a backup is still implemented in Perl and has an active database connection, so just get the WAL segment size from the database. The archive-get command does not have a connection to the database, so get the WAL segment size from pg_control instead. This requires a deeper inspection of pg_control than has been done in the past, so it seemed best to copy the relevant data structures from each version of PostgreSQL and build a generic interface layer to address them. While this approach is a bit verbose, it has the advantage of being relatively simple, and can easily be updated for new versions of PostgreSQL. Since the integration tests generate pg_control files for testing, teach Perl how to generate files with the correct offsets for both 32-bit and 64-bit architectures.
2018-09-25 11:24:42 +02:00
P01 DETAIL: match file from prior backup [TEST_PATH]/db-master/db/base/global/pg_control (8KB, 95%) checksum 4c77c900f7af0d9ab13fa9982051a42e0b637f6c
Add checksum delta for incremental backups. Use checksums rather than timestamps to determine if files have changed. This is useful in cases where the timestamps may not be trustworthy, e.g. when performing an incremental after failing over to a standby. If checksum delta is enabled then checksums will be used for verification of resumed backups, even if they are full. Resumes have always used checksums to verify the files in the repository, enabling delta performs checksums on the database files as well. Note that the user must manually enable this feature in cases were it would be useful or just keep in enabled all the time. A future commit will address automatically enabling the feature in cases where it seems likely to be useful. Contributed by Cynthia Shang.
2018-09-19 17:12:45 +02:00
P01 DETAIL: match file from prior backup [TEST_PATH]/db-master/db/base/base/1/12000 (8KB, 99%) checksum 22c98d248ff548311eda88559e4a8405ed77c003
P01 DETAIL: match file from prior backup [TEST_PATH]/db-master/db/base/postgresql.conf (21B, 99%) checksum 6721d92c9fcdf4248acff1f9a1377127d9064807
P01 INFO: backup file [TEST_PATH]/db-master/db/base/badchecksum.txt (11B, 99%) checksum f927212cd08d11a42a666b2f04235398e9ceeb51
P01 INFO: backup file [TEST_PATH]/db-master/db/base/changecontent.txt (7B, 99%) checksum a094d94583e209556d03c3c5da33131a065f1689
P01 DETAIL: match file from prior backup [TEST_PATH]/db-master/db/base/pg_stat/global.stat (5B, 99%) checksum e350d5ce0153f3e22d5db21cf2a4eff00f3ee877
P01 DETAIL: match file from prior backup [TEST_PATH]/db-master/db/base/changetime.txt (4B, 99%) checksum 88087292ed82e26f3eb824d0bffc05ccf7a30f8d
P01 DETAIL: match file from prior backup [TEST_PATH]/db-master/db/base/base/32768/PG_VERSION (3B, 99%) checksum 184473f470864e067ee3a22e64b47b0a1c356f29
P01 DETAIL: match file from prior backup [TEST_PATH]/db-master/db/base/base/16384/PG_VERSION (3B, 99%) checksum 184473f470864e067ee3a22e64b47b0a1c356f29
P01 DETAIL: match file from prior backup [TEST_PATH]/db-master/db/base/base/1/PG_VERSION (3B, 99%) checksum 184473f470864e067ee3a22e64b47b0a1c356f29
P01 DETAIL: match file from prior backup [TEST_PATH]/db-master/db/base/PG_VERSION (3B, 99%) checksum 184473f470864e067ee3a22e64b47b0a1c356f29
P01 INFO: backup file [TEST_PATH]/db-master/db/base/zerosize.txt (0B, 99%)
The backup command is implemented entirely in C. For the most part this is a direct migration of the Perl code into C except as noted below. A backup can now be initiated from a linked directory. The link will not be stored in the manifest or recreated on restore. If a link or directory does not already exist in the restore location then a directory will be created. The logic for creating backup labels has been improved and it should no longer be possible to get a backup label earlier than the latest backup even with timezone changes or clock skew. This has never been an issue in the field that we know of, but we found it in testing. For online backups all times are fetched from the PostgreSQL primary host (before only copy start was). This doesn't affect backup integrity but it does prevent clock skew between hosts affecting backup duration reporting. Archive copy now works as expected when the archive and backup have different compression settings, i.e. when one is compressed and the other is not. This was a long-standing bug in the Perl code. Resume will now work even if hardlink settings have been changed. Reviewed by Cynthia Shang.
2019-12-14 00:14:26 +02:00
P01 INFO: backup file [TEST_PATH]/db-master/db/base/pg_tblspc/1/[TS_PATH-1]/16384/tablespace1.txt (7B, 99%) checksum d85de07d6421d90aa9191c11c889bfde43680f0f
Added the --checksum-page option. This option allows pgBackRest to validate page checksums in data files when checksums are enabled on PostgreSQL >= 9.3. Note that this functionality requires a C library which may not initially be available in OS packages. The option will automatically be enabled when the library is present and checksums are enabled on the cluster.
2016-12-13 01:54:07 +02:00
P00 WARN: page misalignment in file [TEST_PATH]/db-master/db/base/pg_tblspc/1/[TS_PATH-1]/16384/tablespace1.txt: file size 7 is not divisible by page size 8192
The backup command is implemented entirely in C. For the most part this is a direct migration of the Perl code into C except as noted below. A backup can now be initiated from a linked directory. The link will not be stored in the manifest or recreated on restore. If a link or directory does not already exist in the restore location then a directory will be created. The logic for creating backup labels has been improved and it should no longer be possible to get a backup label earlier than the latest backup even with timezone changes or clock skew. This has never been an issue in the field that we know of, but we found it in testing. For online backups all times are fetched from the PostgreSQL primary host (before only copy start was). This doesn't affect backup integrity but it does prevent clock skew between hosts affecting backup duration reporting. Archive copy now works as expected when the archive and backup have different compression settings, i.e. when one is compressed and the other is not. This was a long-standing bug in the Perl code. Resume will now work even if hardlink settings have been changed. Reviewed by Cynthia Shang.
2019-12-14 00:14:26 +02:00
P01 INFO: backup file [TEST_PATH]/db-master/db/base/pg_tblspc/2/[TS_PATH-1]/32768/tablespace2.txt (7B, 100%) checksum dc7f76e43c46101b47acc55ae4d593a9e6983578
P00 WARN: page misalignment in file [TEST_PATH]/db-master/db/base/pg_tblspc/2/[TS_PATH-1]/32768/tablespace2.txt: file size 7 is not divisible by page size 8192
P00 DETAIL: reference pg_data/PG_VERSION to [BACKUP-FULL-2]
P00 DETAIL: reference pg_data/base/1/12000 to [BACKUP-FULL-2]
P00 DETAIL: reference pg_data/base/1/PG_VERSION to [BACKUP-FULL-2]
P00 DETAIL: reference pg_data/base/16384/17000 to [BACKUP-FULL-2]
P00 DETAIL: reference pg_data/base/16384/PG_VERSION to [BACKUP-FULL-2]
P00 DETAIL: reference pg_data/base/32768/33000 to [BACKUP-FULL-2]
P00 DETAIL: reference pg_data/base/32768/33000.32767 to [BACKUP-FULL-2]
P00 DETAIL: reference pg_data/base/32768/33001 to [BACKUP-FULL-2]
P00 DETAIL: reference pg_data/base/32768/44000_init to [BACKUP-FULL-2]
P00 DETAIL: reference pg_data/base/32768/PG_VERSION to [BACKUP-FULL-2]
P00 DETAIL: reference pg_data/changetime.txt to [BACKUP-FULL-2]
P00 DETAIL: reference pg_data/global/pg_control to [BACKUP-FULL-2]
P00 DETAIL: reference pg_data/pg_stat/global.stat to [BACKUP-FULL-2]
P00 DETAIL: reference pg_data/postgresql.conf to [BACKUP-FULL-2]
P00 DETAIL: reference pg_data/special-!_.*'()&!@;:+,? to [BACKUP-FULL-2]
P00 DETAIL: reference pg_data/zero_from_start to [BACKUP-FULL-2]
Add checksum delta for incremental backups. Use checksums rather than timestamps to determine if files have changed. This is useful in cases where the timestamps may not be trustworthy, e.g. when performing an incremental after failing over to a standby. If checksum delta is enabled then checksums will be used for verification of resumed backups, even if they are full. Resumes have always used checksums to verify the files in the repository, enabling delta performs checksums on the database files as well. Note that the user must manually enable this feature in cases were it would be useful or just keep in enabled all the time. A future commit will address automatically enabling the feature in cases where it seems likely to be useful. Contributed by Cynthia Shang.
2018-09-19 17:12:45 +02:00
P00 INFO: diff backup size = 192KB
Remove integration tests that are now covered in the unit tests. Most of these tests are just checking that errors are thrown when required. These are well covered in various unit tests. The "cannot resume" tests are also well covered in the backup unit tests. Finally, config warnings are well covered in the config unit tests. There is more to be done here, but this accounts for the low-hanging fruit.
2019-12-18 03:14:45 +02:00
P00 INFO: new backup label = [BACKUP-DIFF-1]
Added a predictable log entry to signal that a command has completed successfully. For example a backup ends successfully with: INFO: backup command end: completed successfully.
2016-12-04 21:44:53 +02:00
P00 INFO: backup command end: completed successfully
Integration test improvements for disk and memory efficiency. Set log-level-file=off when more that one test will run. In this case is it impossible to see the logs anyway since they will be automatically cleaned up after the test. This improves performance pretty dramatically since trace-level logging is expensive. If a singe integration test is run then log-level-file is trace by default but can be changed with the --log-level-test-file option. Reduce buffer-size to 64k to save memory during testing and allow more processes to run in parallel. Update log replacement rules so that these options can change without affecting expect logs.
2019-12-17 22:23:07 +02:00
P00 INFO: expire command begin [BACKREST-VERSION]: --buffer-size=[BUFFER-SIZE] --config=[TEST_PATH]/db-master/pgbackrest.conf --lock-path=[TEST_PATH]/db-master/lock --log-level-console=detail --log-level-file=[LOG-LEVEL-FILE] --log-level-stderr=off --log-path=[TEST_PATH]/db-master/log[] --no-log-timestamp --repo1-path=[TEST_PATH]/db-master/repo --stanza=db
Rename retention-* options to repo-retention-*.
2018-02-19 22:07:24 +02:00
P00 INFO: option 'repo1-retention-archive' is not set - archive logs will not be expired
Added a predictable log entry to signal that a command has completed successfully. For example a backup ends successfully with: INFO: backup command end: completed successfully.
2016-12-04 21:44:53 +02:00
P00 INFO: expire command end: completed successfully
Log testing can now be enabled for certain deterministic tests. This works by comparing the generated logs against a previous copy. Currently only enabled for the backup/synthetic tests.
2015-05-05 19:08:48 +02:00
Rename test paths for clarity. This was worked out as part of the test suite refactor [c8f806a] but not committed with it because of the large number of expect logs changes involved. Keeping them separate made it easier to audit the changes in the refactor.
2016-06-24 16:54:31 +02:00
+ supplemental file: [TEST_PATH]/db-master/pgbackrest.conf
----------------------------------------------------------
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
[db]
Rename db-* options to pg-* and backup-* options to repo-* to improve consistency. * repo-* options are now indexed although only one is allowed. * List deprecated option names in documentation and command-line help.
2018-02-04 01:27:38 +02:00
pg1-path=[TEST_PATH]/db-master/db/base
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
New simpler configuration and consistent project/exe/path naming. * The repo-path option now always refers to the repository where backups and archive are stored, whether local or remote, so the repo-remote-path option has been removed. The new spool-path option can be used to define a location for queueing WAL segments when archiving asynchronously. Otherwise, a local repository is no longer required. * Implemented a new config format which should be far simpler to use. See the User Guide and Configuration Reference for details but for a simple configuration all options can now be placed in the stanza section. Options that are shared between stanzas can be placed in the [global] section. More complex configurations can still make use of command sections though this should be a rare use case. * The default configuration filename is now pgbackrest.conf instead of pg_backrest.conf. This was done for consistency with other naming changes but also to prevent old config files from being loaded accidentally. * The default repository name was changed from /var/lib/backup to /var/lib/pgbackrest. * Lock files are now stored in /tmp/pgbackrest by default. These days /run/pgbackrest would be the preferred location but that would require init scripts which are not part of this release. The lock-path option can be used to configure the lock directory. * Log files are now stored in /var/log/pgbackrest by default and no longer have the date appended so they can be managed with logrotate. The log-path option can be used to configure the lock directory. * Executable filename changed from pg_backrest to pgbackrest.
2016-04-14 15:30:54 +02:00
[global]
Integration test improvements for disk and memory efficiency. Set log-level-file=off when more that one test will run. In this case is it impossible to see the logs anyway since they will be automatically cleaned up after the test. This improves performance pretty dramatically since trace-level logging is expensive. If a singe integration test is run then log-level-file is trace by default but can be changed with the --log-level-test-file option. Reduce buffer-size to 64k to save memory during testing and allow more processes to run in parallel. Update log replacement rules so that these options can change without affecting expect logs.
2019-12-17 22:23:07 +02:00
buffer-size=[BUFFER-SIZE]
Fix custom settings for compress-level option being ignored. Reported by Jens Wilke.
2017-10-23 00:05:46 +02:00
compress-level=3
Add infrastructure for multiple compression type support. Add compress-type option and deprecate compress option. Since the compress option is boolean it won't work with multiple compression types. Add logic to cfgLoadUpdateOption() to update compress-type if it is not set directly. The compress option should no longer be referenced outside the cfgLoadUpdateOption() function. Add common/compress/helper module to contain interface functions that work with multiple compression types. Code outside this module should no longer call specific compression drivers, though it may be OK to reference a specific compression type using the new interface (e.g., saving backup history files in gz format). Unit tests only test compression using the gz format because other formats may not be available in all builds. It is the job of integration tests to exercise all compression types. Additional compression types will be added in future commits.
2020-03-06 21:41:03 +02:00
compress-type=none
Storage and IO layer refactor: Refactor storage layer to allow for new repository filesystems using drivers. (Reviewed by Cynthia Shang.) Refactor IO layer to allow for new compression formats, checksum types, and other capabilities using filters. (Reviewed by Cynthia Shang.)
2017-06-09 23:51:41 +02:00
db-timeout=45
lock-path=[TEST_PATH]/db-master/lock
Reduce log level for all expect tests to detail. The C code is designed to be efficient rather than deterministic at the debug log level. As we move more testing from integration to unit tests it makes less sense to try and maintain the expect logs at this log level. Most of the expect logs have already been moved to detail level but mock/all still had tests at debug level. Change the logging defaults in the config file and remove as many references to log-level-console as possible.
2019-05-23 00:23:44 +02:00
log-level-console=detail
Integration test improvements for disk and memory efficiency. Set log-level-file=off when more that one test will run. In this case is it impossible to see the logs anyway since they will be automatically cleaned up after the test. This improves performance pretty dramatically since trace-level logging is expensive. If a singe integration test is run then log-level-file is trace by default but can be changed with the --log-level-test-file option. Reduce buffer-size to 64k to save memory during testing and allow more processes to run in parallel. Update log replacement rules so that these options can change without affecting expect logs.
2019-12-17 22:23:07 +02:00
log-level-file=[LOG-LEVEL-FILE]
Added the log-level-stderr option. Controls whether console log messages are sent to stderr or stdout. By default this is set to warn which represents a change in behavior from previous versions, even though it may be more intuitive. Setting log-level-stderr=off will preserve the old behavior. Suggested by Sascha Biberhofer.
2016-10-05 15:09:30 +02:00
log-level-stderr=off
Storage and IO layer refactor: Refactor storage layer to allow for new repository filesystems using drivers. (Reviewed by Cynthia Shang.) Refactor IO layer to allow for new compression formats, checksum types, and other capabilities using filters. (Reviewed by Cynthia Shang.)
2017-06-09 23:51:41 +02:00
log-path=[TEST_PATH]/db-master/log
Integration test improvements for disk and memory efficiency. Set log-level-file=off when more that one test will run. In this case is it impossible to see the logs anyway since they will be automatically cleaned up after the test. This improves performance pretty dramatically since trace-level logging is expensive. If a singe integration test is run then log-level-file is trace by default but can be changed with the --log-level-test-file option. Reduce buffer-size to 64k to save memory during testing and allow more processes to run in parallel. Update log replacement rules so that these options can change without affecting expect logs.
2019-12-17 22:23:07 +02:00
log-subprocess=[LOG-SUBPROCESS]
Set log-timestamp=n for integration tests. This means less filtering of logs needs to be done and new timestamps can be added without adding new filters.
2018-05-11 17:24:38 +02:00
log-timestamp=n
Storage and IO layer refactor: Refactor storage layer to allow for new repository filesystems using drivers. (Reviewed by Cynthia Shang.) Refactor IO layer to allow for new compression formats, checksum types, and other capabilities using filters. (Reviewed by Cynthia Shang.)
2017-06-09 23:51:41 +02:00
protocol-timeout=60
Rename db-* options to pg-* and backup-* options to repo-* to improve consistency. * repo-* options are now indexed although only one is allowed. * List deprecated option names in documentation and command-line help.
2018-02-04 01:27:38 +02:00
repo1-path=[TEST_PATH]/db-master/repo
Consolidate archive-push tests into new archive test.
2017-06-21 14:38:53 +02:00
spool-path=[TEST_PATH]/db-master/spool
New simpler configuration and consistent project/exe/path naming. * The repo-path option now always refers to the repository where backups and archive are stored, whether local or remote, so the repo-remote-path option has been removed. The new spool-path option can be used to define a location for queueing WAL segments when archiving asynchronously. Otherwise, a local repository is no longer required. * Implemented a new config format which should be far simpler to use. See the User Guide and Configuration Reference for details but for a simple configuration all options can now be placed in the stanza section. Options that are shared between stanzas can be placed in the [global] section. More complex configurations can still make use of command sections though this should be a rare use case. * The default configuration filename is now pgbackrest.conf instead of pg_backrest.conf. This was done for consistency with other naming changes but also to prevent old config files from being loaded accidentally. * The default repository name was changed from /var/lib/backup to /var/lib/pgbackrest. * Lock files are now stored in /tmp/pgbackrest by default. These days /run/pgbackrest would be the preferred location but that would require init scripts which are not part of this release. The lock-path option can be used to configure the lock directory. * Log files are now stored in /var/log/pgbackrest by default and no longer have the date appended so they can be managed with logrotate. The log-path option can be used to configure the lock directory. * Executable filename changed from pg_backrest to pgbackrest.
2016-04-14 15:30:54 +02:00
[global:backup]
archive-copy=y
start-fast=y
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
Remove integration tests that are now covered in the unit tests. Most of these tests are just checking that errors are thrown when required. These are well covered in various unit tests. The "cannot resume" tests are also well covered in the backup unit tests. Finally, config warnings are well covered in the config unit tests. There is more to be done here, but this accounts for the low-hanging fruit.
2019-12-18 03:14:45 +02:00
+ supplemental file: [TEST_PATH]/db-master/repo/backup/db/[BACKUP-DIFF-1]/backup.manifest
Rename test paths for clarity. This was worked out as part of the test suite refactor [c8f806a] but not committed with it because of the large number of expect logs changes involved. Keeping them separate made it easier to audit the changes in the refactor.
2016-06-24 16:54:31 +02:00
-----------------------------------------------------------------------------------------
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
[backrest]
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
backrest-format=5
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
backrest-version="[VERSION-1]"
[backup]
Remove integration tests that are now covered in the unit tests. Most of these tests are just checking that errors are thrown when required. These are well covered in various unit tests. The "cannot resume" tests are also well covered in the backup unit tests. Finally, config warnings are well covered in the config unit tests. There is more to be done here, but this accounts for the low-hanging fruit.
2019-12-18 03:14:45 +02:00
backup-label="[BACKUP-DIFF-1]"
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
backup-prior="[BACKUP-FULL-2]"
backup-timestamp-copy-start=[TIMESTAMP]
backup-timestamp-start=[TIMESTAMP]
backup-timestamp-stop=[TIMESTAMP]
backup-type="diff"
[backup:db]
Database version refactoring. * Refactor db version constants into a separate module. * Update synthetic backup tests to PostgreSQL 9.4.
2016-08-12 04:35:24 +02:00
db-catalog-version=201409291
db-control-version=942
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
db-id=1
Automate generation of WAL and pg_control test files. The existing static files would not work with 32-bit or big-endian systems so create functions to generate these files dynamically rather than creating a bunch of new static files.
2017-11-19 03:02:54 +02:00
db-system-id=1000000000000000094
Database version refactoring. * Refactor db version constants into a separate module. * Update synthetic backup tests to PostgreSQL 9.4.
2016-08-12 04:35:24 +02:00
db-version="9.4"
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
[backup:option]
Set archive-check option in manifest correctly when offline. Archive check does not run when in offline backup mode but the option was set to true in the manifest. It's harmless since these options are informational only but it could cause confusion when debugging.
2019-11-28 15:27:21 +02:00
option-archive-check=false
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
option-archive-copy=true
Backup from a standby cluster. A connection to the primary cluster is still required to start/stop the backup and copy files that are not replicated, but the vast majority of files are copied from the standby in order to reduce load on the master.
2016-08-25 17:25:46 +02:00
option-backup-standby=false
Integration test improvements for disk and memory efficiency. Set log-level-file=off when more that one test will run. In this case is it impossible to see the logs anyway since they will be automatically cleaned up after the test. This improves performance pretty dramatically since trace-level logging is expensive. If a singe integration test is run then log-level-file is trace by default but can be changed with the --log-level-test-file option. Reduce buffer-size to 64k to save memory during testing and allow more processes to run in parallel. Update log replacement rules so that these options can change without affecting expect logs.
2019-12-17 22:23:07 +02:00
option-buffer-size=[BUFFER-SIZE]
Added the --checksum-page option. This option allows pgBackRest to validate page checksums in data files when checksums are enabled on PostgreSQL >= 9.3. Note that this functionality requires a C library which may not initially be available in OS packages. The option will automatically be enabled when the library is present and checksums are enabled on the cluster.
2016-12-13 01:54:07 +02:00
option-checksum-page=true
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
option-compress=false
Add additional options to backup.manifest for debugging purposes. Add the buffer-size, compress-level, compress-level-network, and process-max options to the backup:option section in backup.manifest to aid in debugging. It may also make sense to propagate these options up to backup.info so they can be displayed in the info command, but for now this is deemed sufficient. Contributed by blogh.
2019-03-10 11:03:52 +02:00
option-compress-level=3
option-compress-level-network=3
Add infrastructure for multiple compression type support. Add compress-type option and deprecate compress option. Since the compress option is boolean it won't work with multiple compression types. Add logic to cfgLoadUpdateOption() to update compress-type if it is not set directly. The compress option should no longer be referenced outside the cfgLoadUpdateOption() function. Add common/compress/helper module to contain interface functions that work with multiple compression types. Code outside this module should no longer call specific compression drivers, though it may be OK to reference a specific compression type using the new interface (e.g., saving backup history files in gz format). Unit tests only test compression using the gz format because other formats may not be available in all builds. It is the job of integration tests to exercise all compression types. Additional compression types will be added in future commits.
2020-03-06 21:41:03 +02:00
option-compress-type="none"
Add checksum delta for incremental backups. Use checksums rather than timestamps to determine if files have changed. This is useful in cases where the timestamps may not be trustworthy, e.g. when performing an incremental after failing over to a standby. If checksum delta is enabled then checksums will be used for verification of resumed backups, even if they are full. Resumes have always used checksums to verify the files in the repository, enabling delta performs checksums on the database files as well. Note that the user must manually enable this feature in cases were it would be useful or just keep in enabled all the time. A future commit will address automatically enabling the feature in cases where it seems likely to be useful. Contributed by Cynthia Shang.
2018-09-19 17:12:45 +02:00
option-delta=true
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
option-hardlink=false
Closed #167: Rename `--no-start-stop` option to `--no-online`.
2016-02-20 02:32:02 +02:00
option-online=false
Add additional options to backup.manifest for debugging purposes. Add the buffer-size, compress-level, compress-level-network, and process-max options to the backup:option section in backup.manifest to aid in debugging. It may also make sense to propagate these options up to backup.info so they can be displayed in the info command, but for now this is deemed sufficient. Contributed by blogh.
2019-03-10 11:03:52 +02:00
option-process-max=1
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
[backup:target]
Rename test paths for clarity. This was worked out as part of the test suite refactor [c8f806a] but not committed with it because of the large number of expect logs changes involved. Keeping them separate made it easier to audit the changes in the refactor.
2016-06-24 16:54:31 +02:00
pg_data={"path":"[TEST_PATH]/db-master/db/base","type":"path"}
pg_tblspc/1={"path":"[TEST_PATH]/db-master/db/tablespace/ts1","tablespace-id":"1","tablespace-name":"ts1","type":"link"}
pg_tblspc/2={"path":"[TEST_PATH]/db-master/db/tablespace/ts2","tablespace-id":"2","tablespace-name":"ts2","type":"link"}
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
[target:file]
Fixed a issue where tablespaces were copied from the master during standby backup.
2016-09-04 15:19:44 +02:00
pg_data/PG_VERSION={"checksum":"184473f470864e067ee3a22e64b47b0a1c356f29","master":true,"reference":"[BACKUP-FULL-2]","size":[SIZE],"timestamp":[TIMESTAMP-1]}
pg_data/badchecksum.txt={"checksum":"f927212cd08d11a42a666b2f04235398e9ceeb51","master":true,"size":[SIZE],"timestamp":[TIMESTAMP-1]}
Added the --checksum-page option. This option allows pgBackRest to validate page checksums in data files when checksums are enabled on PostgreSQL >= 9.3. Note that this functionality requires a C library which may not initially be available in OS packages. The option will automatically be enabled when the library is present and checksums are enabled on the cluster.
2016-12-13 01:54:07 +02:00
pg_data/base/1/12000={"checksum":"22c98d248ff548311eda88559e4a8405ed77c003","checksum-page":true,"reference":"[BACKUP-FULL-2]","size":[SIZE],"timestamp":[TIMESTAMP-1]}
Database version refactoring. * Refactor db version constants into a separate module. * Update synthetic backup tests to PostgreSQL 9.4.
2016-08-12 04:35:24 +02:00
pg_data/base/1/PG_VERSION={"checksum":"184473f470864e067ee3a22e64b47b0a1c356f29","mode":"0660","reference":"[BACKUP-FULL-2]","size":[SIZE],"timestamp":[TIMESTAMP-1]}
Added the --checksum-page option. This option allows pgBackRest to validate page checksums in data files when checksums are enabled on PostgreSQL >= 9.3. Note that this functionality requires a C library which may not initially be available in OS packages. The option will automatically be enabled when the library is present and checksums are enabled on the cluster.
2016-12-13 01:54:07 +02:00
pg_data/base/16384/17000={"checksum":"e0101dd8ffb910c9c202ca35b5f828bcb9697bed","checksum-page":false,"checksum-page-error":[1],"reference":"[BACKUP-FULL-2]","size":[SIZE],"timestamp":[TIMESTAMP-1]}
Database version refactoring. * Refactor db version constants into a separate module. * Update synthetic backup tests to PostgreSQL 9.4.
2016-08-12 04:35:24 +02:00
pg_data/base/16384/PG_VERSION={"checksum":"184473f470864e067ee3a22e64b47b0a1c356f29","reference":"[BACKUP-FULL-2]","size":[SIZE],"timestamp":[TIMESTAMP-1]}
Fix incorrect offline upper bound for ignoring page checksum errors. For offline backups the upper bound was being set to 0x0000FFFF0000FFFF rather than UINT64_MAX. This meant that page checksum errors might be ignored for databases with a lot of past WAL in offline mode. Online mode is not affected since the upper bound is retrieved from pg_start_backup().
2019-07-11 15:13:56 +02:00
pg_data/base/32768/33000={"checksum":"7a16d165e4775f7c92e8cdf60c0af57313f0bf90","checksum-page":true,"reference":"[BACKUP-FULL-2]","size":[SIZE],"timestamp":[TIMESTAMP-1]}
pg_data/base/32768/33000.32767={"checksum":"6e99b589e550e68e934fd235ccba59fe5b592a9e","checksum-page":true,"reference":"[BACKUP-FULL-2]","size":[SIZE],"timestamp":[TIMESTAMP-1]}
Added the --checksum-page option. This option allows pgBackRest to validate page checksums in data files when checksums are enabled on PostgreSQL >= 9.3. Note that this functionality requires a C library which may not initially be available in OS packages. The option will automatically be enabled when the library is present and checksums are enabled on the cluster.
2016-12-13 01:54:07 +02:00
pg_data/base/32768/33001={"checksum":"6bf316f11d28c28914ea9be92c00de9bea6d9a6b","checksum-page":false,"checksum-page-error":[0,[3,5],7],"reference":"[BACKUP-FULL-2]","size":[SIZE],"timestamp":[TIMESTAMP-1]}
Fix incorrect offline upper bound for ignoring page checksum errors. For offline backups the upper bound was being set to 0x0000FFFF0000FFFF rather than UINT64_MAX. This meant that page checksum errors might be ignored for databases with a lot of past WAL in offline mode. Online mode is not affected since the upper bound is retrieved from pg_start_backup().
2019-07-11 15:13:56 +02:00
pg_data/base/32768/44000_init={"checksum":"7a16d165e4775f7c92e8cdf60c0af57313f0bf90","checksum-page":true,"reference":"[BACKUP-FULL-2]","size":[SIZE],"timestamp":[TIMESTAMP-1]}
Database version refactoring. * Refactor db version constants into a separate module. * Update synthetic backup tests to PostgreSQL 9.4.
2016-08-12 04:35:24 +02:00
pg_data/base/32768/PG_VERSION={"checksum":"184473f470864e067ee3a22e64b47b0a1c356f29","reference":"[BACKUP-FULL-2]","size":[SIZE],"timestamp":[TIMESTAMP-1]}
Add checksum delta for incremental backups. Use checksums rather than timestamps to determine if files have changed. This is useful in cases where the timestamps may not be trustworthy, e.g. when performing an incremental after failing over to a standby. If checksum delta is enabled then checksums will be used for verification of resumed backups, even if they are full. Resumes have always used checksums to verify the files in the repository, enabling delta performs checksums on the database files as well. Note that the user must manually enable this feature in cases were it would be useful or just keep in enabled all the time. A future commit will address automatically enabling the feature in cases where it seems likely to be useful. Contributed by Cynthia Shang.
2018-09-19 17:12:45 +02:00
pg_data/changecontent.txt={"checksum":"a094d94583e209556d03c3c5da33131a065f1689","master":true,"size":[SIZE],"timestamp":[TIMESTAMP-1]}
pg_data/changetime.txt={"checksum":"88087292ed82e26f3eb824d0bffc05ccf7a30f8d","master":true,"reference":"[BACKUP-FULL-2]","size":[SIZE],"timestamp":[TIMESTAMP-2]}
Support configurable WAL segment size. PostgreSQL 11 introduces configurable WAL segment sizes, from 1MB to 1GB. There are two areas that needed to be updated to support this: building the archive-get queue and checking that WAL has been archived after a backup. Both operations require the WAL segment size to properly build a list. Checking the archive after a backup is still implemented in Perl and has an active database connection, so just get the WAL segment size from the database. The archive-get command does not have a connection to the database, so get the WAL segment size from pg_control instead. This requires a deeper inspection of pg_control than has been done in the past, so it seemed best to copy the relevant data structures from each version of PostgreSQL and build a generic interface layer to address them. While this approach is a bit verbose, it has the advantage of being relatively simple, and can easily be updated for new versions of PostgreSQL. Since the integration tests generate pg_control files for testing, teach Perl how to generate files with the correct offsets for both 32-bit and 64-bit architectures.
2018-09-25 11:24:42 +02:00
pg_data/global/pg_control={"checksum":"4c77c900f7af0d9ab13fa9982051a42e0b637f6c","master":true,"reference":"[BACKUP-FULL-2]","size":[SIZE],"timestamp":[TIMESTAMP-2]}
Fixed a issue where tablespaces were copied from the master during standby backup.
2016-09-04 15:19:44 +02:00
pg_data/pg_stat/global.stat={"checksum":"e350d5ce0153f3e22d5db21cf2a4eff00f3ee877","master":true,"reference":"[BACKUP-FULL-2]","size":[SIZE],"timestamp":[TIMESTAMP-2]}
pg_data/postgresql.conf={"checksum":"6721d92c9fcdf4248acff1f9a1377127d9064807","master":true,"reference":"[BACKUP-FULL-2]","size":[SIZE],"timestamp":[TIMESTAMP-2]}
Use minio for integration testing. ScalityS3 has not received any maintenance in years and is slow to start which is bad for testing. Replace it with minio which starts quickly and ships as a single executable or a tiny container. Minio has stricter limits on allowable characters but should still provide enough coverage to show that our encoding is working correctly. This commit also includes the upgrade to openssl 1.1.1 in the Ubuntu 18.04 container.
2019-07-03 04:20:35 +02:00
pg_data/special-!_.*'()&!@;:+,?={"master":true,"reference":"[BACKUP-FULL-2]","size":[SIZE],"timestamp":[TIMESTAMP-1]}
Allow zero-size files in backup manifest to reference a prior manifest regardless of timestamp delta. Contributed by Cynthia Shang.
2018-08-24 22:50:33 +02:00
pg_data/zero_from_start={"master":true,"reference":"[BACKUP-FULL-2]","size":[SIZE],"timestamp":[TIMESTAMP-1]}
Add zero-length file to mock/all test.
2018-07-05 21:40:50 +02:00
pg_data/zerosize.txt={"master":true,"size":[SIZE],"timestamp":[TIMESTAMP-2]}
Added the --checksum-page option. This option allows pgBackRest to validate page checksums in data files when checksums are enabled on PostgreSQL >= 9.3. Note that this functionality requires a C library which may not initially be available in OS packages. The option will automatically be enabled when the library is present and checksums are enabled on the cluster.
2016-12-13 01:54:07 +02:00
pg_tblspc/1/[TS_PATH-1]/16384/tablespace1.txt={"checksum":"d85de07d6421d90aa9191c11c889bfde43680f0f","checksum-page":false,"size":[SIZE],"timestamp":[TIMESTAMP-1]}
pg_tblspc/2/[TS_PATH-1]/32768/tablespace2.txt={"checksum":"dc7f76e43c46101b47acc55ae4d593a9e6983578","checksum-page":false,"size":[SIZE],"timestamp":[TIMESTAMP-1]}
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
[target:file:default]
Continuous integration using travis-ci.
2017-02-21 15:59:23 +02:00
group="[GROUP-1]"
Fixed a issue where tablespaces were copied from the master during standby backup.
2016-09-04 15:19:44 +02:00
master=false
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
mode="0600"
Regression tests can now be run as any properly-configured user, not just vagrant.
2016-10-01 19:39:44 +02:00
user="[USER-1]"
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
[target:link]
Rename test paths for clarity. This was worked out as part of the test suite refactor [c8f806a] but not committed with it because of the large number of expect logs changes involved. Keeping them separate made it easier to audit the changes in the refactor.
2016-06-24 16:54:31 +02:00
pg_data/pg_tblspc/1={"destination":"[TEST_PATH]/db-master/db/tablespace/ts1"}
pg_data/pg_tblspc/2={"destination":"[TEST_PATH]/db-master/db/tablespace/ts2"}
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
[target:link:default]
Continuous integration using travis-ci.
2017-02-21 15:59:23 +02:00
group="[GROUP-1]"
Regression tests can now be run as any properly-configured user, not just vagrant.
2016-10-01 19:39:44 +02:00
user="[USER-1]"
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
[target:path]
Fixed an issue with invalid backup groups being set correctly on restore. If the backup cannot map a group to a name it stores the group in the manifest as false then uses either the owner of $PGDATA to set the group during restore or failing that the group of the current user. This logic was not working correctly because the selected group was overwriting the user on restore leaving the group undefined and the user incorrectly set to the group. (Reported by Jeff McCormick.)
2017-11-20 19:46:35 +02:00
pg_data={"mode":"0777"}
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
pg_data/base={}
Allow selective restore of databases from a cluster backup. This feature can result in major space and time savings when only specific databases are restored. Unrestored databases will not be accessible but must be manually dropped before they will be removed from the shared catalogue.
2016-05-11 15:21:39 +02:00
pg_data/base/1={}
pg_data/base/16384={}
pg_data/base/32768={}
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
pg_data/global={}
pg_data/pg_clog={}
Exclude directories during backup that are cleaned, recreated, or zeroed by PostgreSQL at startup. These include (depending on the version where they were introduced): pgsql_tmp, pg_dynshmem, pg_notify, pg_replslot, pg_serial, pg_snapshots, pg_stat_tmp, pg_subtrans. The postgresql.auto.conf.tmp file is now excluded in addition to files that were already excluded: backup_label.old, postmaster.opts, postmaster.pid, recovery.conf, recovery.done.
2016-08-16 15:35:16 +02:00
pg_data/pg_dynshmem={}
Allow arbitrary directories and/or files to be excluded from a backup. Misuse of this feature can lead to inconsistent backups so read the --exclude documentation carefully before using.
2018-08-27 21:51:05 +02:00
pg_data/pg_log={}
Exclude directories during backup that are cleaned, recreated, or zeroed by PostgreSQL at startup. These include (depending on the version where they were introduced): pgsql_tmp, pg_dynshmem, pg_notify, pg_replslot, pg_serial, pg_snapshots, pg_stat_tmp, pg_subtrans. The postgresql.auto.conf.tmp file is now excluded in addition to files that were already excluded: backup_label.old, postmaster.opts, postmaster.pid, recovery.conf, recovery.done.
2016-08-16 15:35:16 +02:00
pg_data/pg_notify={}
pg_data/pg_replslot={}
pg_data/pg_serial={}
pg_data/pg_snapshots={}
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
pg_data/pg_stat={}
Exclude directories during backup that are cleaned, recreated, or zeroed by PostgreSQL at startup. These include (depending on the version where they were introduced): pgsql_tmp, pg_dynshmem, pg_notify, pg_replslot, pg_serial, pg_snapshots, pg_stat_tmp, pg_subtrans. The postgresql.auto.conf.tmp file is now excluded in addition to files that were already excluded: backup_label.old, postmaster.opts, postmaster.pid, recovery.conf, recovery.done.
2016-08-16 15:35:16 +02:00
pg_data/pg_stat_tmp={}
pg_data/pg_subtrans={}
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
pg_data/pg_tblspc={}
Fixed an issue with invalid backup groups being set correctly on restore. If the backup cannot map a group to a name it stores the group in the manifest as false then uses either the owner of $PGDATA to set the group during restore or failing that the group of the current user. This logic was not working correctly because the selected group was overwriting the user on restore leaving the group undefined and the user incorrectly set to the group. (Reported by Jeff McCormick.)
2017-11-20 19:46:35 +02:00
pg_tblspc={"mode":"0777"}
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
pg_tblspc/1={}
pg_tblspc/1/[TS_PATH-1]={}
Allow selective restore of databases from a cluster backup. This feature can result in major space and time savings when only specific databases are restored. Unrestored databases will not be accessible but must be manually dropped before they will be removed from the shared catalogue.
2016-05-11 15:21:39 +02:00
pg_tblspc/1/[TS_PATH-1]/16384={}
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
pg_tblspc/2={}
pg_tblspc/2/[TS_PATH-1]={}
Allow selective restore of databases from a cluster backup. This feature can result in major space and time savings when only specific databases are restored. Unrestored databases will not be accessible but must be manually dropped before they will be removed from the shared catalogue.
2016-05-11 15:21:39 +02:00
pg_tblspc/2/[TS_PATH-1]/32768={}
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
[target:path:default]
Continuous integration using travis-ci.
2017-02-21 15:59:23 +02:00
group="[GROUP-1]"
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
mode="0700"
Regression tests can now be run as any properly-configured user, not just vagrant.
2016-10-01 19:39:44 +02:00
user="[USER-1]"
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
Move info file checksum to the end of the file. Putting the checksum at the beginning of the file made it impossible to stream the file out when saving. The entire file had to be held in memory while it was checksummed so the checksum could be written at the beginning. Instead place the checksum at the end. This does not break the existing Perl or C code since the read is not order dependent. There are no plans to improve the Perl code to take advantage of this change, but it will make the C implementation more efficient. Reviewed by Cynthia Shang.
2019-08-22 01:45:48 +02:00
[backrest]
backrest-checksum="[CHECKSUM]"
Rename test paths for clarity. This was worked out as part of the test suite refactor [c8f806a] but not committed with it because of the large number of expect logs changes involved. Keeping them separate made it easier to audit the changes in the refactor.
2016-06-24 16:54:31 +02:00
+ supplemental file: [TEST_PATH]/db-master/repo/backup/db/backup.info
---------------------------------------------------------------------
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
[backrest]
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
backrest-format=5
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
backrest-version="[VERSION-1]"
[backup:current]
Set archive-check option in manifest correctly when offline. Archive check does not run when in offline backup mode but the option was set to true in the manifest. It's harmless since these options are informational only but it could cause confusion when debugging.
2019-11-28 15:27:21 +02:00
[BACKUP-FULL-2]={"backrest-format":5,"backrest-version":"[VERSION-1]","backup-archive-start":null,"backup-archive-stop":null,"backup-info-repo-size":[SIZE],"backup-info-repo-size-delta":[DELTA],"backup-info-size":[SIZE],"backup-info-size-delta":[DELTA],"backup-timestamp-start":[TIMESTAMP],"backup-timestamp-stop":[TIMESTAMP],"backup-type":"full","db-id":1,"option-archive-check":false,"option-archive-copy":true,"option-backup-standby":false,"option-checksum-page":true,"option-compress":false,"option-hardlink":false,"option-online":false}
Remove integration tests that are now covered in the unit tests. Most of these tests are just checking that errors are thrown when required. These are well covered in various unit tests. The "cannot resume" tests are also well covered in the backup unit tests. Finally, config warnings are well covered in the config unit tests. There is more to be done here, but this accounts for the low-hanging fruit.
2019-12-18 03:14:45 +02:00
[BACKUP-INCR-2]={"backrest-format":5,"backrest-version":"[VERSION-1]","backup-archive-start":null,"backup-archive-stop":null,"backup-info-repo-size":[SIZE],"backup-info-repo-size-delta":[DELTA],"backup-info-size":[SIZE],"backup-info-size-delta":[DELTA],"backup-prior":"[BACKUP-FULL-2]","backup-reference":["[BACKUP-FULL-2]"],"backup-timestamp-start":[TIMESTAMP],"backup-timestamp-stop":[TIMESTAMP],"backup-type":"incr","db-id":1,"option-archive-check":false,"option-archive-copy":true,"option-backup-standby":false,"option-checksum-page":true,"option-compress":false,"option-hardlink":false,"option-online":false}
[BACKUP-DIFF-1]={"backrest-format":5,"backrest-version":"[VERSION-1]","backup-archive-start":null,"backup-archive-stop":null,"backup-info-repo-size":[SIZE],"backup-info-repo-size-delta":[DELTA],"backup-info-size":[SIZE],"backup-info-size-delta":[DELTA],"backup-prior":"[BACKUP-FULL-2]","backup-reference":["[BACKUP-FULL-2]"],"backup-timestamp-start":[TIMESTAMP],"backup-timestamp-stop":[TIMESTAMP],"backup-type":"diff","db-id":1,"option-archive-check":false,"option-archive-copy":true,"option-backup-standby":false,"option-checksum-page":true,"option-compress":false,"option-hardlink":false,"option-online":false}
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
[db]
Database version refactoring. * Refactor db version constants into a separate module. * Update synthetic backup tests to PostgreSQL 9.4.
2016-08-12 04:35:24 +02:00
db-catalog-version=201409291
db-control-version=942
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
db-id=1
Automate generation of WAL and pg_control test files. The existing static files would not work with 32-bit or big-endian systems so create functions to generate these files dynamically rather than creating a bunch of new static files.
2017-11-19 03:02:54 +02:00
db-system-id=1000000000000000094
Database version refactoring. * Refactor db version constants into a separate module. * Update synthetic backup tests to PostgreSQL 9.4.
2016-08-12 04:35:24 +02:00
db-version="9.4"
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
[db:history]
Automate generation of WAL and pg_control test files. The existing static files would not work with 32-bit or big-endian systems so create functions to generate these files dynamically rather than creating a bunch of new static files.
2017-11-19 03:02:54 +02:00
1={"db-catalog-version":201409291,"db-control-version":942,"db-system-id":1000000000000000094,"db-version":"9.4"}
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
Improve performance of info file load/save. Info files required three copies in memory to be loaded (the original string, an ini representation, and the final info object). Not only was this memory inefficient but the Ini object does sequential scans when searching for keys making large files very slow to load. This has not been an issue since archive.info and backup.info are very small, but it becomes a big deal when loading manifests with hundreds of thousands of files. Instead of holding copies of the data in memory, use a callback to deliver the ini data directly to the object when loading. Use a similar method for save to avoid having an intermediate copy. Save is a bit complex because sections/keys must be written in alpha order or older versions of pgBackRest will not calculate the correct checksum. Also move the load retry logic to helper functions rather than embedding it in the Info object. This allows for more flexibility in loading and ensures that stack traces will be available when developing unit tests. Reviewed by Cynthia Shang.
2019-09-06 19:48:28 +02:00
[backrest]
backrest-checksum="[CHECKSUM]"
Remove integration tests that are now covered in the unit tests. Most of these tests are just checking that errors are thrown when required. These are well covered in various unit tests. The "cannot resume" tests are also well covered in the backup unit tests. Finally, config warnings are well covered in the config unit tests. There is more to be done here, but this accounts for the low-hanging fruit.
2019-12-18 03:14:45 +02:00
restore, backup '[BACKUP-DIFF-1]', remap - remap all paths (db-master host)
> [CONTAINER-EXEC] db-master [BACKREST-BIN] --config=[TEST_PATH]/db-master/pgbackrest.conf --set=[BACKUP-DIFF-1] --stanza=db restore
Log testing can now be enabled for certain deterministic tests. This works by comparing the generated logs against a previous copy. Currently only enabled for the backup/synthetic tests.
2015-05-05 19:08:48 +02:00
------------------------------------------------------------------------------------------------------------------------------------
Remove compress/compress-level options from commands where unused. These commands (e.g. restore, archive-get) never used the compress options but allowed them to be passed on the command line. Now they will error when these options are passed on the command line. If these errors occur then remove the unused options.
2020-02-27 19:25:32 +02:00
P00 INFO: restore command begin [BACKREST-VERSION]: --buffer-size=[BUFFER-SIZE] --config=[TEST_PATH]/db-master/pgbackrest.conf --lock-path=[TEST_PATH]/db-master/lock --log-level-console=detail --log-level-file=[LOG-LEVEL-FILE] --log-level-stderr=off --log-path=[TEST_PATH]/db-master/log[] --no-log-timestamp --pg1-path=[TEST_PATH]/db-master/db/base-2 --protocol-timeout=60 --repo1-path=[TEST_PATH]/db-master/repo --set=[BACKUP-DIFF-1] --stanza=db --tablespace-map=1=[TEST_PATH]/db-master/db/tablespace/ts1-2 --tablespace-map=2=[TEST_PATH]/db-master/db/tablespace/ts2-2
Remove integration tests that are now covered in the unit tests. Most of these tests are just checking that errors are thrown when required. These are well covered in various unit tests. The "cannot resume" tests are also well covered in the backup unit tests. Finally, config warnings are well covered in the config unit tests. There is more to be done here, but this accounts for the low-hanging fruit.
2019-12-18 03:14:45 +02:00
P00 INFO: restore backup set [BACKUP-DIFF-1]
The restore command is implemented entirely in C. For the most part this is a direct migration of the Perl code into C. There is one important behavioral change with regard to how file permissions are handled. The Perl code tried to set ownership as it was in the manifest even when running as an unprivileged user. This usually just led to errors and frustration. The C code works like this: If a restore is run as a non-root user (the typical scenario) then all files restored will belong to the user/group executing pgBackRest. If existing files are not owned by the executing user/group then an error will result if the ownership cannot be updated to the executing user/group. In that case the file ownership will need to be updated by a privileged user before the restore can be retried. If a restore is run as the root user then pgBackRest will attempt to recreate the ownership recorded in the manifest when the backup was made. Only user/group names are stored in the manifest so the same names must exist on the restore host for this to work. If the user/group name cannot be found locally then the user/group of the PostgreSQL data directory will be used and finally root if the data directory user/group cannot be mapped to a name. Reviewed by Cynthia Shang.
2019-09-26 13:52:02 +02:00
P00 INFO: remap data directory to '[TEST_PATH]/db-master/db/base-2'
P00 INFO: map tablespace 'pg_tblspc/1' to '[TEST_PATH]/db-master/db/tablespace/ts1-2'
P00 INFO: map tablespace 'pg_tblspc/2' to '[TEST_PATH]/db-master/db/tablespace/ts2-2'
P00 DETAIL: check '[TEST_PATH]/db-master/db/base-2' exists
P00 DETAIL: check '[TEST_PATH]/db-master/db/tablespace/ts1-2/[TS_PATH-1]' exists
P00 DETAIL: check '[TEST_PATH]/db-master/db/tablespace/ts2-2/[TS_PATH-1]' exists
P00 DETAIL: update mode for '[TEST_PATH]/db-master/db/base-2' to 0777
P00 DETAIL: create path '[TEST_PATH]/db-master/db/base-2/base'
P00 DETAIL: create path '[TEST_PATH]/db-master/db/base-2/base/1'
P00 DETAIL: create path '[TEST_PATH]/db-master/db/base-2/base/16384'
P00 DETAIL: create path '[TEST_PATH]/db-master/db/base-2/base/32768'
P00 DETAIL: create path '[TEST_PATH]/db-master/db/base-2/global'
P00 DETAIL: create path '[TEST_PATH]/db-master/db/base-2/pg_clog'
P00 DETAIL: create path '[TEST_PATH]/db-master/db/base-2/pg_dynshmem'
P00 DETAIL: create path '[TEST_PATH]/db-master/db/base-2/pg_log'
P00 DETAIL: create path '[TEST_PATH]/db-master/db/base-2/pg_notify'
P00 DETAIL: create path '[TEST_PATH]/db-master/db/base-2/pg_replslot'
P00 DETAIL: create path '[TEST_PATH]/db-master/db/base-2/pg_serial'
P00 DETAIL: create path '[TEST_PATH]/db-master/db/base-2/pg_snapshots'
P00 DETAIL: create path '[TEST_PATH]/db-master/db/base-2/pg_stat'
P00 DETAIL: create path '[TEST_PATH]/db-master/db/base-2/pg_stat_tmp'
P00 DETAIL: create path '[TEST_PATH]/db-master/db/base-2/pg_subtrans'
P00 DETAIL: create path '[TEST_PATH]/db-master/db/base-2/pg_tblspc'
P00 DETAIL: create symlink '[TEST_PATH]/db-master/db/base-2/pg_tblspc/1' to '[TEST_PATH]/db-master/db/tablespace/ts1-2'
P00 DETAIL: create path '[TEST_PATH]/db-master/db/base-2/pg_tblspc/1/[TS_PATH-1]/16384'
P00 DETAIL: create symlink '[TEST_PATH]/db-master/db/base-2/pg_tblspc/2' to '[TEST_PATH]/db-master/db/tablespace/ts2-2'
P00 DETAIL: create path '[TEST_PATH]/db-master/db/base-2/pg_tblspc/2/[TS_PATH-1]/32768'
Exclude temporary and unlogged relation (table/index) files from backup. Implemented using the same logic as the patches adding this feature to PostgreSQL, 8694cc96 and 920a5e50. Temporary relation exclusion is enabled in PostgreSQL ≥ 9.0. Unlogged relation exclusion is enabled in PostgreSQL ≥ 9.1, where the feature was introduced. Contributed by Cynthia Shang.
2018-07-31 00:53:34 +02:00
P01 INFO: restore file [TEST_PATH]/db-master/db/base-2/base/32768/33001 (64KB, 33%) checksum 6bf316f11d28c28914ea9be92c00de9bea6d9a6b
Fix incorrect offline upper bound for ignoring page checksum errors. For offline backups the upper bound was being set to 0x0000FFFF0000FFFF rather than UINT64_MAX. This meant that page checksum errors might be ignored for databases with a lot of past WAL in offline mode. Online mode is not affected since the upper bound is retrieved from pg_start_backup().
2019-07-11 15:13:56 +02:00
P01 INFO: restore file [TEST_PATH]/db-master/db/base-2/base/32768/44000_init (32KB, 49%) checksum 7a16d165e4775f7c92e8cdf60c0af57313f0bf90
P01 INFO: restore file [TEST_PATH]/db-master/db/base-2/base/32768/33000.32767 (32KB, 66%) checksum 6e99b589e550e68e934fd235ccba59fe5b592a9e
P01 INFO: restore file [TEST_PATH]/db-master/db/base-2/base/32768/33000 (32KB, 83%) checksum 7a16d165e4775f7c92e8cdf60c0af57313f0bf90
Exclude temporary and unlogged relation (table/index) files from backup. Implemented using the same logic as the patches adding this feature to PostgreSQL, 8694cc96 and 920a5e50. Temporary relation exclusion is enabled in PostgreSQL ≥ 9.0. Unlogged relation exclusion is enabled in PostgreSQL ≥ 9.1, where the feature was introduced. Contributed by Cynthia Shang.
2018-07-31 00:53:34 +02:00
P01 INFO: restore file [TEST_PATH]/db-master/db/base-2/base/16384/17000 (16KB, 91%) checksum e0101dd8ffb910c9c202ca35b5f828bcb9697bed
Support configurable WAL segment size. PostgreSQL 11 introduces configurable WAL segment sizes, from 1MB to 1GB. There are two areas that needed to be updated to support this: building the archive-get queue and checking that WAL has been archived after a backup. Both operations require the WAL segment size to properly build a list. Checking the archive after a backup is still implemented in Perl and has an active database connection, so just get the WAL segment size from the database. The archive-get command does not have a connection to the database, so get the WAL segment size from pg_control instead. This requires a deeper inspection of pg_control than has been done in the past, so it seemed best to copy the relevant data structures from each version of PostgreSQL and build a generic interface layer to address them. While this approach is a bit verbose, it has the advantage of being relatively simple, and can easily be updated for new versions of PostgreSQL. Since the integration tests generate pg_control files for testing, teach Perl how to generate files with the correct offsets for both 32-bit and 64-bit architectures.
2018-09-25 11:24:42 +02:00
P01 INFO: restore file [TEST_PATH]/db-master/db/base-2/global/pg_control.pgbackrest.tmp (8KB, 95%) checksum 4c77c900f7af0d9ab13fa9982051a42e0b637f6c
Add more realistic data files to synthetic backup and restore tests.
2016-12-10 16:13:55 +02:00
P01 INFO: restore file [TEST_PATH]/db-master/db/base-2/base/1/12000 (8KB, 99%) checksum 22c98d248ff548311eda88559e4a8405ed77c003
For simplicity, the pg_control file is now copied with the rest of the files instead of by itself of at the end of the process. The backup command does not require this behavior and the restore copies to a temporary file which is renamed at the end of the restore.
2016-11-30 21:23:05 +02:00
P01 INFO: restore file [TEST_PATH]/db-master/db/base-2/postgresql.conf (21B, 99%) checksum 6721d92c9fcdf4248acff1f9a1377127d9064807
P01 INFO: restore file [TEST_PATH]/db-master/db/base-2/badchecksum.txt (11B, 99%) checksum f927212cd08d11a42a666b2f04235398e9ceeb51
Add checksum delta for incremental backups. Use checksums rather than timestamps to determine if files have changed. This is useful in cases where the timestamps may not be trustworthy, e.g. when performing an incremental after failing over to a standby. If checksum delta is enabled then checksums will be used for verification of resumed backups, even if they are full. Resumes have always used checksums to verify the files in the repository, enabling delta performs checksums on the database files as well. Note that the user must manually enable this feature in cases were it would be useful or just keep in enabled all the time. A future commit will address automatically enabling the feature in cases where it seems likely to be useful. Contributed by Cynthia Shang.
2018-09-19 17:12:45 +02:00
P01 INFO: restore file [TEST_PATH]/db-master/db/base-2/changecontent.txt (7B, 99%) checksum a094d94583e209556d03c3c5da33131a065f1689
For simplicity, the pg_control file is now copied with the rest of the files instead of by itself of at the end of the process. The backup command does not require this behavior and the restore copies to a temporary file which is renamed at the end of the restore.
2016-11-30 21:23:05 +02:00
P01 INFO: restore file [TEST_PATH]/db-master/db/base-2/pg_stat/global.stat (5B, 99%) checksum e350d5ce0153f3e22d5db21cf2a4eff00f3ee877
Add checksum delta for incremental backups. Use checksums rather than timestamps to determine if files have changed. This is useful in cases where the timestamps may not be trustworthy, e.g. when performing an incremental after failing over to a standby. If checksum delta is enabled then checksums will be used for verification of resumed backups, even if they are full. Resumes have always used checksums to verify the files in the repository, enabling delta performs checksums on the database files as well. Note that the user must manually enable this feature in cases were it would be useful or just keep in enabled all the time. A future commit will address automatically enabling the feature in cases where it seems likely to be useful. Contributed by Cynthia Shang.
2018-09-19 17:12:45 +02:00
P01 INFO: restore file [TEST_PATH]/db-master/db/base-2/changetime.txt (4B, 99%) checksum 88087292ed82e26f3eb824d0bffc05ccf7a30f8d
For simplicity, the pg_control file is now copied with the rest of the files instead of by itself of at the end of the process. The backup command does not require this behavior and the restore copies to a temporary file which is renamed at the end of the restore.
2016-11-30 21:23:05 +02:00
P01 INFO: restore file [TEST_PATH]/db-master/db/base-2/base/32768/PG_VERSION (3B, 99%) checksum 184473f470864e067ee3a22e64b47b0a1c356f29
P01 INFO: restore file [TEST_PATH]/db-master/db/base-2/base/16384/PG_VERSION (3B, 99%) checksum 184473f470864e067ee3a22e64b47b0a1c356f29
P01 INFO: restore file [TEST_PATH]/db-master/db/base-2/base/1/PG_VERSION (3B, 99%) checksum 184473f470864e067ee3a22e64b47b0a1c356f29
P01 INFO: restore file [TEST_PATH]/db-master/db/base-2/PG_VERSION (3B, 99%) checksum 184473f470864e067ee3a22e64b47b0a1c356f29
Add zero-length file to mock/all test.
2018-07-05 21:40:50 +02:00
P01 INFO: restore file [TEST_PATH]/db-master/db/base-2/zerosize.txt (0B, 99%)
Allow zero-size files in backup manifest to reference a prior manifest regardless of timestamp delta. Contributed by Cynthia Shang.
2018-08-24 22:50:33 +02:00
P01 INFO: restore file [TEST_PATH]/db-master/db/base-2/zero_from_start (0B, 99%)
Use minio for integration testing. ScalityS3 has not received any maintenance in years and is slow to start which is bad for testing. Replace it with minio which starts quickly and ships as a single executable or a tiny container. Minio has stricter limits on allowable characters but should still provide enough coverage to show that our encoding is working correctly. This commit also includes the upgrade to openssl 1.1.1 in the Ubuntu 18.04 container.
2019-07-03 04:20:35 +02:00
P01 INFO: restore file [TEST_PATH]/db-master/db/base-2/special-!_.*'()&!@;:+,? (0B, 99%)
The restore command is implemented entirely in C. For the most part this is a direct migration of the Perl code into C. There is one important behavioral change with regard to how file permissions are handled. The Perl code tried to set ownership as it was in the manifest even when running as an unprivileged user. This usually just led to errors and frustration. The C code works like this: If a restore is run as a non-root user (the typical scenario) then all files restored will belong to the user/group executing pgBackRest. If existing files are not owned by the executing user/group then an error will result if the ownership cannot be updated to the executing user/group. In that case the file ownership will need to be updated by a privileged user before the restore can be retried. If a restore is run as the root user then pgBackRest will attempt to recreate the ownership recorded in the manifest when the backup was made. Only user/group names are stored in the manifest so the same names must exist on the restore host for this to work. If the user/group name cannot be found locally then the user/group of the PostgreSQL data directory will be used and finally root if the data directory user/group cannot be mapped to a name. Reviewed by Cynthia Shang.
2019-09-26 13:52:02 +02:00
P01 INFO: restore file [TEST_PATH]/db-master/db/base-2/pg_tblspc/1/[TS_PATH-1]/16384/tablespace1.txt (7B, 99%) checksum d85de07d6421d90aa9191c11c889bfde43680f0f
P01 INFO: restore file [TEST_PATH]/db-master/db/base-2/pg_tblspc/2/[TS_PATH-1]/32768/tablespace2.txt (7B, 100%) checksum dc7f76e43c46101b47acc55ae4d593a9e6983578
For simplicity, the pg_control file is now copied with the rest of the files instead of by itself of at the end of the process. The backup command does not require this behavior and the restore copies to a temporary file which is renamed at the end of the restore.
2016-11-30 21:23:05 +02:00
P00 INFO: write [TEST_PATH]/db-master/db/base-2/recovery.conf
The restore command is implemented entirely in C. For the most part this is a direct migration of the Perl code into C. There is one important behavioral change with regard to how file permissions are handled. The Perl code tried to set ownership as it was in the manifest even when running as an unprivileged user. This usually just led to errors and frustration. The C code works like this: If a restore is run as a non-root user (the typical scenario) then all files restored will belong to the user/group executing pgBackRest. If existing files are not owned by the executing user/group then an error will result if the ownership cannot be updated to the executing user/group. In that case the file ownership will need to be updated by a privileged user before the restore can be retried. If a restore is run as the root user then pgBackRest will attempt to recreate the ownership recorded in the manifest when the backup was made. Only user/group names are stored in the manifest so the same names must exist on the restore host for this to work. If the user/group name cannot be found locally then the user/group of the PostgreSQL data directory will be used and finally root if the data directory user/group cannot be mapped to a name. Reviewed by Cynthia Shang.
2019-09-26 13:52:02 +02:00
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base-2'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base-2/base'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base-2/base/1'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base-2/base/16384'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base-2/base/32768'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base-2/pg_clog'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base-2/pg_dynshmem'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base-2/pg_log'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base-2/pg_notify'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base-2/pg_replslot'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base-2/pg_serial'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base-2/pg_snapshots'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base-2/pg_stat'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base-2/pg_stat_tmp'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base-2/pg_subtrans'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base-2/pg_tblspc'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base-2/pg_tblspc/1'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base-2/pg_tblspc/1/[TS_PATH-1]'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base-2/pg_tblspc/1/[TS_PATH-1]/16384'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base-2/pg_tblspc/2'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base-2/pg_tblspc/2/[TS_PATH-1]'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base-2/pg_tblspc/2/[TS_PATH-1]/32768'
Storage and IO layer refactor: Refactor storage layer to allow for new repository filesystems using drivers. (Reviewed by Cynthia Shang.) Refactor IO layer to allow for new compression formats, checksum types, and other capabilities using filters. (Reviewed by Cynthia Shang.)
2017-06-09 23:51:41 +02:00
P00 INFO: restore global/pg_control (performed last to ensure aborted restores cannot be started)
The restore command is implemented entirely in C. For the most part this is a direct migration of the Perl code into C. There is one important behavioral change with regard to how file permissions are handled. The Perl code tried to set ownership as it was in the manifest even when running as an unprivileged user. This usually just led to errors and frustration. The C code works like this: If a restore is run as a non-root user (the typical scenario) then all files restored will belong to the user/group executing pgBackRest. If existing files are not owned by the executing user/group then an error will result if the ownership cannot be updated to the executing user/group. In that case the file ownership will need to be updated by a privileged user before the restore can be retried. If a restore is run as the root user then pgBackRest will attempt to recreate the ownership recorded in the manifest when the backup was made. Only user/group names are stored in the manifest so the same names must exist on the restore host for this to work. If the user/group name cannot be found locally then the user/group of the PostgreSQL data directory will be used and finally root if the data directory user/group cannot be mapped to a name. Reviewed by Cynthia Shang.
2019-09-26 13:52:02 +02:00
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base-2/global'
Added a predictable log entry to signal that a command has completed successfully. For example a backup ends successfully with: INFO: backup command end: completed successfully.
2016-12-04 21:44:53 +02:00
P00 INFO: restore command end: completed successfully
Log testing can now be enabled for certain deterministic tests. This works by comparing the generated logs against a previous copy. Currently only enabled for the backup/synthetic tests.
2015-05-05 19:08:48 +02:00
Rename test paths for clarity. This was worked out as part of the test suite refactor [c8f806a] but not committed with it because of the large number of expect logs changes involved. Keeping them separate made it easier to audit the changes in the refactor.
2016-06-24 16:54:31 +02:00
+ supplemental file: [TEST_PATH]/db-master/db/base-2/recovery.conf
------------------------------------------------------------------
PostgreSQL 12 support. Recovery settings are now written into postgresql.auto.conf instead of recovery.conf. Existing recovery_target* settings will be commented out to help avoid conflicts. A comment is added before recovery settings to identify them as written by pgBackRest since it is unclear how, in general, old settings will be removed. recovery.signal and standby.signal are automatically created based on the recovery settings.
2019-10-01 19:20:43 +02:00
# Recovery settings generated by pgBackRest restore on [TIMESTAMP]
Use path list in the backup manifest to do restore path syncs. Remove recursive path sync functionality since it is no longer used.
2018-05-01 17:05:37 +02:00
restore_command = '[BACKREST-BIN] --config=[TEST_PATH]/db-master/pgbackrest.conf --stanza=db archive-get %f "%p"'
Added vagrant configuration for Ubuntu 12.04. Changes to regression tests to work better with VMs.
2015-07-07 03:00:48 +02:00
Remove integration tests that are now covered in the unit tests. Most of these tests are just checking that errors are thrown when required. These are well covered in various unit tests. The "cannot resume" tests are also well covered in the backup unit tests. Finally, config warnings are well covered in the config unit tests. There is more to be done here, but this accounts for the low-hanging fruit.
2019-12-18 03:14:45 +02:00
restore delta, backup '[BACKUP-DIFF-1]', remap - ensure file in tblspc root remains after --delta (db-master host)
> [CONTAINER-EXEC] db-master [BACKREST-BIN] --config=[TEST_PATH]/db-master/pgbackrest.conf --delta --set=[BACKUP-DIFF-1] --stanza=db restore
Closed #58: Get catalog number for better tablespace copying When backing up and restoring tablespaces pgBackRest only operates on the subdirectory created for the version of PostgreSQL being run against. Since multiple versions can live in a tablespace (especially during a binary upgrade) this prevents too many files from being copied during a backup and other versions possibly being wiped out during a `--delta` restore. This only applies to PostgreSQL >= 9.0 -- before that only one PostgreSQL version could use a tablespace.
2016-02-02 21:33:15 +02:00
------------------------------------------------------------------------------------------------------------------------------------
Remove compress/compress-level options from commands where unused. These commands (e.g. restore, archive-get) never used the compress options but allowed them to be passed on the command line. Now they will error when these options are passed on the command line. If these errors occur then remove the unused options.
2020-02-27 19:25:32 +02:00
P00 INFO: restore command begin [BACKREST-VERSION]: --buffer-size=[BUFFER-SIZE] --config=[TEST_PATH]/db-master/pgbackrest.conf --delta --lock-path=[TEST_PATH]/db-master/lock --log-level-console=detail --log-level-file=[LOG-LEVEL-FILE] --log-level-stderr=off --log-path=[TEST_PATH]/db-master/log[] --no-log-timestamp --pg1-path=[TEST_PATH]/db-master/db/base-2 --protocol-timeout=60 --repo1-path=[TEST_PATH]/db-master/repo --set=[BACKUP-DIFF-1] --stanza=db --tablespace-map=1=[TEST_PATH]/db-master/db/tablespace/ts1-2 --tablespace-map=2=[TEST_PATH]/db-master/db/tablespace/ts2-2
Remove integration tests that are now covered in the unit tests. Most of these tests are just checking that errors are thrown when required. These are well covered in various unit tests. The "cannot resume" tests are also well covered in the backup unit tests. Finally, config warnings are well covered in the config unit tests. There is more to be done here, but this accounts for the low-hanging fruit.
2019-12-18 03:14:45 +02:00
P00 INFO: restore backup set [BACKUP-DIFF-1]
The restore command is implemented entirely in C. For the most part this is a direct migration of the Perl code into C. There is one important behavioral change with regard to how file permissions are handled. The Perl code tried to set ownership as it was in the manifest even when running as an unprivileged user. This usually just led to errors and frustration. The C code works like this: If a restore is run as a non-root user (the typical scenario) then all files restored will belong to the user/group executing pgBackRest. If existing files are not owned by the executing user/group then an error will result if the ownership cannot be updated to the executing user/group. In that case the file ownership will need to be updated by a privileged user before the restore can be retried. If a restore is run as the root user then pgBackRest will attempt to recreate the ownership recorded in the manifest when the backup was made. Only user/group names are stored in the manifest so the same names must exist on the restore host for this to work. If the user/group name cannot be found locally then the user/group of the PostgreSQL data directory will be used and finally root if the data directory user/group cannot be mapped to a name. Reviewed by Cynthia Shang.
2019-09-26 13:52:02 +02:00
P00 INFO: remap data directory to '[TEST_PATH]/db-master/db/base-2'
P00 INFO: map tablespace 'pg_tblspc/1' to '[TEST_PATH]/db-master/db/tablespace/ts1-2'
P00 INFO: map tablespace 'pg_tblspc/2' to '[TEST_PATH]/db-master/db/tablespace/ts2-2'
P00 DETAIL: check '[TEST_PATH]/db-master/db/base-2' exists
P00 DETAIL: check '[TEST_PATH]/db-master/db/tablespace/ts1-2/[TS_PATH-1]' exists
P00 DETAIL: check '[TEST_PATH]/db-master/db/tablespace/ts2-2/[TS_PATH-1]' exists
P00 DETAIL: remove 'global/pg_control' so cluster will not start if restore does not complete
P00 INFO: remove invalid files/links/paths from '[TEST_PATH]/db-master/db/base-2'
P00 DETAIL: remove invalid file '[TEST_PATH]/db-master/db/base-2/recovery.conf'
P00 INFO: remove invalid files/links/paths from '[TEST_PATH]/db-master/db/tablespace/ts1-2/[TS_PATH-1]'
P00 INFO: remove invalid files/links/paths from '[TEST_PATH]/db-master/db/tablespace/ts2-2/[TS_PATH-1]'
Exclude temporary and unlogged relation (table/index) files from backup. Implemented using the same logic as the patches adding this feature to PostgreSQL, 8694cc96 and 920a5e50. Temporary relation exclusion is enabled in PostgreSQL ≥ 9.0. Unlogged relation exclusion is enabled in PostgreSQL ≥ 9.1, where the feature was introduced. Contributed by Cynthia Shang.
2018-07-31 00:53:34 +02:00
P01 DETAIL: restore file [TEST_PATH]/db-master/db/base-2/base/32768/33001 - exists and matches backup (64KB, 33%) checksum 6bf316f11d28c28914ea9be92c00de9bea6d9a6b
Fix incorrect offline upper bound for ignoring page checksum errors. For offline backups the upper bound was being set to 0x0000FFFF0000FFFF rather than UINT64_MAX. This meant that page checksum errors might be ignored for databases with a lot of past WAL in offline mode. Online mode is not affected since the upper bound is retrieved from pg_start_backup().
2019-07-11 15:13:56 +02:00
P01 DETAIL: restore file [TEST_PATH]/db-master/db/base-2/base/32768/44000_init - exists and matches backup (32KB, 49%) checksum 7a16d165e4775f7c92e8cdf60c0af57313f0bf90
P01 DETAIL: restore file [TEST_PATH]/db-master/db/base-2/base/32768/33000.32767 - exists and matches backup (32KB, 66%) checksum 6e99b589e550e68e934fd235ccba59fe5b592a9e
P01 DETAIL: restore file [TEST_PATH]/db-master/db/base-2/base/32768/33000 - exists and matches backup (32KB, 83%) checksum 7a16d165e4775f7c92e8cdf60c0af57313f0bf90
Exclude temporary and unlogged relation (table/index) files from backup. Implemented using the same logic as the patches adding this feature to PostgreSQL, 8694cc96 and 920a5e50. Temporary relation exclusion is enabled in PostgreSQL ≥ 9.0. Unlogged relation exclusion is enabled in PostgreSQL ≥ 9.1, where the feature was introduced. Contributed by Cynthia Shang.
2018-07-31 00:53:34 +02:00
P01 DETAIL: restore file [TEST_PATH]/db-master/db/base-2/base/16384/17000 - exists and matches backup (16KB, 91%) checksum e0101dd8ffb910c9c202ca35b5f828bcb9697bed
Support configurable WAL segment size. PostgreSQL 11 introduces configurable WAL segment sizes, from 1MB to 1GB. There are two areas that needed to be updated to support this: building the archive-get queue and checking that WAL has been archived after a backup. Both operations require the WAL segment size to properly build a list. Checking the archive after a backup is still implemented in Perl and has an active database connection, so just get the WAL segment size from the database. The archive-get command does not have a connection to the database, so get the WAL segment size from pg_control instead. This requires a deeper inspection of pg_control than has been done in the past, so it seemed best to copy the relevant data structures from each version of PostgreSQL and build a generic interface layer to address them. While this approach is a bit verbose, it has the advantage of being relatively simple, and can easily be updated for new versions of PostgreSQL. Since the integration tests generate pg_control files for testing, teach Perl how to generate files with the correct offsets for both 32-bit and 64-bit architectures.
2018-09-25 11:24:42 +02:00
P01 INFO: restore file [TEST_PATH]/db-master/db/base-2/global/pg_control.pgbackrest.tmp (8KB, 95%) checksum 4c77c900f7af0d9ab13fa9982051a42e0b637f6c
Add more realistic data files to synthetic backup and restore tests.
2016-12-10 16:13:55 +02:00
P01 DETAIL: restore file [TEST_PATH]/db-master/db/base-2/base/1/12000 - exists and matches backup (8KB, 99%) checksum 22c98d248ff548311eda88559e4a8405ed77c003
For simplicity, the pg_control file is now copied with the rest of the files instead of by itself of at the end of the process. The backup command does not require this behavior and the restore copies to a temporary file which is renamed at the end of the restore.
2016-11-30 21:23:05 +02:00
P01 DETAIL: restore file [TEST_PATH]/db-master/db/base-2/postgresql.conf - exists and matches backup (21B, 99%) checksum 6721d92c9fcdf4248acff1f9a1377127d9064807
P01 DETAIL: restore file [TEST_PATH]/db-master/db/base-2/badchecksum.txt - exists and matches backup (11B, 99%) checksum f927212cd08d11a42a666b2f04235398e9ceeb51
Add checksum delta for incremental backups. Use checksums rather than timestamps to determine if files have changed. This is useful in cases where the timestamps may not be trustworthy, e.g. when performing an incremental after failing over to a standby. If checksum delta is enabled then checksums will be used for verification of resumed backups, even if they are full. Resumes have always used checksums to verify the files in the repository, enabling delta performs checksums on the database files as well. Note that the user must manually enable this feature in cases were it would be useful or just keep in enabled all the time. A future commit will address automatically enabling the feature in cases where it seems likely to be useful. Contributed by Cynthia Shang.
2018-09-19 17:12:45 +02:00
P01 DETAIL: restore file [TEST_PATH]/db-master/db/base-2/changecontent.txt - exists and matches backup (7B, 99%) checksum a094d94583e209556d03c3c5da33131a065f1689
For simplicity, the pg_control file is now copied with the rest of the files instead of by itself of at the end of the process. The backup command does not require this behavior and the restore copies to a temporary file which is renamed at the end of the restore.
2016-11-30 21:23:05 +02:00
P01 DETAIL: restore file [TEST_PATH]/db-master/db/base-2/pg_stat/global.stat - exists and matches backup (5B, 99%) checksum e350d5ce0153f3e22d5db21cf2a4eff00f3ee877
Add checksum delta for incremental backups. Use checksums rather than timestamps to determine if files have changed. This is useful in cases where the timestamps may not be trustworthy, e.g. when performing an incremental after failing over to a standby. If checksum delta is enabled then checksums will be used for verification of resumed backups, even if they are full. Resumes have always used checksums to verify the files in the repository, enabling delta performs checksums on the database files as well. Note that the user must manually enable this feature in cases were it would be useful or just keep in enabled all the time. A future commit will address automatically enabling the feature in cases where it seems likely to be useful. Contributed by Cynthia Shang.
2018-09-19 17:12:45 +02:00
P01 DETAIL: restore file [TEST_PATH]/db-master/db/base-2/changetime.txt - exists and matches backup (4B, 99%) checksum 88087292ed82e26f3eb824d0bffc05ccf7a30f8d
For simplicity, the pg_control file is now copied with the rest of the files instead of by itself of at the end of the process. The backup command does not require this behavior and the restore copies to a temporary file which is renamed at the end of the restore.
2016-11-30 21:23:05 +02:00
P01 DETAIL: restore file [TEST_PATH]/db-master/db/base-2/base/32768/PG_VERSION - exists and matches backup (3B, 99%) checksum 184473f470864e067ee3a22e64b47b0a1c356f29
P01 DETAIL: restore file [TEST_PATH]/db-master/db/base-2/base/16384/PG_VERSION - exists and matches backup (3B, 99%) checksum 184473f470864e067ee3a22e64b47b0a1c356f29
P01 DETAIL: restore file [TEST_PATH]/db-master/db/base-2/base/1/PG_VERSION - exists and matches backup (3B, 99%) checksum 184473f470864e067ee3a22e64b47b0a1c356f29
P01 DETAIL: restore file [TEST_PATH]/db-master/db/base-2/PG_VERSION - exists and matches backup (3B, 99%) checksum 184473f470864e067ee3a22e64b47b0a1c356f29
Add zero-length file to mock/all test.
2018-07-05 21:40:50 +02:00
P01 DETAIL: restore file [TEST_PATH]/db-master/db/base-2/zerosize.txt - exists and is zero size (0B, 99%)
Allow zero-size files in backup manifest to reference a prior manifest regardless of timestamp delta. Contributed by Cynthia Shang.
2018-08-24 22:50:33 +02:00
P01 DETAIL: restore file [TEST_PATH]/db-master/db/base-2/zero_from_start - exists and is zero size (0B, 99%)
Use minio for integration testing. ScalityS3 has not received any maintenance in years and is slow to start which is bad for testing. Replace it with minio which starts quickly and ships as a single executable or a tiny container. Minio has stricter limits on allowable characters but should still provide enough coverage to show that our encoding is working correctly. This commit also includes the upgrade to openssl 1.1.1 in the Ubuntu 18.04 container.
2019-07-03 04:20:35 +02:00
P01 DETAIL: restore file [TEST_PATH]/db-master/db/base-2/special-!_.*'()&!@;:+,? - exists and is zero size (0B, 99%)
The restore command is implemented entirely in C. For the most part this is a direct migration of the Perl code into C. There is one important behavioral change with regard to how file permissions are handled. The Perl code tried to set ownership as it was in the manifest even when running as an unprivileged user. This usually just led to errors and frustration. The C code works like this: If a restore is run as a non-root user (the typical scenario) then all files restored will belong to the user/group executing pgBackRest. If existing files are not owned by the executing user/group then an error will result if the ownership cannot be updated to the executing user/group. In that case the file ownership will need to be updated by a privileged user before the restore can be retried. If a restore is run as the root user then pgBackRest will attempt to recreate the ownership recorded in the manifest when the backup was made. Only user/group names are stored in the manifest so the same names must exist on the restore host for this to work. If the user/group name cannot be found locally then the user/group of the PostgreSQL data directory will be used and finally root if the data directory user/group cannot be mapped to a name. Reviewed by Cynthia Shang.
2019-09-26 13:52:02 +02:00
P01 DETAIL: restore file [TEST_PATH]/db-master/db/base-2/pg_tblspc/1/[TS_PATH-1]/16384/tablespace1.txt - exists and matches backup (7B, 99%) checksum d85de07d6421d90aa9191c11c889bfde43680f0f
P01 DETAIL: restore file [TEST_PATH]/db-master/db/base-2/pg_tblspc/2/[TS_PATH-1]/32768/tablespace2.txt - exists and matches backup (7B, 100%) checksum dc7f76e43c46101b47acc55ae4d593a9e6983578
For simplicity, the pg_control file is now copied with the rest of the files instead of by itself of at the end of the process. The backup command does not require this behavior and the restore copies to a temporary file which is renamed at the end of the restore.
2016-11-30 21:23:05 +02:00
P00 INFO: write [TEST_PATH]/db-master/db/base-2/recovery.conf
The restore command is implemented entirely in C. For the most part this is a direct migration of the Perl code into C. There is one important behavioral change with regard to how file permissions are handled. The Perl code tried to set ownership as it was in the manifest even when running as an unprivileged user. This usually just led to errors and frustration. The C code works like this: If a restore is run as a non-root user (the typical scenario) then all files restored will belong to the user/group executing pgBackRest. If existing files are not owned by the executing user/group then an error will result if the ownership cannot be updated to the executing user/group. In that case the file ownership will need to be updated by a privileged user before the restore can be retried. If a restore is run as the root user then pgBackRest will attempt to recreate the ownership recorded in the manifest when the backup was made. Only user/group names are stored in the manifest so the same names must exist on the restore host for this to work. If the user/group name cannot be found locally then the user/group of the PostgreSQL data directory will be used and finally root if the data directory user/group cannot be mapped to a name. Reviewed by Cynthia Shang.
2019-09-26 13:52:02 +02:00
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base-2'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base-2/base'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base-2/base/1'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base-2/base/16384'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base-2/base/32768'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base-2/pg_clog'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base-2/pg_dynshmem'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base-2/pg_log'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base-2/pg_notify'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base-2/pg_replslot'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base-2/pg_serial'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base-2/pg_snapshots'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base-2/pg_stat'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base-2/pg_stat_tmp'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base-2/pg_subtrans'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base-2/pg_tblspc'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base-2/pg_tblspc/1'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base-2/pg_tblspc/1/[TS_PATH-1]'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base-2/pg_tblspc/1/[TS_PATH-1]/16384'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base-2/pg_tblspc/2'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base-2/pg_tblspc/2/[TS_PATH-1]'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base-2/pg_tblspc/2/[TS_PATH-1]/32768'
Storage and IO layer refactor: Refactor storage layer to allow for new repository filesystems using drivers. (Reviewed by Cynthia Shang.) Refactor IO layer to allow for new compression formats, checksum types, and other capabilities using filters. (Reviewed by Cynthia Shang.)
2017-06-09 23:51:41 +02:00
P00 INFO: restore global/pg_control (performed last to ensure aborted restores cannot be started)
The restore command is implemented entirely in C. For the most part this is a direct migration of the Perl code into C. There is one important behavioral change with regard to how file permissions are handled. The Perl code tried to set ownership as it was in the manifest even when running as an unprivileged user. This usually just led to errors and frustration. The C code works like this: If a restore is run as a non-root user (the typical scenario) then all files restored will belong to the user/group executing pgBackRest. If existing files are not owned by the executing user/group then an error will result if the ownership cannot be updated to the executing user/group. In that case the file ownership will need to be updated by a privileged user before the restore can be retried. If a restore is run as the root user then pgBackRest will attempt to recreate the ownership recorded in the manifest when the backup was made. Only user/group names are stored in the manifest so the same names must exist on the restore host for this to work. If the user/group name cannot be found locally then the user/group of the PostgreSQL data directory will be used and finally root if the data directory user/group cannot be mapped to a name. Reviewed by Cynthia Shang.
2019-09-26 13:52:02 +02:00
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base-2/global'
Added a predictable log entry to signal that a command has completed successfully. For example a backup ends successfully with: INFO: backup command end: completed successfully.
2016-12-04 21:44:53 +02:00
P00 INFO: restore command end: completed successfully
Closed #58: Get catalog number for better tablespace copying When backing up and restoring tablespaces pgBackRest only operates on the subdirectory created for the version of PostgreSQL being run against. Since multiple versions can live in a tablespace (especially during a binary upgrade) this prevents too many files from being copied during a backup and other versions possibly being wiped out during a `--delta` restore. This only applies to PostgreSQL >= 9.0 -- before that only one PostgreSQL version could use a tablespace.
2016-02-02 21:33:15 +02:00
Rename test paths for clarity. This was worked out as part of the test suite refactor [c8f806a] but not committed with it because of the large number of expect logs changes involved. Keeping them separate made it easier to audit the changes in the refactor.
2016-06-24 16:54:31 +02:00
+ supplemental file: [TEST_PATH]/db-master/db/base-2/recovery.conf
------------------------------------------------------------------
PostgreSQL 12 support. Recovery settings are now written into postgresql.auto.conf instead of recovery.conf. Existing recovery_target* settings will be commented out to help avoid conflicts. A comment is added before recovery settings to identify them as written by pgBackRest since it is unclear how, in general, old settings will be removed. recovery.signal and standby.signal are automatically created based on the recovery settings.
2019-10-01 19:20:43 +02:00
# Recovery settings generated by pgBackRest restore on [TIMESTAMP]
Reduce log level for all expect tests to detail. The C code is designed to be efficient rather than deterministic at the debug log level. As we move more testing from integration to unit tests it makes less sense to try and maintain the expect logs at this log level. Most of the expect logs have already been moved to detail level but mock/all still had tests at debug level. Change the logging defaults in the config file and remove as many references to log-level-console as possible.
2019-05-23 00:23:44 +02:00
restore_command = '[BACKREST-BIN] --config=[TEST_PATH]/db-master/pgbackrest.conf --stanza=db archive-get %f "%p"'
Closed #58: Get catalog number for better tablespace copying When backing up and restoring tablespaces pgBackRest only operates on the subdirectory created for the version of PostgreSQL being run against. Since multiple versions can live in a tablespace (especially during a binary upgrade) this prevents too many files from being copied during a backup and other versions possibly being wiped out during a `--delta` restore. This only applies to PostgreSQL >= 9.0 -- before that only one PostgreSQL version could use a tablespace.
2016-02-02 21:33:15 +02:00
Major refactor of the test suite. * Make the code more modular and object-oriented. * Multiple Docker containers can now be created for a single test to simulate more realistic environments.
2016-06-24 14:12:58 +02:00
incr backup - add files and remove tablespace 2 (db-master host)
Reduce log level for all expect tests to detail. The C code is designed to be efficient rather than deterministic at the debug log level. As we move more testing from integration to unit tests it makes less sense to try and maintain the expect logs at this log level. Most of the expect logs have already been moved to detail level but mock/all still had tests at debug level. Change the logging defaults in the config file and remove as many references to log-level-console as possible.
2019-05-23 00:23:44 +02:00
> [CONTAINER-EXEC] db-master [BACKREST-BIN] --config=[TEST_PATH]/db-master/pgbackrest.conf --no-online --process-max=1 --stanza=db backup
Log testing can now be enabled for certain deterministic tests. This works by comparing the generated logs against a previous copy. Currently only enabled for the backup/synthetic tests.
2015-05-05 19:08:48 +02:00
------------------------------------------------------------------------------------------------------------------------------------
Add infrastructure for multiple compression type support. Add compress-type option and deprecate compress option. Since the compress option is boolean it won't work with multiple compression types. Add logic to cfgLoadUpdateOption() to update compress-type if it is not set directly. The compress option should no longer be referenced outside the cfgLoadUpdateOption() function. Add common/compress/helper module to contain interface functions that work with multiple compression types. Code outside this module should no longer call specific compression drivers, though it may be OK to reference a specific compression type using the new interface (e.g., saving backup history files in gz format). Unit tests only test compression using the gz format because other formats may not be available in all builds. It is the job of integration tests to exercise all compression types. Additional compression types will be added in future commits.
2020-03-06 21:41:03 +02:00
P00 INFO: backup command begin [BACKREST-VERSION]: --buffer-size=[BUFFER-SIZE] --compress-level=3 --compress-type=none --config=[TEST_PATH]/db-master/pgbackrest.conf --db-timeout=45 --lock-path=[TEST_PATH]/db-master/lock --log-level-console=detail --log-level-file=[LOG-LEVEL-FILE] --log-level-stderr=off --log-path=[TEST_PATH]/db-master/log[] --no-log-timestamp --no-online --pg1-path=[TEST_PATH]/db-master/db/base-2 --process-max=1 --protocol-timeout=60 --repo1-path=[TEST_PATH]/db-master/repo --stanza=db --start-fast
Add time-based retention for full backups. The --repo-retention-full-type option allows retention of full backups based on a time period, specified in days. The new option will default to 'count' and therefore will not affect current installations. Setting repo-retention-full-type to 'time' will allow the user to use a time period, in days, to indicate full backup retention. Using this method, a full backup can be expired only if the time the backup completed is older than the number of days set with repo-retention-full (calculated from the moment the 'expire' command is run) and at least one full backup meets the retention period. If archive retention has not been configured, then the default settings will expire archives that are prior to the oldest retained full backup. For example, if there are three full backups ending in times that are 25 days old (F1), 20 days old (F2) and 10 days old (F3), then if the full retention period is 15 days, then only F1 will be expired; F2 will be retained because F1 is not at least 15 days old.
2020-05-08 21:25:03 +02:00
P00 WARN: option 'repo1-retention-full' is not set for 'repo1-retention-full-type=count', the repository may run out of space
Rename retention-* options to repo-retention-*.
2018-02-19 22:07:24 +02:00
HINT: to retain full backups indefinitely (without warning), set option 'repo1-retention-full' to the maximum.
Remove integration tests that are now covered in the unit tests. Most of these tests are just checking that errors are thrown when required. These are well covered in various unit tests. The "cannot resume" tests are also well covered in the backup unit tests. Finally, config warnings are well covered in the config unit tests. There is more to be done here, but this accounts for the low-hanging fruit.
2019-12-18 03:14:45 +02:00
P00 INFO: last backup label = [BACKUP-DIFF-1], version = 0.00
P00 WARN: incr backup cannot alter 'checksum-page' option to 'false', reset to 'true' from [BACKUP-DIFF-1]
The backup command is implemented entirely in C. For the most part this is a direct migration of the Perl code into C except as noted below. A backup can now be initiated from a linked directory. The link will not be stored in the manifest or recreated on restore. If a link or directory does not already exist in the restore location then a directory will be created. The logic for creating backup labels has been improved and it should no longer be possible to get a backup label earlier than the latest backup even with timezone changes or clock skew. This has never been an issue in the field that we know of, but we found it in testing. For online backups all times are fetched from the PostgreSQL primary host (before only copy start was). This doesn't affect backup integrity but it does prevent clock skew between hosts affecting backup duration reporting. Archive copy now works as expected when the archive and backup have different compression settings, i.e. when one is compressed and the other is not. This was a long-standing bug in the Perl code. Resume will now work even if hardlink settings have been changed. Reviewed by Cynthia Shang.
2019-12-14 00:14:26 +02:00
P01 INFO: backup file [TEST_PATH]/db-master/db/base-2/base/base2.txt (5B, 38%) checksum 09b5e31766be1dba1ec27de82f975c1b6eea2a92
P01 INFO: backup file [TEST_PATH]/db-master/db/base-2/pg_tblspc/2/[TS_PATH-1]/32768/tablespace2b.txt (8B, 100%) checksum e324463005236d83e6e54795dbddd20a74533bf3
Added the --checksum-page option. This option allows pgBackRest to validate page checksums in data files when checksums are enabled on PostgreSQL >= 9.3. Note that this functionality requires a C library which may not initially be available in OS packages. The option will automatically be enabled when the library is present and checksums are enabled on the cluster.
2016-12-13 01:54:07 +02:00
P00 WARN: page misalignment in file [TEST_PATH]/db-master/db/base-2/pg_tblspc/2/[TS_PATH-1]/32768/tablespace2b.txt: file size 8 is not divisible by page size 8192
The backup command is implemented entirely in C. For the most part this is a direct migration of the Perl code into C except as noted below. A backup can now be initiated from a linked directory. The link will not be stored in the manifest or recreated on restore. If a link or directory does not already exist in the restore location then a directory will be created. The logic for creating backup labels has been improved and it should no longer be possible to get a backup label earlier than the latest backup even with timezone changes or clock skew. This has never been an issue in the field that we know of, but we found it in testing. For online backups all times are fetched from the PostgreSQL primary host (before only copy start was). This doesn't affect backup integrity but it does prevent clock skew between hosts affecting backup duration reporting. Archive copy now works as expected when the archive and backup have different compression settings, i.e. when one is compressed and the other is not. This was a long-standing bug in the Perl code. Resume will now work even if hardlink settings have been changed. Reviewed by Cynthia Shang.
2019-12-14 00:14:26 +02:00
P00 DETAIL: reference pg_data/PG_VERSION to [BACKUP-FULL-2]
Remove integration tests that are now covered in the unit tests. Most of these tests are just checking that errors are thrown when required. These are well covered in various unit tests. The "cannot resume" tests are also well covered in the backup unit tests. Finally, config warnings are well covered in the config unit tests. There is more to be done here, but this accounts for the low-hanging fruit.
2019-12-18 03:14:45 +02:00
P00 DETAIL: reference pg_data/badchecksum.txt to [BACKUP-DIFF-1]
The backup command is implemented entirely in C. For the most part this is a direct migration of the Perl code into C except as noted below. A backup can now be initiated from a linked directory. The link will not be stored in the manifest or recreated on restore. If a link or directory does not already exist in the restore location then a directory will be created. The logic for creating backup labels has been improved and it should no longer be possible to get a backup label earlier than the latest backup even with timezone changes or clock skew. This has never been an issue in the field that we know of, but we found it in testing. For online backups all times are fetched from the PostgreSQL primary host (before only copy start was). This doesn't affect backup integrity but it does prevent clock skew between hosts affecting backup duration reporting. Archive copy now works as expected when the archive and backup have different compression settings, i.e. when one is compressed and the other is not. This was a long-standing bug in the Perl code. Resume will now work even if hardlink settings have been changed. Reviewed by Cynthia Shang.
2019-12-14 00:14:26 +02:00
P00 DETAIL: reference pg_data/base/1/12000 to [BACKUP-FULL-2]
P00 DETAIL: reference pg_data/base/1/PG_VERSION to [BACKUP-FULL-2]
P00 DETAIL: reference pg_data/base/16384/17000 to [BACKUP-FULL-2]
P00 DETAIL: reference pg_data/base/16384/PG_VERSION to [BACKUP-FULL-2]
P00 DETAIL: reference pg_data/base/32768/33000 to [BACKUP-FULL-2]
P00 DETAIL: reference pg_data/base/32768/33000.32767 to [BACKUP-FULL-2]
P00 DETAIL: reference pg_data/base/32768/33001 to [BACKUP-FULL-2]
P00 DETAIL: reference pg_data/base/32768/44000_init to [BACKUP-FULL-2]
P00 DETAIL: reference pg_data/base/32768/PG_VERSION to [BACKUP-FULL-2]
Remove integration tests that are now covered in the unit tests. Most of these tests are just checking that errors are thrown when required. These are well covered in various unit tests. The "cannot resume" tests are also well covered in the backup unit tests. Finally, config warnings are well covered in the config unit tests. There is more to be done here, but this accounts for the low-hanging fruit.
2019-12-18 03:14:45 +02:00
P00 DETAIL: reference pg_data/changecontent.txt to [BACKUP-DIFF-1]
The backup command is implemented entirely in C. For the most part this is a direct migration of the Perl code into C except as noted below. A backup can now be initiated from a linked directory. The link will not be stored in the manifest or recreated on restore. If a link or directory does not already exist in the restore location then a directory will be created. The logic for creating backup labels has been improved and it should no longer be possible to get a backup label earlier than the latest backup even with timezone changes or clock skew. This has never been an issue in the field that we know of, but we found it in testing. For online backups all times are fetched from the PostgreSQL primary host (before only copy start was). This doesn't affect backup integrity but it does prevent clock skew between hosts affecting backup duration reporting. Archive copy now works as expected when the archive and backup have different compression settings, i.e. when one is compressed and the other is not. This was a long-standing bug in the Perl code. Resume will now work even if hardlink settings have been changed. Reviewed by Cynthia Shang.
2019-12-14 00:14:26 +02:00
P00 DETAIL: reference pg_data/changetime.txt to [BACKUP-FULL-2]
P00 DETAIL: reference pg_data/global/pg_control to [BACKUP-FULL-2]
P00 DETAIL: reference pg_data/pg_stat/global.stat to [BACKUP-FULL-2]
P00 DETAIL: reference pg_data/postgresql.conf to [BACKUP-FULL-2]
P00 DETAIL: reference pg_data/special-!_.*'()&!@;:+,? to [BACKUP-FULL-2]
P00 DETAIL: reference pg_data/zero_from_start to [BACKUP-FULL-2]
Remove integration tests that are now covered in the unit tests. Most of these tests are just checking that errors are thrown when required. These are well covered in various unit tests. The "cannot resume" tests are also well covered in the backup unit tests. Finally, config warnings are well covered in the config unit tests. There is more to be done here, but this accounts for the low-hanging fruit.
2019-12-18 03:14:45 +02:00
P00 DETAIL: reference pg_data/zerosize.txt to [BACKUP-DIFF-1]
P00 DETAIL: reference pg_tblspc/2/[TS_PATH-1]/32768/tablespace2.txt to [BACKUP-DIFF-1]
Show Process ID in log output instead of filtering it out with the timestamp.
2016-09-14 23:30:57 +02:00
P00 INFO: incr backup size = 13B
P00 INFO: new backup label = [BACKUP-INCR-3]
Added a predictable log entry to signal that a command has completed successfully. For example a backup ends successfully with: INFO: backup command end: completed successfully.
2016-12-04 21:44:53 +02:00
P00 INFO: backup command end: completed successfully
Integration test improvements for disk and memory efficiency. Set log-level-file=off when more that one test will run. In this case is it impossible to see the logs anyway since they will be automatically cleaned up after the test. This improves performance pretty dramatically since trace-level logging is expensive. If a singe integration test is run then log-level-file is trace by default but can be changed with the --log-level-test-file option. Reduce buffer-size to 64k to save memory during testing and allow more processes to run in parallel. Update log replacement rules so that these options can change without affecting expect logs.
2019-12-17 22:23:07 +02:00
P00 INFO: expire command begin [BACKREST-VERSION]: --buffer-size=[BUFFER-SIZE] --config=[TEST_PATH]/db-master/pgbackrest.conf --lock-path=[TEST_PATH]/db-master/lock --log-level-console=detail --log-level-file=[LOG-LEVEL-FILE] --log-level-stderr=off --log-path=[TEST_PATH]/db-master/log[] --no-log-timestamp --repo1-path=[TEST_PATH]/db-master/repo --stanza=db
Rename retention-* options to repo-retention-*.
2018-02-19 22:07:24 +02:00
P00 INFO: option 'repo1-retention-archive' is not set - archive logs will not be expired
Added a predictable log entry to signal that a command has completed successfully. For example a backup ends successfully with: INFO: backup command end: completed successfully.
2016-12-04 21:44:53 +02:00
P00 INFO: expire command end: completed successfully
Log testing can now be enabled for certain deterministic tests. This works by comparing the generated logs against a previous copy. Currently only enabled for the backup/synthetic tests.
2015-05-05 19:08:48 +02:00
Rename test paths for clarity. This was worked out as part of the test suite refactor [c8f806a] but not committed with it because of the large number of expect logs changes involved. Keeping them separate made it easier to audit the changes in the refactor.
2016-06-24 16:54:31 +02:00
+ supplemental file: [TEST_PATH]/db-master/pgbackrest.conf
----------------------------------------------------------
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
[db]
Rename db-* options to pg-* and backup-* options to repo-* to improve consistency. * repo-* options are now indexed although only one is allowed. * List deprecated option names in documentation and command-line help.
2018-02-04 01:27:38 +02:00
pg1-path=[TEST_PATH]/db-master/db/base-2
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
New simpler configuration and consistent project/exe/path naming. * The repo-path option now always refers to the repository where backups and archive are stored, whether local or remote, so the repo-remote-path option has been removed. The new spool-path option can be used to define a location for queueing WAL segments when archiving asynchronously. Otherwise, a local repository is no longer required. * Implemented a new config format which should be far simpler to use. See the User Guide and Configuration Reference for details but for a simple configuration all options can now be placed in the stanza section. Options that are shared between stanzas can be placed in the [global] section. More complex configurations can still make use of command sections though this should be a rare use case. * The default configuration filename is now pgbackrest.conf instead of pg_backrest.conf. This was done for consistency with other naming changes but also to prevent old config files from being loaded accidentally. * The default repository name was changed from /var/lib/backup to /var/lib/pgbackrest. * Lock files are now stored in /tmp/pgbackrest by default. These days /run/pgbackrest would be the preferred location but that would require init scripts which are not part of this release. The lock-path option can be used to configure the lock directory. * Log files are now stored in /var/log/pgbackrest by default and no longer have the date appended so they can be managed with logrotate. The log-path option can be used to configure the lock directory. * Executable filename changed from pg_backrest to pgbackrest.
2016-04-14 15:30:54 +02:00
[db:restore]
Rename test paths for clarity. This was worked out as part of the test suite refactor [c8f806a] but not committed with it because of the large number of expect logs changes involved. Keeping them separate made it easier to audit the changes in the refactor.
2016-06-24 16:54:31 +02:00
tablespace-map=1=[TEST_PATH]/db-master/db/tablespace/ts1-2
tablespace-map=2=[TEST_PATH]/db-master/db/tablespace/ts2-2
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
New simpler configuration and consistent project/exe/path naming. * The repo-path option now always refers to the repository where backups and archive are stored, whether local or remote, so the repo-remote-path option has been removed. The new spool-path option can be used to define a location for queueing WAL segments when archiving asynchronously. Otherwise, a local repository is no longer required. * Implemented a new config format which should be far simpler to use. See the User Guide and Configuration Reference for details but for a simple configuration all options can now be placed in the stanza section. Options that are shared between stanzas can be placed in the [global] section. More complex configurations can still make use of command sections though this should be a rare use case. * The default configuration filename is now pgbackrest.conf instead of pg_backrest.conf. This was done for consistency with other naming changes but also to prevent old config files from being loaded accidentally. * The default repository name was changed from /var/lib/backup to /var/lib/pgbackrest. * Lock files are now stored in /tmp/pgbackrest by default. These days /run/pgbackrest would be the preferred location but that would require init scripts which are not part of this release. The lock-path option can be used to configure the lock directory. * Log files are now stored in /var/log/pgbackrest by default and no longer have the date appended so they can be managed with logrotate. The log-path option can be used to configure the lock directory. * Executable filename changed from pg_backrest to pgbackrest.
2016-04-14 15:30:54 +02:00
[global]
Integration test improvements for disk and memory efficiency. Set log-level-file=off when more that one test will run. In this case is it impossible to see the logs anyway since they will be automatically cleaned up after the test. This improves performance pretty dramatically since trace-level logging is expensive. If a singe integration test is run then log-level-file is trace by default but can be changed with the --log-level-test-file option. Reduce buffer-size to 64k to save memory during testing and allow more processes to run in parallel. Update log replacement rules so that these options can change without affecting expect logs.
2019-12-17 22:23:07 +02:00
buffer-size=[BUFFER-SIZE]
Fix custom settings for compress-level option being ignored. Reported by Jens Wilke.
2017-10-23 00:05:46 +02:00
compress-level=3
Add infrastructure for multiple compression type support. Add compress-type option and deprecate compress option. Since the compress option is boolean it won't work with multiple compression types. Add logic to cfgLoadUpdateOption() to update compress-type if it is not set directly. The compress option should no longer be referenced outside the cfgLoadUpdateOption() function. Add common/compress/helper module to contain interface functions that work with multiple compression types. Code outside this module should no longer call specific compression drivers, though it may be OK to reference a specific compression type using the new interface (e.g., saving backup history files in gz format). Unit tests only test compression using the gz format because other formats may not be available in all builds. It is the job of integration tests to exercise all compression types. Additional compression types will be added in future commits.
2020-03-06 21:41:03 +02:00
compress-type=none
Storage and IO layer refactor: Refactor storage layer to allow for new repository filesystems using drivers. (Reviewed by Cynthia Shang.) Refactor IO layer to allow for new compression formats, checksum types, and other capabilities using filters. (Reviewed by Cynthia Shang.)
2017-06-09 23:51:41 +02:00
db-timeout=45
lock-path=[TEST_PATH]/db-master/lock
Reduce log level for all expect tests to detail. The C code is designed to be efficient rather than deterministic at the debug log level. As we move more testing from integration to unit tests it makes less sense to try and maintain the expect logs at this log level. Most of the expect logs have already been moved to detail level but mock/all still had tests at debug level. Change the logging defaults in the config file and remove as many references to log-level-console as possible.
2019-05-23 00:23:44 +02:00
log-level-console=detail
Integration test improvements for disk and memory efficiency. Set log-level-file=off when more that one test will run. In this case is it impossible to see the logs anyway since they will be automatically cleaned up after the test. This improves performance pretty dramatically since trace-level logging is expensive. If a singe integration test is run then log-level-file is trace by default but can be changed with the --log-level-test-file option. Reduce buffer-size to 64k to save memory during testing and allow more processes to run in parallel. Update log replacement rules so that these options can change without affecting expect logs.
2019-12-17 22:23:07 +02:00
log-level-file=[LOG-LEVEL-FILE]
Added the log-level-stderr option. Controls whether console log messages are sent to stderr or stdout. By default this is set to warn which represents a change in behavior from previous versions, even though it may be more intuitive. Setting log-level-stderr=off will preserve the old behavior. Suggested by Sascha Biberhofer.
2016-10-05 15:09:30 +02:00
log-level-stderr=off
Storage and IO layer refactor: Refactor storage layer to allow for new repository filesystems using drivers. (Reviewed by Cynthia Shang.) Refactor IO layer to allow for new compression formats, checksum types, and other capabilities using filters. (Reviewed by Cynthia Shang.)
2017-06-09 23:51:41 +02:00
log-path=[TEST_PATH]/db-master/log
Integration test improvements for disk and memory efficiency. Set log-level-file=off when more that one test will run. In this case is it impossible to see the logs anyway since they will be automatically cleaned up after the test. This improves performance pretty dramatically since trace-level logging is expensive. If a singe integration test is run then log-level-file is trace by default but can be changed with the --log-level-test-file option. Reduce buffer-size to 64k to save memory during testing and allow more processes to run in parallel. Update log replacement rules so that these options can change without affecting expect logs.
2019-12-17 22:23:07 +02:00
log-subprocess=[LOG-SUBPROCESS]
Set log-timestamp=n for integration tests. This means less filtering of logs needs to be done and new timestamps can be added without adding new filters.
2018-05-11 17:24:38 +02:00
log-timestamp=n
Storage and IO layer refactor: Refactor storage layer to allow for new repository filesystems using drivers. (Reviewed by Cynthia Shang.) Refactor IO layer to allow for new compression formats, checksum types, and other capabilities using filters. (Reviewed by Cynthia Shang.)
2017-06-09 23:51:41 +02:00
protocol-timeout=60
Rename db-* options to pg-* and backup-* options to repo-* to improve consistency. * repo-* options are now indexed although only one is allowed. * List deprecated option names in documentation and command-line help.
2018-02-04 01:27:38 +02:00
repo1-path=[TEST_PATH]/db-master/repo
Consolidate archive-push tests into new archive test.
2017-06-21 14:38:53 +02:00
spool-path=[TEST_PATH]/db-master/spool
New simpler configuration and consistent project/exe/path naming. * The repo-path option now always refers to the repository where backups and archive are stored, whether local or remote, so the repo-remote-path option has been removed. The new spool-path option can be used to define a location for queueing WAL segments when archiving asynchronously. Otherwise, a local repository is no longer required. * Implemented a new config format which should be far simpler to use. See the User Guide and Configuration Reference for details but for a simple configuration all options can now be placed in the stanza section. Options that are shared between stanzas can be placed in the [global] section. More complex configurations can still make use of command sections though this should be a rare use case. * The default configuration filename is now pgbackrest.conf instead of pg_backrest.conf. This was done for consistency with other naming changes but also to prevent old config files from being loaded accidentally. * The default repository name was changed from /var/lib/backup to /var/lib/pgbackrest. * Lock files are now stored in /tmp/pgbackrest by default. These days /run/pgbackrest would be the preferred location but that would require init scripts which are not part of this release. The lock-path option can be used to configure the lock directory. * Log files are now stored in /var/log/pgbackrest by default and no longer have the date appended so they can be managed with logrotate. The log-path option can be used to configure the lock directory. * Executable filename changed from pg_backrest to pgbackrest.
2016-04-14 15:30:54 +02:00
[global:backup]
archive-copy=y
start-fast=y
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
Rename test paths for clarity. This was worked out as part of the test suite refactor [c8f806a] but not committed with it because of the large number of expect logs changes involved. Keeping them separate made it easier to audit the changes in the refactor.
2016-06-24 16:54:31 +02:00
+ supplemental file: [TEST_PATH]/db-master/repo/backup/db/[BACKUP-INCR-3]/backup.manifest
-----------------------------------------------------------------------------------------
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
[backrest]
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
backrest-format=5
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
backrest-version="[VERSION-1]"
[backup]
backup-label="[BACKUP-INCR-3]"
Remove integration tests that are now covered in the unit tests. Most of these tests are just checking that errors are thrown when required. These are well covered in various unit tests. The "cannot resume" tests are also well covered in the backup unit tests. Finally, config warnings are well covered in the config unit tests. There is more to be done here, but this accounts for the low-hanging fruit.
2019-12-18 03:14:45 +02:00
backup-prior="[BACKUP-DIFF-1]"
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
backup-timestamp-copy-start=[TIMESTAMP]
backup-timestamp-start=[TIMESTAMP]
backup-timestamp-stop=[TIMESTAMP]
backup-type="incr"
[backup:db]
Database version refactoring. * Refactor db version constants into a separate module. * Update synthetic backup tests to PostgreSQL 9.4.
2016-08-12 04:35:24 +02:00
db-catalog-version=201409291
db-control-version=942
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
db-id=1
Automate generation of WAL and pg_control test files. The existing static files would not work with 32-bit or big-endian systems so create functions to generate these files dynamically rather than creating a bunch of new static files.
2017-11-19 03:02:54 +02:00
db-system-id=1000000000000000094
Database version refactoring. * Refactor db version constants into a separate module. * Update synthetic backup tests to PostgreSQL 9.4.
2016-08-12 04:35:24 +02:00
db-version="9.4"
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
[backup:option]
Set archive-check option in manifest correctly when offline. Archive check does not run when in offline backup mode but the option was set to true in the manifest. It's harmless since these options are informational only but it could cause confusion when debugging.
2019-11-28 15:27:21 +02:00
option-archive-check=false
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
option-archive-copy=true
Backup from a standby cluster. A connection to the primary cluster is still required to start/stop the backup and copy files that are not replicated, but the vast majority of files are copied from the standby in order to reduce load on the master.
2016-08-25 17:25:46 +02:00
option-backup-standby=false
Integration test improvements for disk and memory efficiency. Set log-level-file=off when more that one test will run. In this case is it impossible to see the logs anyway since they will be automatically cleaned up after the test. This improves performance pretty dramatically since trace-level logging is expensive. If a singe integration test is run then log-level-file is trace by default but can be changed with the --log-level-test-file option. Reduce buffer-size to 64k to save memory during testing and allow more processes to run in parallel. Update log replacement rules so that these options can change without affecting expect logs.
2019-12-17 22:23:07 +02:00
option-buffer-size=[BUFFER-SIZE]
Added the --checksum-page option. This option allows pgBackRest to validate page checksums in data files when checksums are enabled on PostgreSQL >= 9.3. Note that this functionality requires a C library which may not initially be available in OS packages. The option will automatically be enabled when the library is present and checksums are enabled on the cluster.
2016-12-13 01:54:07 +02:00
option-checksum-page=true
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
option-compress=false
Add additional options to backup.manifest for debugging purposes. Add the buffer-size, compress-level, compress-level-network, and process-max options to the backup:option section in backup.manifest to aid in debugging. It may also make sense to propagate these options up to backup.info so they can be displayed in the info command, but for now this is deemed sufficient. Contributed by blogh.
2019-03-10 11:03:52 +02:00
option-compress-level=3
option-compress-level-network=3
Add infrastructure for multiple compression type support. Add compress-type option and deprecate compress option. Since the compress option is boolean it won't work with multiple compression types. Add logic to cfgLoadUpdateOption() to update compress-type if it is not set directly. The compress option should no longer be referenced outside the cfgLoadUpdateOption() function. Add common/compress/helper module to contain interface functions that work with multiple compression types. Code outside this module should no longer call specific compression drivers, though it may be OK to reference a specific compression type using the new interface (e.g., saving backup history files in gz format). Unit tests only test compression using the gz format because other formats may not be available in all builds. It is the job of integration tests to exercise all compression types. Additional compression types will be added in future commits.
2020-03-06 21:41:03 +02:00
option-compress-type="none"
Add checksum delta for incremental backups. Use checksums rather than timestamps to determine if files have changed. This is useful in cases where the timestamps may not be trustworthy, e.g. when performing an incremental after failing over to a standby. If checksum delta is enabled then checksums will be used for verification of resumed backups, even if they are full. Resumes have always used checksums to verify the files in the repository, enabling delta performs checksums on the database files as well. Note that the user must manually enable this feature in cases were it would be useful or just keep in enabled all the time. A future commit will address automatically enabling the feature in cases where it seems likely to be useful. Contributed by Cynthia Shang.
2018-09-19 17:12:45 +02:00
option-delta=false
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
option-hardlink=false
Closed #167: Rename `--no-start-stop` option to `--no-online`.
2016-02-20 02:32:02 +02:00
option-online=false
Add additional options to backup.manifest for debugging purposes. Add the buffer-size, compress-level, compress-level-network, and process-max options to the backup:option section in backup.manifest to aid in debugging. It may also make sense to propagate these options up to backup.info so they can be displayed in the info command, but for now this is deemed sufficient. Contributed by blogh.
2019-03-10 11:03:52 +02:00
option-process-max=1
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
[backup:target]
Rename test paths for clarity. This was worked out as part of the test suite refactor [c8f806a] but not committed with it because of the large number of expect logs changes involved. Keeping them separate made it easier to audit the changes in the refactor.
2016-06-24 16:54:31 +02:00
pg_data={"path":"[TEST_PATH]/db-master/db/base-2","type":"path"}
pg_tblspc/2={"path":"[TEST_PATH]/db-master/db/tablespace/ts2-2","tablespace-id":"2","tablespace-name":"ts2","type":"link"}
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
[target:file]
Fixed a issue where tablespaces were copied from the master during standby backup.
2016-09-04 15:19:44 +02:00
pg_data/PG_VERSION={"checksum":"184473f470864e067ee3a22e64b47b0a1c356f29","master":true,"reference":"[BACKUP-FULL-2]","size":[SIZE],"timestamp":[TIMESTAMP-1]}
Remove integration tests that are now covered in the unit tests. Most of these tests are just checking that errors are thrown when required. These are well covered in various unit tests. The "cannot resume" tests are also well covered in the backup unit tests. Finally, config warnings are well covered in the config unit tests. There is more to be done here, but this accounts for the low-hanging fruit.
2019-12-18 03:14:45 +02:00
pg_data/badchecksum.txt={"checksum":"f927212cd08d11a42a666b2f04235398e9ceeb51","master":true,"reference":"[BACKUP-DIFF-1]","size":[SIZE],"timestamp":[TIMESTAMP-1]}
Added the --checksum-page option. This option allows pgBackRest to validate page checksums in data files when checksums are enabled on PostgreSQL >= 9.3. Note that this functionality requires a C library which may not initially be available in OS packages. The option will automatically be enabled when the library is present and checksums are enabled on the cluster.
2016-12-13 01:54:07 +02:00
pg_data/base/1/12000={"checksum":"22c98d248ff548311eda88559e4a8405ed77c003","checksum-page":true,"reference":"[BACKUP-FULL-2]","size":[SIZE],"timestamp":[TIMESTAMP-1]}
Database version refactoring. * Refactor db version constants into a separate module. * Update synthetic backup tests to PostgreSQL 9.4.
2016-08-12 04:35:24 +02:00
pg_data/base/1/PG_VERSION={"checksum":"184473f470864e067ee3a22e64b47b0a1c356f29","mode":"0660","reference":"[BACKUP-FULL-2]","size":[SIZE],"timestamp":[TIMESTAMP-1]}
Added the --checksum-page option. This option allows pgBackRest to validate page checksums in data files when checksums are enabled on PostgreSQL >= 9.3. Note that this functionality requires a C library which may not initially be available in OS packages. The option will automatically be enabled when the library is present and checksums are enabled on the cluster.
2016-12-13 01:54:07 +02:00
pg_data/base/16384/17000={"checksum":"e0101dd8ffb910c9c202ca35b5f828bcb9697bed","checksum-page":false,"checksum-page-error":[1],"reference":"[BACKUP-FULL-2]","size":[SIZE],"timestamp":[TIMESTAMP-1]}
Database version refactoring. * Refactor db version constants into a separate module. * Update synthetic backup tests to PostgreSQL 9.4.
2016-08-12 04:35:24 +02:00
pg_data/base/16384/PG_VERSION={"checksum":"184473f470864e067ee3a22e64b47b0a1c356f29","reference":"[BACKUP-FULL-2]","size":[SIZE],"timestamp":[TIMESTAMP-1]}
Fix incorrect offline upper bound for ignoring page checksum errors. For offline backups the upper bound was being set to 0x0000FFFF0000FFFF rather than UINT64_MAX. This meant that page checksum errors might be ignored for databases with a lot of past WAL in offline mode. Online mode is not affected since the upper bound is retrieved from pg_start_backup().
2019-07-11 15:13:56 +02:00
pg_data/base/32768/33000={"checksum":"7a16d165e4775f7c92e8cdf60c0af57313f0bf90","checksum-page":true,"reference":"[BACKUP-FULL-2]","size":[SIZE],"timestamp":[TIMESTAMP-1]}
pg_data/base/32768/33000.32767={"checksum":"6e99b589e550e68e934fd235ccba59fe5b592a9e","checksum-page":true,"reference":"[BACKUP-FULL-2]","size":[SIZE],"timestamp":[TIMESTAMP-1]}
Added the --checksum-page option. This option allows pgBackRest to validate page checksums in data files when checksums are enabled on PostgreSQL >= 9.3. Note that this functionality requires a C library which may not initially be available in OS packages. The option will automatically be enabled when the library is present and checksums are enabled on the cluster.
2016-12-13 01:54:07 +02:00
pg_data/base/32768/33001={"checksum":"6bf316f11d28c28914ea9be92c00de9bea6d9a6b","checksum-page":false,"checksum-page-error":[0,[3,5],7],"reference":"[BACKUP-FULL-2]","size":[SIZE],"timestamp":[TIMESTAMP-1]}
Fix incorrect offline upper bound for ignoring page checksum errors. For offline backups the upper bound was being set to 0x0000FFFF0000FFFF rather than UINT64_MAX. This meant that page checksum errors might be ignored for databases with a lot of past WAL in offline mode. Online mode is not affected since the upper bound is retrieved from pg_start_backup().
2019-07-11 15:13:56 +02:00
pg_data/base/32768/44000_init={"checksum":"7a16d165e4775f7c92e8cdf60c0af57313f0bf90","checksum-page":true,"reference":"[BACKUP-FULL-2]","size":[SIZE],"timestamp":[TIMESTAMP-1]}
Database version refactoring. * Refactor db version constants into a separate module. * Update synthetic backup tests to PostgreSQL 9.4.
2016-08-12 04:35:24 +02:00
pg_data/base/32768/PG_VERSION={"checksum":"184473f470864e067ee3a22e64b47b0a1c356f29","reference":"[BACKUP-FULL-2]","size":[SIZE],"timestamp":[TIMESTAMP-1]}
Update Perl page checksum expression. This expression determines which files contain page checksums but it was also including the directory above the relation directories. In a real PostgreSQL installation this not a problem because these directories don't contain any files. However, our tests place a file in `base` which the Perl code thought should have page checksums while the new C code says no. Update the expression to document the change and avoid churn in the expect logs later.
2019-11-25 14:37:09 +02:00
pg_data/base/base2.txt={"checksum":"09b5e31766be1dba1ec27de82f975c1b6eea2a92","size":[SIZE],"timestamp":[TIMESTAMP-1]}
Remove integration tests that are now covered in the unit tests. Most of these tests are just checking that errors are thrown when required. These are well covered in various unit tests. The "cannot resume" tests are also well covered in the backup unit tests. Finally, config warnings are well covered in the config unit tests. There is more to be done here, but this accounts for the low-hanging fruit.
2019-12-18 03:14:45 +02:00
pg_data/changecontent.txt={"checksum":"a094d94583e209556d03c3c5da33131a065f1689","master":true,"reference":"[BACKUP-DIFF-1]","size":[SIZE],"timestamp":[TIMESTAMP-1]}
Add checksum delta for incremental backups. Use checksums rather than timestamps to determine if files have changed. This is useful in cases where the timestamps may not be trustworthy, e.g. when performing an incremental after failing over to a standby. If checksum delta is enabled then checksums will be used for verification of resumed backups, even if they are full. Resumes have always used checksums to verify the files in the repository, enabling delta performs checksums on the database files as well. Note that the user must manually enable this feature in cases were it would be useful or just keep in enabled all the time. A future commit will address automatically enabling the feature in cases where it seems likely to be useful. Contributed by Cynthia Shang.
2018-09-19 17:12:45 +02:00
pg_data/changetime.txt={"checksum":"88087292ed82e26f3eb824d0bffc05ccf7a30f8d","master":true,"reference":"[BACKUP-FULL-2]","size":[SIZE],"timestamp":[TIMESTAMP-2]}
Support configurable WAL segment size. PostgreSQL 11 introduces configurable WAL segment sizes, from 1MB to 1GB. There are two areas that needed to be updated to support this: building the archive-get queue and checking that WAL has been archived after a backup. Both operations require the WAL segment size to properly build a list. Checking the archive after a backup is still implemented in Perl and has an active database connection, so just get the WAL segment size from the database. The archive-get command does not have a connection to the database, so get the WAL segment size from pg_control instead. This requires a deeper inspection of pg_control than has been done in the past, so it seemed best to copy the relevant data structures from each version of PostgreSQL and build a generic interface layer to address them. While this approach is a bit verbose, it has the advantage of being relatively simple, and can easily be updated for new versions of PostgreSQL. Since the integration tests generate pg_control files for testing, teach Perl how to generate files with the correct offsets for both 32-bit and 64-bit architectures.
2018-09-25 11:24:42 +02:00
pg_data/global/pg_control={"checksum":"4c77c900f7af0d9ab13fa9982051a42e0b637f6c","master":true,"reference":"[BACKUP-FULL-2]","size":[SIZE],"timestamp":[TIMESTAMP-2]}
Fixed a issue where tablespaces were copied from the master during standby backup.
2016-09-04 15:19:44 +02:00
pg_data/pg_stat/global.stat={"checksum":"e350d5ce0153f3e22d5db21cf2a4eff00f3ee877","master":true,"reference":"[BACKUP-FULL-2]","size":[SIZE],"timestamp":[TIMESTAMP-2]}
pg_data/postgresql.conf={"checksum":"6721d92c9fcdf4248acff1f9a1377127d9064807","master":true,"reference":"[BACKUP-FULL-2]","size":[SIZE],"timestamp":[TIMESTAMP-2]}
Use minio for integration testing. ScalityS3 has not received any maintenance in years and is slow to start which is bad for testing. Replace it with minio which starts quickly and ships as a single executable or a tiny container. Minio has stricter limits on allowable characters but should still provide enough coverage to show that our encoding is working correctly. This commit also includes the upgrade to openssl 1.1.1 in the Ubuntu 18.04 container.
2019-07-03 04:20:35 +02:00
pg_data/special-!_.*'()&!@;:+,?={"master":true,"reference":"[BACKUP-FULL-2]","size":[SIZE],"timestamp":[TIMESTAMP-1]}
Allow zero-size files in backup manifest to reference a prior manifest regardless of timestamp delta. Contributed by Cynthia Shang.
2018-08-24 22:50:33 +02:00
pg_data/zero_from_start={"master":true,"reference":"[BACKUP-FULL-2]","size":[SIZE],"timestamp":[TIMESTAMP-1]}
Remove integration tests that are now covered in the unit tests. Most of these tests are just checking that errors are thrown when required. These are well covered in various unit tests. The "cannot resume" tests are also well covered in the backup unit tests. Finally, config warnings are well covered in the config unit tests. There is more to be done here, but this accounts for the low-hanging fruit.
2019-12-18 03:14:45 +02:00
pg_data/zerosize.txt={"master":true,"reference":"[BACKUP-DIFF-1]","size":[SIZE],"timestamp":[TIMESTAMP-2]}
pg_tblspc/2/[TS_PATH-1]/32768/tablespace2.txt={"checksum":"dc7f76e43c46101b47acc55ae4d593a9e6983578","checksum-page":false,"reference":"[BACKUP-DIFF-1]","size":[SIZE],"timestamp":[TIMESTAMP-1]}
Added the --checksum-page option. This option allows pgBackRest to validate page checksums in data files when checksums are enabled on PostgreSQL >= 9.3. Note that this functionality requires a C library which may not initially be available in OS packages. The option will automatically be enabled when the library is present and checksums are enabled on the cluster.
2016-12-13 01:54:07 +02:00
pg_tblspc/2/[TS_PATH-1]/32768/tablespace2b.txt={"checksum":"e324463005236d83e6e54795dbddd20a74533bf3","checksum-page":false,"size":[SIZE],"timestamp":[TIMESTAMP-1]}
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
[target:file:default]
Continuous integration using travis-ci.
2017-02-21 15:59:23 +02:00
group="[GROUP-1]"
Fixed a issue where tablespaces were copied from the master during standby backup.
2016-09-04 15:19:44 +02:00
master=false
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
mode="0600"
Regression tests can now be run as any properly-configured user, not just vagrant.
2016-10-01 19:39:44 +02:00
user="[USER-1]"
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
[target:link]
Rename test paths for clarity. This was worked out as part of the test suite refactor [c8f806a] but not committed with it because of the large number of expect logs changes involved. Keeping them separate made it easier to audit the changes in the refactor.
2016-06-24 16:54:31 +02:00
pg_data/pg_tblspc/2={"destination":"[TEST_PATH]/db-master/db/tablespace/ts2-2"}
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
[target:link:default]
Continuous integration using travis-ci.
2017-02-21 15:59:23 +02:00
group="[GROUP-1]"
Regression tests can now be run as any properly-configured user, not just vagrant.
2016-10-01 19:39:44 +02:00
user="[USER-1]"
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
[target:path]
The restore command is implemented entirely in C. For the most part this is a direct migration of the Perl code into C. There is one important behavioral change with regard to how file permissions are handled. The Perl code tried to set ownership as it was in the manifest even when running as an unprivileged user. This usually just led to errors and frustration. The C code works like this: If a restore is run as a non-root user (the typical scenario) then all files restored will belong to the user/group executing pgBackRest. If existing files are not owned by the executing user/group then an error will result if the ownership cannot be updated to the executing user/group. In that case the file ownership will need to be updated by a privileged user before the restore can be retried. If a restore is run as the root user then pgBackRest will attempt to recreate the ownership recorded in the manifest when the backup was made. Only user/group names are stored in the manifest so the same names must exist on the restore host for this to work. If the user/group name cannot be found locally then the user/group of the PostgreSQL data directory will be used and finally root if the data directory user/group cannot be mapped to a name. Reviewed by Cynthia Shang.
2019-09-26 13:52:02 +02:00
pg_data={"mode":"0777"}
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
pg_data/base={}
Allow selective restore of databases from a cluster backup. This feature can result in major space and time savings when only specific databases are restored. Unrestored databases will not be accessible but must be manually dropped before they will be removed from the shared catalogue.
2016-05-11 15:21:39 +02:00
pg_data/base/1={}
pg_data/base/16384={}
pg_data/base/32768={}
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
pg_data/global={}
pg_data/pg_clog={}
Exclude directories during backup that are cleaned, recreated, or zeroed by PostgreSQL at startup. These include (depending on the version where they were introduced): pgsql_tmp, pg_dynshmem, pg_notify, pg_replslot, pg_serial, pg_snapshots, pg_stat_tmp, pg_subtrans. The postgresql.auto.conf.tmp file is now excluded in addition to files that were already excluded: backup_label.old, postmaster.opts, postmaster.pid, recovery.conf, recovery.done.
2016-08-16 15:35:16 +02:00
pg_data/pg_dynshmem={}
Allow arbitrary directories and/or files to be excluded from a backup. Misuse of this feature can lead to inconsistent backups so read the --exclude documentation carefully before using.
2018-08-27 21:51:05 +02:00
pg_data/pg_log={}
Exclude directories during backup that are cleaned, recreated, or zeroed by PostgreSQL at startup. These include (depending on the version where they were introduced): pgsql_tmp, pg_dynshmem, pg_notify, pg_replslot, pg_serial, pg_snapshots, pg_stat_tmp, pg_subtrans. The postgresql.auto.conf.tmp file is now excluded in addition to files that were already excluded: backup_label.old, postmaster.opts, postmaster.pid, recovery.conf, recovery.done.
2016-08-16 15:35:16 +02:00
pg_data/pg_notify={}
pg_data/pg_replslot={}
pg_data/pg_serial={}
pg_data/pg_snapshots={}
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
pg_data/pg_stat={}
Exclude directories during backup that are cleaned, recreated, or zeroed by PostgreSQL at startup. These include (depending on the version where they were introduced): pgsql_tmp, pg_dynshmem, pg_notify, pg_replslot, pg_serial, pg_snapshots, pg_stat_tmp, pg_subtrans. The postgresql.auto.conf.tmp file is now excluded in addition to files that were already excluded: backup_label.old, postmaster.opts, postmaster.pid, recovery.conf, recovery.done.
2016-08-16 15:35:16 +02:00
pg_data/pg_stat_tmp={}
pg_data/pg_subtrans={}
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
pg_data/pg_tblspc={}
The restore command is implemented entirely in C. For the most part this is a direct migration of the Perl code into C. There is one important behavioral change with regard to how file permissions are handled. The Perl code tried to set ownership as it was in the manifest even when running as an unprivileged user. This usually just led to errors and frustration. The C code works like this: If a restore is run as a non-root user (the typical scenario) then all files restored will belong to the user/group executing pgBackRest. If existing files are not owned by the executing user/group then an error will result if the ownership cannot be updated to the executing user/group. In that case the file ownership will need to be updated by a privileged user before the restore can be retried. If a restore is run as the root user then pgBackRest will attempt to recreate the ownership recorded in the manifest when the backup was made. Only user/group names are stored in the manifest so the same names must exist on the restore host for this to work. If the user/group name cannot be found locally then the user/group of the PostgreSQL data directory will be used and finally root if the data directory user/group cannot be mapped to a name. Reviewed by Cynthia Shang.
2019-09-26 13:52:02 +02:00
pg_tblspc={"mode":"0777"}
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
pg_tblspc/2={}
pg_tblspc/2/[TS_PATH-1]={}
Allow selective restore of databases from a cluster backup. This feature can result in major space and time savings when only specific databases are restored. Unrestored databases will not be accessible but must be manually dropped before they will be removed from the shared catalogue.
2016-05-11 15:21:39 +02:00
pg_tblspc/2/[TS_PATH-1]/32768={}
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
[target:path:default]
Continuous integration using travis-ci.
2017-02-21 15:59:23 +02:00
group="[GROUP-1]"
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
mode="0700"
Regression tests can now be run as any properly-configured user, not just vagrant.
2016-10-01 19:39:44 +02:00
user="[USER-1]"
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
Move info file checksum to the end of the file. Putting the checksum at the beginning of the file made it impossible to stream the file out when saving. The entire file had to be held in memory while it was checksummed so the checksum could be written at the beginning. Instead place the checksum at the end. This does not break the existing Perl or C code since the read is not order dependent. There are no plans to improve the Perl code to take advantage of this change, but it will make the C implementation more efficient. Reviewed by Cynthia Shang.
2019-08-22 01:45:48 +02:00
[backrest]
backrest-checksum="[CHECKSUM]"
Rename test paths for clarity. This was worked out as part of the test suite refactor [c8f806a] but not committed with it because of the large number of expect logs changes involved. Keeping them separate made it easier to audit the changes in the refactor.
2016-06-24 16:54:31 +02:00
+ supplemental file: [TEST_PATH]/db-master/repo/backup/db/backup.info
---------------------------------------------------------------------
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
[backrest]
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
backrest-format=5
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
backrest-version="[VERSION-1]"
[backup:current]
Set archive-check option in manifest correctly when offline. Archive check does not run when in offline backup mode but the option was set to true in the manifest. It's harmless since these options are informational only but it could cause confusion when debugging.
2019-11-28 15:27:21 +02:00
[BACKUP-FULL-2]={"backrest-format":5,"backrest-version":"[VERSION-1]","backup-archive-start":null,"backup-archive-stop":null,"backup-info-repo-size":[SIZE],"backup-info-repo-size-delta":[DELTA],"backup-info-size":[SIZE],"backup-info-size-delta":[DELTA],"backup-timestamp-start":[TIMESTAMP],"backup-timestamp-stop":[TIMESTAMP],"backup-type":"full","db-id":1,"option-archive-check":false,"option-archive-copy":true,"option-backup-standby":false,"option-checksum-page":true,"option-compress":false,"option-hardlink":false,"option-online":false}
Remove integration tests that are now covered in the unit tests. Most of these tests are just checking that errors are thrown when required. These are well covered in various unit tests. The "cannot resume" tests are also well covered in the backup unit tests. Finally, config warnings are well covered in the config unit tests. There is more to be done here, but this accounts for the low-hanging fruit.
2019-12-18 03:14:45 +02:00
[BACKUP-INCR-2]={"backrest-format":5,"backrest-version":"[VERSION-1]","backup-archive-start":null,"backup-archive-stop":null,"backup-info-repo-size":[SIZE],"backup-info-repo-size-delta":[DELTA],"backup-info-size":[SIZE],"backup-info-size-delta":[DELTA],"backup-prior":"[BACKUP-FULL-2]","backup-reference":["[BACKUP-FULL-2]"],"backup-timestamp-start":[TIMESTAMP],"backup-timestamp-stop":[TIMESTAMP],"backup-type":"incr","db-id":1,"option-archive-check":false,"option-archive-copy":true,"option-backup-standby":false,"option-checksum-page":true,"option-compress":false,"option-hardlink":false,"option-online":false}
[BACKUP-DIFF-1]={"backrest-format":5,"backrest-version":"[VERSION-1]","backup-archive-start":null,"backup-archive-stop":null,"backup-info-repo-size":[SIZE],"backup-info-repo-size-delta":[DELTA],"backup-info-size":[SIZE],"backup-info-size-delta":[DELTA],"backup-prior":"[BACKUP-FULL-2]","backup-reference":["[BACKUP-FULL-2]"],"backup-timestamp-start":[TIMESTAMP],"backup-timestamp-stop":[TIMESTAMP],"backup-type":"diff","db-id":1,"option-archive-check":false,"option-archive-copy":true,"option-backup-standby":false,"option-checksum-page":true,"option-compress":false,"option-hardlink":false,"option-online":false}
[BACKUP-INCR-3]={"backrest-format":5,"backrest-version":"[VERSION-1]","backup-archive-start":null,"backup-archive-stop":null,"backup-info-repo-size":[SIZE],"backup-info-repo-size-delta":[DELTA],"backup-info-size":[SIZE],"backup-info-size-delta":[DELTA],"backup-prior":"[BACKUP-DIFF-1]","backup-reference":["[BACKUP-FULL-2]","[BACKUP-DIFF-1]"],"backup-timestamp-start":[TIMESTAMP],"backup-timestamp-stop":[TIMESTAMP],"backup-type":"incr","db-id":1,"option-archive-check":false,"option-archive-copy":true,"option-backup-standby":false,"option-checksum-page":true,"option-compress":false,"option-hardlink":false,"option-online":false}
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
[db]
Database version refactoring. * Refactor db version constants into a separate module. * Update synthetic backup tests to PostgreSQL 9.4.
2016-08-12 04:35:24 +02:00
db-catalog-version=201409291
db-control-version=942
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
db-id=1
Automate generation of WAL and pg_control test files. The existing static files would not work with 32-bit or big-endian systems so create functions to generate these files dynamically rather than creating a bunch of new static files.
2017-11-19 03:02:54 +02:00
db-system-id=1000000000000000094
Database version refactoring. * Refactor db version constants into a separate module. * Update synthetic backup tests to PostgreSQL 9.4.
2016-08-12 04:35:24 +02:00
db-version="9.4"
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
[db:history]
Automate generation of WAL and pg_control test files. The existing static files would not work with 32-bit or big-endian systems so create functions to generate these files dynamically rather than creating a bunch of new static files.
2017-11-19 03:02:54 +02:00
1={"db-catalog-version":201409291,"db-control-version":942,"db-system-id":1000000000000000094,"db-version":"9.4"}
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
Improve performance of info file load/save. Info files required three copies in memory to be loaded (the original string, an ini representation, and the final info object). Not only was this memory inefficient but the Ini object does sequential scans when searching for keys making large files very slow to load. This has not been an issue since archive.info and backup.info are very small, but it becomes a big deal when loading manifests with hundreds of thousands of files. Instead of holding copies of the data in memory, use a callback to deliver the ini data directly to the object when loading. Use a similar method for save to avoid having an intermediate copy. Save is a bit complex because sections/keys must be written in alpha order or older versions of pgBackRest will not calculate the correct checksum. Also move the load retry logic to helper functions rather than embedding it in the Info object. This allows for more flexibility in loading and ensures that stack traces will be available when developing unit tests. Reviewed by Cynthia Shang.
2019-09-06 19:48:28 +02:00
[backrest]
backrest-checksum="[CHECKSUM]"
Major refactor of the test suite. * Make the code more modular and object-oriented. * Multiple Docker containers can now be created for a single test to simulate more realistic environments.
2016-06-24 14:12:58 +02:00
incr backup - update files (db-master host)
Reduce log level for all expect tests to detail. The C code is designed to be efficient rather than deterministic at the debug log level. As we move more testing from integration to unit tests it makes less sense to try and maintain the expect logs at this log level. Most of the expect logs have already been moved to detail level but mock/all still had tests at debug level. Change the logging defaults in the config file and remove as many references to log-level-console as possible.
2019-05-23 00:23:44 +02:00
> [CONTAINER-EXEC] db-master [BACKREST-BIN] --config=[TEST_PATH]/db-master/pgbackrest.conf --no-online --stanza=db backup
Log testing can now be enabled for certain deterministic tests. This works by comparing the generated logs against a previous copy. Currently only enabled for the backup/synthetic tests.
2015-05-05 19:08:48 +02:00
------------------------------------------------------------------------------------------------------------------------------------
Add infrastructure for multiple compression type support. Add compress-type option and deprecate compress option. Since the compress option is boolean it won't work with multiple compression types. Add logic to cfgLoadUpdateOption() to update compress-type if it is not set directly. The compress option should no longer be referenced outside the cfgLoadUpdateOption() function. Add common/compress/helper module to contain interface functions that work with multiple compression types. Code outside this module should no longer call specific compression drivers, though it may be OK to reference a specific compression type using the new interface (e.g., saving backup history files in gz format). Unit tests only test compression using the gz format because other formats may not be available in all builds. It is the job of integration tests to exercise all compression types. Additional compression types will be added in future commits.
2020-03-06 21:41:03 +02:00
P00 INFO: backup command begin [BACKREST-VERSION]: --buffer-size=[BUFFER-SIZE] --compress-level=3 --compress-type=none --config=[TEST_PATH]/db-master/pgbackrest.conf --db-timeout=45 --lock-path=[TEST_PATH]/db-master/lock --log-level-console=detail --log-level-file=[LOG-LEVEL-FILE] --log-level-stderr=off --log-path=[TEST_PATH]/db-master/log[] --no-log-timestamp --no-online --pg1-path=[TEST_PATH]/db-master/db/base-2 --protocol-timeout=60 --repo1-path=[TEST_PATH]/db-master/repo --stanza=db --start-fast
Add time-based retention for full backups. The --repo-retention-full-type option allows retention of full backups based on a time period, specified in days. The new option will default to 'count' and therefore will not affect current installations. Setting repo-retention-full-type to 'time' will allow the user to use a time period, in days, to indicate full backup retention. Using this method, a full backup can be expired only if the time the backup completed is older than the number of days set with repo-retention-full (calculated from the moment the 'expire' command is run) and at least one full backup meets the retention period. If archive retention has not been configured, then the default settings will expire archives that are prior to the oldest retained full backup. For example, if there are three full backups ending in times that are 25 days old (F1), 20 days old (F2) and 10 days old (F3), then if the full retention period is 15 days, then only F1 will be expired; F2 will be retained because F1 is not at least 15 days old.
2020-05-08 21:25:03 +02:00
P00 WARN: option 'repo1-retention-full' is not set for 'repo1-retention-full-type=count', the repository may run out of space
Rename retention-* options to repo-retention-*.
2018-02-19 22:07:24 +02:00
HINT: to retain full backups indefinitely (without warning), set option 'repo1-retention-full' to the maximum.
Show Process ID in log output instead of filtering it out with the timestamp.
2016-09-14 23:30:57 +02:00
P00 INFO: last backup label = [BACKUP-INCR-3], version = [VERSION-1]
Added the --checksum-page option. This option allows pgBackRest to validate page checksums in data files when checksums are enabled on PostgreSQL >= 9.3. Note that this functionality requires a C library which may not initially be available in OS packages. The option will automatically be enabled when the library is present and checksums are enabled on the cluster.
2016-12-13 01:54:07 +02:00
P00 WARN: incr backup cannot alter 'checksum-page' option to 'false', reset to 'true' from [BACKUP-INCR-3]
The backup command is implemented entirely in C. For the most part this is a direct migration of the Perl code into C except as noted below. A backup can now be initiated from a linked directory. The link will not be stored in the manifest or recreated on restore. If a link or directory does not already exist in the restore location then a directory will be created. The logic for creating backup labels has been improved and it should no longer be possible to get a backup label earlier than the latest backup even with timezone changes or clock skew. This has never been an issue in the field that we know of, but we found it in testing. For online backups all times are fetched from the PostgreSQL primary host (before only copy start was). This doesn't affect backup integrity but it does prevent clock skew between hosts affecting backup duration reporting. Archive copy now works as expected when the archive and backup have different compression settings, i.e. when one is compressed and the other is not. This was a long-standing bug in the Perl code. Resume will now work even if hardlink settings have been changed. Reviewed by Cynthia Shang.
2019-12-14 00:14:26 +02:00
P00 WARN: file 'base/16384/17000' has same timestamp as prior but different size, enabling delta checksum
Automatically enable backup checksum delta when anomalies (e.g. timeline switch) are detected. There are a number of cases where a checksum delta is more appropriate than the default time-based delta: * Timeline has switched since the prior backup * File timestamp is older than recorded in the prior backup * File size changed but timestamp did not * File timestamp is in the future compared to the start of the backup * Online option has changed since the prior backup A practical example is that checksum delta will be enabled after a failover to standby due to the timeline switch. In this case, timestamps can't be trusted and our recommendation has been to run a full backup, which can impact the retention schedule and requires manual intervention. Now, a checksum delta will be performed if the backup type is incr/diff. This means more CPU will be used during the backup but the backup size will be smaller and the retention schedule will not be impacted. Contributed by Cynthia Shang.
2018-11-01 17:31:25 +02:00
P01 DETAIL: match file from prior backup [TEST_PATH]/db-master/db/base-2/base/32768/33001 (64KB, 36%) checksum 6bf316f11d28c28914ea9be92c00de9bea6d9a6b
Fix incorrect offline upper bound for ignoring page checksum errors. For offline backups the upper bound was being set to 0x0000FFFF0000FFFF rather than UINT64_MAX. This meant that page checksum errors might be ignored for databases with a lot of past WAL in offline mode. Online mode is not affected since the upper bound is retrieved from pg_start_backup().
2019-07-11 15:13:56 +02:00
P01 DETAIL: match file from prior backup [TEST_PATH]/db-master/db/base-2/base/32768/44000_init (32KB, 54%) checksum 7a16d165e4775f7c92e8cdf60c0af57313f0bf90
P01 DETAIL: match file from prior backup [TEST_PATH]/db-master/db/base-2/base/32768/33000.32767 (32KB, 72%) checksum 6e99b589e550e68e934fd235ccba59fe5b592a9e
P01 DETAIL: match file from prior backup [TEST_PATH]/db-master/db/base-2/base/32768/33000 (32KB, 90%) checksum 7a16d165e4775f7c92e8cdf60c0af57313f0bf90
Automatically enable backup checksum delta when anomalies (e.g. timeline switch) are detected. There are a number of cases where a checksum delta is more appropriate than the default time-based delta: * Timeline has switched since the prior backup * File timestamp is older than recorded in the prior backup * File size changed but timestamp did not * File timestamp is in the future compared to the start of the backup * Online option has changed since the prior backup A practical example is that checksum delta will be enabled after a failover to standby due to the timeline switch. In this case, timestamps can't be trusted and our recommendation has been to run a full backup, which can impact the retention schedule and requires manual intervention. Now, a checksum delta will be performed if the backup type is incr/diff. This means more CPU will be used during the backup but the backup size will be smaller and the retention schedule will not be impacted. Contributed by Cynthia Shang.
2018-11-01 17:31:25 +02:00
P01 DETAIL: match file from prior backup [TEST_PATH]/db-master/db/base-2/global/pg_control (8KB, 95%) checksum 4c77c900f7af0d9ab13fa9982051a42e0b637f6c
P01 DETAIL: match file from prior backup [TEST_PATH]/db-master/db/base-2/base/1/12000 (8KB, 99%) checksum 22c98d248ff548311eda88559e4a8405ed77c003
P01 DETAIL: match file from prior backup [TEST_PATH]/db-master/db/base-2/postgresql.conf (21B, 99%) checksum 6721d92c9fcdf4248acff1f9a1377127d9064807
P01 DETAIL: match file from prior backup [TEST_PATH]/db-master/db/base-2/badchecksum.txt (11B, 99%) checksum f927212cd08d11a42a666b2f04235398e9ceeb51
P01 INFO: backup file [TEST_PATH]/db-master/db/base-2/base/16384/17000 (8B, 99%) checksum 9a53d532e27785e681766c98516a5e93f096a501
Added the --checksum-page option. This option allows pgBackRest to validate page checksums in data files when checksums are enabled on PostgreSQL >= 9.3. Note that this functionality requires a C library which may not initially be available in OS packages. The option will automatically be enabled when the library is present and checksums are enabled on the cluster.
2016-12-13 01:54:07 +02:00
P00 WARN: page misalignment in file [TEST_PATH]/db-master/db/base-2/base/16384/17000: file size 8 is not divisible by page size 8192
Automatically enable backup checksum delta when anomalies (e.g. timeline switch) are detected. There are a number of cases where a checksum delta is more appropriate than the default time-based delta: * Timeline has switched since the prior backup * File timestamp is older than recorded in the prior backup * File size changed but timestamp did not * File timestamp is in the future compared to the start of the backup * Online option has changed since the prior backup A practical example is that checksum delta will be enabled after a failover to standby due to the timeline switch. In this case, timestamps can't be trusted and our recommendation has been to run a full backup, which can impact the retention schedule and requires manual intervention. Now, a checksum delta will be performed if the backup type is incr/diff. This means more CPU will be used during the backup but the backup size will be smaller and the retention schedule will not be impacted. Contributed by Cynthia Shang.
2018-11-01 17:31:25 +02:00
P01 DETAIL: match file from prior backup [TEST_PATH]/db-master/db/base-2/changecontent.txt (7B, 99%) checksum a094d94583e209556d03c3c5da33131a065f1689
P01 DETAIL: match file from prior backup [TEST_PATH]/db-master/db/base-2/pg_stat/global.stat (5B, 99%) checksum e350d5ce0153f3e22d5db21cf2a4eff00f3ee877
P01 DETAIL: match file from prior backup [TEST_PATH]/db-master/db/base-2/base/base2.txt (5B, 99%) checksum 09b5e31766be1dba1ec27de82f975c1b6eea2a92
P01 DETAIL: match file from prior backup [TEST_PATH]/db-master/db/base-2/changetime.txt (4B, 99%) checksum 88087292ed82e26f3eb824d0bffc05ccf7a30f8d
P01 DETAIL: match file from prior backup [TEST_PATH]/db-master/db/base-2/base/32768/PG_VERSION (3B, 99%) checksum 184473f470864e067ee3a22e64b47b0a1c356f29
P01 DETAIL: match file from prior backup [TEST_PATH]/db-master/db/base-2/base/16384/PG_VERSION (3B, 99%) checksum 184473f470864e067ee3a22e64b47b0a1c356f29
P01 DETAIL: match file from prior backup [TEST_PATH]/db-master/db/base-2/base/1/PG_VERSION (3B, 99%) checksum 184473f470864e067ee3a22e64b47b0a1c356f29
P01 DETAIL: match file from prior backup [TEST_PATH]/db-master/db/base-2/PG_VERSION (3B, 99%) checksum 184473f470864e067ee3a22e64b47b0a1c356f29
P01 DETAIL: match file from prior backup [TEST_PATH]/db-master/db/base-2/pg_tblspc/2/[TS_PATH-1]/32768/tablespace2b.txt (8B, 99%) checksum e324463005236d83e6e54795dbddd20a74533bf3
P01 DETAIL: match file from prior backup [TEST_PATH]/db-master/db/base-2/pg_tblspc/2/[TS_PATH-1]/32768/tablespace2.txt (7B, 100%) checksum dc7f76e43c46101b47acc55ae4d593a9e6983578
The backup command is implemented entirely in C. For the most part this is a direct migration of the Perl code into C except as noted below. A backup can now be initiated from a linked directory. The link will not be stored in the manifest or recreated on restore. If a link or directory does not already exist in the restore location then a directory will be created. The logic for creating backup labels has been improved and it should no longer be possible to get a backup label earlier than the latest backup even with timezone changes or clock skew. This has never been an issue in the field that we know of, but we found it in testing. For online backups all times are fetched from the PostgreSQL primary host (before only copy start was). This doesn't affect backup integrity but it does prevent clock skew between hosts affecting backup duration reporting. Archive copy now works as expected when the archive and backup have different compression settings, i.e. when one is compressed and the other is not. This was a long-standing bug in the Perl code. Resume will now work even if hardlink settings have been changed. Reviewed by Cynthia Shang.
2019-12-14 00:14:26 +02:00
P00 DETAIL: reference pg_data/PG_VERSION to [BACKUP-FULL-2]
Remove integration tests that are now covered in the unit tests. Most of these tests are just checking that errors are thrown when required. These are well covered in various unit tests. The "cannot resume" tests are also well covered in the backup unit tests. Finally, config warnings are well covered in the config unit tests. There is more to be done here, but this accounts for the low-hanging fruit.
2019-12-18 03:14:45 +02:00
P00 DETAIL: reference pg_data/badchecksum.txt to [BACKUP-DIFF-1]
The backup command is implemented entirely in C. For the most part this is a direct migration of the Perl code into C except as noted below. A backup can now be initiated from a linked directory. The link will not be stored in the manifest or recreated on restore. If a link or directory does not already exist in the restore location then a directory will be created. The logic for creating backup labels has been improved and it should no longer be possible to get a backup label earlier than the latest backup even with timezone changes or clock skew. This has never been an issue in the field that we know of, but we found it in testing. For online backups all times are fetched from the PostgreSQL primary host (before only copy start was). This doesn't affect backup integrity but it does prevent clock skew between hosts affecting backup duration reporting. Archive copy now works as expected when the archive and backup have different compression settings, i.e. when one is compressed and the other is not. This was a long-standing bug in the Perl code. Resume will now work even if hardlink settings have been changed. Reviewed by Cynthia Shang.
2019-12-14 00:14:26 +02:00
P00 DETAIL: reference pg_data/base/1/12000 to [BACKUP-FULL-2]
P00 DETAIL: reference pg_data/base/1/PG_VERSION to [BACKUP-FULL-2]
P00 DETAIL: reference pg_data/base/16384/PG_VERSION to [BACKUP-FULL-2]
P00 DETAIL: reference pg_data/base/32768/33000 to [BACKUP-FULL-2]
P00 DETAIL: reference pg_data/base/32768/33000.32767 to [BACKUP-FULL-2]
P00 DETAIL: reference pg_data/base/32768/33001 to [BACKUP-FULL-2]
P00 DETAIL: reference pg_data/base/32768/44000_init to [BACKUP-FULL-2]
P00 DETAIL: reference pg_data/base/32768/PG_VERSION to [BACKUP-FULL-2]
P00 DETAIL: reference pg_data/base/base2.txt to [BACKUP-INCR-3]
Remove integration tests that are now covered in the unit tests. Most of these tests are just checking that errors are thrown when required. These are well covered in various unit tests. The "cannot resume" tests are also well covered in the backup unit tests. Finally, config warnings are well covered in the config unit tests. There is more to be done here, but this accounts for the low-hanging fruit.
2019-12-18 03:14:45 +02:00
P00 DETAIL: reference pg_data/changecontent.txt to [BACKUP-DIFF-1]
The backup command is implemented entirely in C. For the most part this is a direct migration of the Perl code into C except as noted below. A backup can now be initiated from a linked directory. The link will not be stored in the manifest or recreated on restore. If a link or directory does not already exist in the restore location then a directory will be created. The logic for creating backup labels has been improved and it should no longer be possible to get a backup label earlier than the latest backup even with timezone changes or clock skew. This has never been an issue in the field that we know of, but we found it in testing. For online backups all times are fetched from the PostgreSQL primary host (before only copy start was). This doesn't affect backup integrity but it does prevent clock skew between hosts affecting backup duration reporting. Archive copy now works as expected when the archive and backup have different compression settings, i.e. when one is compressed and the other is not. This was a long-standing bug in the Perl code. Resume will now work even if hardlink settings have been changed. Reviewed by Cynthia Shang.
2019-12-14 00:14:26 +02:00
P00 DETAIL: reference pg_data/changetime.txt to [BACKUP-FULL-2]
P00 DETAIL: reference pg_data/global/pg_control to [BACKUP-FULL-2]
P00 DETAIL: reference pg_data/pg_stat/global.stat to [BACKUP-FULL-2]
P00 DETAIL: reference pg_data/postgresql.conf to [BACKUP-FULL-2]
P00 DETAIL: reference pg_data/special-!_.*'()&!@;:+,? to [BACKUP-FULL-2]
P00 DETAIL: reference pg_data/zero_from_start to [BACKUP-FULL-2]
Remove integration tests that are now covered in the unit tests. Most of these tests are just checking that errors are thrown when required. These are well covered in various unit tests. The "cannot resume" tests are also well covered in the backup unit tests. Finally, config warnings are well covered in the config unit tests. There is more to be done here, but this accounts for the low-hanging fruit.
2019-12-18 03:14:45 +02:00
P00 DETAIL: reference pg_data/zerosize.txt to [BACKUP-DIFF-1]
P00 DETAIL: reference pg_tblspc/2/[TS_PATH-1]/32768/tablespace2.txt to [BACKUP-DIFF-1]
The backup command is implemented entirely in C. For the most part this is a direct migration of the Perl code into C except as noted below. A backup can now be initiated from a linked directory. The link will not be stored in the manifest or recreated on restore. If a link or directory does not already exist in the restore location then a directory will be created. The logic for creating backup labels has been improved and it should no longer be possible to get a backup label earlier than the latest backup even with timezone changes or clock skew. This has never been an issue in the field that we know of, but we found it in testing. For online backups all times are fetched from the PostgreSQL primary host (before only copy start was). This doesn't affect backup integrity but it does prevent clock skew between hosts affecting backup duration reporting. Archive copy now works as expected when the archive and backup have different compression settings, i.e. when one is compressed and the other is not. This was a long-standing bug in the Perl code. Resume will now work even if hardlink settings have been changed. Reviewed by Cynthia Shang.
2019-12-14 00:14:26 +02:00
P00 DETAIL: reference pg_tblspc/2/[TS_PATH-1]/32768/tablespace2b.txt to [BACKUP-INCR-3]
Automatically enable backup checksum delta when anomalies (e.g. timeline switch) are detected. There are a number of cases where a checksum delta is more appropriate than the default time-based delta: * Timeline has switched since the prior backup * File timestamp is older than recorded in the prior backup * File size changed but timestamp did not * File timestamp is in the future compared to the start of the backup * Online option has changed since the prior backup A practical example is that checksum delta will be enabled after a failover to standby due to the timeline switch. In this case, timestamps can't be trusted and our recommendation has been to run a full backup, which can impact the retention schedule and requires manual intervention. Now, a checksum delta will be performed if the backup type is incr/diff. This means more CPU will be used during the backup but the backup size will be smaller and the retention schedule will not be impacted. Contributed by Cynthia Shang.
2018-11-01 17:31:25 +02:00
P00 INFO: incr backup size = 176KB
Show Process ID in log output instead of filtering it out with the timestamp.
2016-09-14 23:30:57 +02:00
P00 INFO: new backup label = [BACKUP-INCR-4]
Added a predictable log entry to signal that a command has completed successfully. For example a backup ends successfully with: INFO: backup command end: completed successfully.
2016-12-04 21:44:53 +02:00
P00 INFO: backup command end: completed successfully
Integration test improvements for disk and memory efficiency. Set log-level-file=off when more that one test will run. In this case is it impossible to see the logs anyway since they will be automatically cleaned up after the test. This improves performance pretty dramatically since trace-level logging is expensive. If a singe integration test is run then log-level-file is trace by default but can be changed with the --log-level-test-file option. Reduce buffer-size to 64k to save memory during testing and allow more processes to run in parallel. Update log replacement rules so that these options can change without affecting expect logs.
2019-12-17 22:23:07 +02:00
P00 INFO: expire command begin [BACKREST-VERSION]: --buffer-size=[BUFFER-SIZE] --config=[TEST_PATH]/db-master/pgbackrest.conf --lock-path=[TEST_PATH]/db-master/lock --log-level-console=detail --log-level-file=[LOG-LEVEL-FILE] --log-level-stderr=off --log-path=[TEST_PATH]/db-master/log[] --no-log-timestamp --repo1-path=[TEST_PATH]/db-master/repo --stanza=db
Rename retention-* options to repo-retention-*.
2018-02-19 22:07:24 +02:00
P00 INFO: option 'repo1-retention-archive' is not set - archive logs will not be expired
Added a predictable log entry to signal that a command has completed successfully. For example a backup ends successfully with: INFO: backup command end: completed successfully.
2016-12-04 21:44:53 +02:00
P00 INFO: expire command end: completed successfully
Log testing can now be enabled for certain deterministic tests. This works by comparing the generated logs against a previous copy. Currently only enabled for the backup/synthetic tests.
2015-05-05 19:08:48 +02:00
Rename test paths for clarity. This was worked out as part of the test suite refactor [c8f806a] but not committed with it because of the large number of expect logs changes involved. Keeping them separate made it easier to audit the changes in the refactor.
2016-06-24 16:54:31 +02:00
+ supplemental file: [TEST_PATH]/db-master/pgbackrest.conf
----------------------------------------------------------
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
[db]
Rename db-* options to pg-* and backup-* options to repo-* to improve consistency. * repo-* options are now indexed although only one is allowed. * List deprecated option names in documentation and command-line help.
2018-02-04 01:27:38 +02:00
pg1-path=[TEST_PATH]/db-master/db/base-2
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
New simpler configuration and consistent project/exe/path naming. * The repo-path option now always refers to the repository where backups and archive are stored, whether local or remote, so the repo-remote-path option has been removed. The new spool-path option can be used to define a location for queueing WAL segments when archiving asynchronously. Otherwise, a local repository is no longer required. * Implemented a new config format which should be far simpler to use. See the User Guide and Configuration Reference for details but for a simple configuration all options can now be placed in the stanza section. Options that are shared between stanzas can be placed in the [global] section. More complex configurations can still make use of command sections though this should be a rare use case. * The default configuration filename is now pgbackrest.conf instead of pg_backrest.conf. This was done for consistency with other naming changes but also to prevent old config files from being loaded accidentally. * The default repository name was changed from /var/lib/backup to /var/lib/pgbackrest. * Lock files are now stored in /tmp/pgbackrest by default. These days /run/pgbackrest would be the preferred location but that would require init scripts which are not part of this release. The lock-path option can be used to configure the lock directory. * Log files are now stored in /var/log/pgbackrest by default and no longer have the date appended so they can be managed with logrotate. The log-path option can be used to configure the lock directory. * Executable filename changed from pg_backrest to pgbackrest.
2016-04-14 15:30:54 +02:00
[db:restore]
Rename test paths for clarity. This was worked out as part of the test suite refactor [c8f806a] but not committed with it because of the large number of expect logs changes involved. Keeping them separate made it easier to audit the changes in the refactor.
2016-06-24 16:54:31 +02:00
tablespace-map=1=[TEST_PATH]/db-master/db/tablespace/ts1-2
tablespace-map=2=[TEST_PATH]/db-master/db/tablespace/ts2-2
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
New simpler configuration and consistent project/exe/path naming. * The repo-path option now always refers to the repository where backups and archive are stored, whether local or remote, so the repo-remote-path option has been removed. The new spool-path option can be used to define a location for queueing WAL segments when archiving asynchronously. Otherwise, a local repository is no longer required. * Implemented a new config format which should be far simpler to use. See the User Guide and Configuration Reference for details but for a simple configuration all options can now be placed in the stanza section. Options that are shared between stanzas can be placed in the [global] section. More complex configurations can still make use of command sections though this should be a rare use case. * The default configuration filename is now pgbackrest.conf instead of pg_backrest.conf. This was done for consistency with other naming changes but also to prevent old config files from being loaded accidentally. * The default repository name was changed from /var/lib/backup to /var/lib/pgbackrest. * Lock files are now stored in /tmp/pgbackrest by default. These days /run/pgbackrest would be the preferred location but that would require init scripts which are not part of this release. The lock-path option can be used to configure the lock directory. * Log files are now stored in /var/log/pgbackrest by default and no longer have the date appended so they can be managed with logrotate. The log-path option can be used to configure the lock directory. * Executable filename changed from pg_backrest to pgbackrest.
2016-04-14 15:30:54 +02:00
[global]
Integration test improvements for disk and memory efficiency. Set log-level-file=off when more that one test will run. In this case is it impossible to see the logs anyway since they will be automatically cleaned up after the test. This improves performance pretty dramatically since trace-level logging is expensive. If a singe integration test is run then log-level-file is trace by default but can be changed with the --log-level-test-file option. Reduce buffer-size to 64k to save memory during testing and allow more processes to run in parallel. Update log replacement rules so that these options can change without affecting expect logs.
2019-12-17 22:23:07 +02:00
buffer-size=[BUFFER-SIZE]
Fix custom settings for compress-level option being ignored. Reported by Jens Wilke.
2017-10-23 00:05:46 +02:00
compress-level=3
Add infrastructure for multiple compression type support. Add compress-type option and deprecate compress option. Since the compress option is boolean it won't work with multiple compression types. Add logic to cfgLoadUpdateOption() to update compress-type if it is not set directly. The compress option should no longer be referenced outside the cfgLoadUpdateOption() function. Add common/compress/helper module to contain interface functions that work with multiple compression types. Code outside this module should no longer call specific compression drivers, though it may be OK to reference a specific compression type using the new interface (e.g., saving backup history files in gz format). Unit tests only test compression using the gz format because other formats may not be available in all builds. It is the job of integration tests to exercise all compression types. Additional compression types will be added in future commits.
2020-03-06 21:41:03 +02:00
compress-type=none
Storage and IO layer refactor: Refactor storage layer to allow for new repository filesystems using drivers. (Reviewed by Cynthia Shang.) Refactor IO layer to allow for new compression formats, checksum types, and other capabilities using filters. (Reviewed by Cynthia Shang.)
2017-06-09 23:51:41 +02:00
db-timeout=45
lock-path=[TEST_PATH]/db-master/lock
Reduce log level for all expect tests to detail. The C code is designed to be efficient rather than deterministic at the debug log level. As we move more testing from integration to unit tests it makes less sense to try and maintain the expect logs at this log level. Most of the expect logs have already been moved to detail level but mock/all still had tests at debug level. Change the logging defaults in the config file and remove as many references to log-level-console as possible.
2019-05-23 00:23:44 +02:00
log-level-console=detail
Integration test improvements for disk and memory efficiency. Set log-level-file=off when more that one test will run. In this case is it impossible to see the logs anyway since they will be automatically cleaned up after the test. This improves performance pretty dramatically since trace-level logging is expensive. If a singe integration test is run then log-level-file is trace by default but can be changed with the --log-level-test-file option. Reduce buffer-size to 64k to save memory during testing and allow more processes to run in parallel. Update log replacement rules so that these options can change without affecting expect logs.
2019-12-17 22:23:07 +02:00
log-level-file=[LOG-LEVEL-FILE]
Added the log-level-stderr option. Controls whether console log messages are sent to stderr or stdout. By default this is set to warn which represents a change in behavior from previous versions, even though it may be more intuitive. Setting log-level-stderr=off will preserve the old behavior. Suggested by Sascha Biberhofer.
2016-10-05 15:09:30 +02:00
log-level-stderr=off
Storage and IO layer refactor: Refactor storage layer to allow for new repository filesystems using drivers. (Reviewed by Cynthia Shang.) Refactor IO layer to allow for new compression formats, checksum types, and other capabilities using filters. (Reviewed by Cynthia Shang.)
2017-06-09 23:51:41 +02:00
log-path=[TEST_PATH]/db-master/log
Integration test improvements for disk and memory efficiency. Set log-level-file=off when more that one test will run. In this case is it impossible to see the logs anyway since they will be automatically cleaned up after the test. This improves performance pretty dramatically since trace-level logging is expensive. If a singe integration test is run then log-level-file is trace by default but can be changed with the --log-level-test-file option. Reduce buffer-size to 64k to save memory during testing and allow more processes to run in parallel. Update log replacement rules so that these options can change without affecting expect logs.
2019-12-17 22:23:07 +02:00
log-subprocess=[LOG-SUBPROCESS]
Set log-timestamp=n for integration tests. This means less filtering of logs needs to be done and new timestamps can be added without adding new filters.
2018-05-11 17:24:38 +02:00
log-timestamp=n
Storage and IO layer refactor: Refactor storage layer to allow for new repository filesystems using drivers. (Reviewed by Cynthia Shang.) Refactor IO layer to allow for new compression formats, checksum types, and other capabilities using filters. (Reviewed by Cynthia Shang.)
2017-06-09 23:51:41 +02:00
protocol-timeout=60
Rename db-* options to pg-* and backup-* options to repo-* to improve consistency. * repo-* options are now indexed although only one is allowed. * List deprecated option names in documentation and command-line help.
2018-02-04 01:27:38 +02:00
repo1-path=[TEST_PATH]/db-master/repo
Consolidate archive-push tests into new archive test.
2017-06-21 14:38:53 +02:00
spool-path=[TEST_PATH]/db-master/spool
New simpler configuration and consistent project/exe/path naming. * The repo-path option now always refers to the repository where backups and archive are stored, whether local or remote, so the repo-remote-path option has been removed. The new spool-path option can be used to define a location for queueing WAL segments when archiving asynchronously. Otherwise, a local repository is no longer required. * Implemented a new config format which should be far simpler to use. See the User Guide and Configuration Reference for details but for a simple configuration all options can now be placed in the stanza section. Options that are shared between stanzas can be placed in the [global] section. More complex configurations can still make use of command sections though this should be a rare use case. * The default configuration filename is now pgbackrest.conf instead of pg_backrest.conf. This was done for consistency with other naming changes but also to prevent old config files from being loaded accidentally. * The default repository name was changed from /var/lib/backup to /var/lib/pgbackrest. * Lock files are now stored in /tmp/pgbackrest by default. These days /run/pgbackrest would be the preferred location but that would require init scripts which are not part of this release. The lock-path option can be used to configure the lock directory. * Log files are now stored in /var/log/pgbackrest by default and no longer have the date appended so they can be managed with logrotate. The log-path option can be used to configure the lock directory. * Executable filename changed from pg_backrest to pgbackrest.
2016-04-14 15:30:54 +02:00
[global:backup]
archive-copy=y
start-fast=y
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
Rename test paths for clarity. This was worked out as part of the test suite refactor [c8f806a] but not committed with it because of the large number of expect logs changes involved. Keeping them separate made it easier to audit the changes in the refactor.
2016-06-24 16:54:31 +02:00
+ supplemental file: [TEST_PATH]/db-master/repo/backup/db/[BACKUP-INCR-4]/backup.manifest
-----------------------------------------------------------------------------------------
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
[backrest]
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
backrest-format=5
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
backrest-version="[VERSION-1]"
[backup]
backup-label="[BACKUP-INCR-4]"
backup-prior="[BACKUP-INCR-3]"
backup-timestamp-copy-start=[TIMESTAMP]
backup-timestamp-start=[TIMESTAMP]
backup-timestamp-stop=[TIMESTAMP]
backup-type="incr"
[backup:db]
Database version refactoring. * Refactor db version constants into a separate module. * Update synthetic backup tests to PostgreSQL 9.4.
2016-08-12 04:35:24 +02:00
db-catalog-version=201409291
db-control-version=942
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
db-id=1
Automate generation of WAL and pg_control test files. The existing static files would not work with 32-bit or big-endian systems so create functions to generate these files dynamically rather than creating a bunch of new static files.
2017-11-19 03:02:54 +02:00
db-system-id=1000000000000000094
Database version refactoring. * Refactor db version constants into a separate module. * Update synthetic backup tests to PostgreSQL 9.4.
2016-08-12 04:35:24 +02:00
db-version="9.4"
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
[backup:option]
Set archive-check option in manifest correctly when offline. Archive check does not run when in offline backup mode but the option was set to true in the manifest. It's harmless since these options are informational only but it could cause confusion when debugging.
2019-11-28 15:27:21 +02:00
option-archive-check=false
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
option-archive-copy=true
Backup from a standby cluster. A connection to the primary cluster is still required to start/stop the backup and copy files that are not replicated, but the vast majority of files are copied from the standby in order to reduce load on the master.
2016-08-25 17:25:46 +02:00
option-backup-standby=false
Integration test improvements for disk and memory efficiency. Set log-level-file=off when more that one test will run. In this case is it impossible to see the logs anyway since they will be automatically cleaned up after the test. This improves performance pretty dramatically since trace-level logging is expensive. If a singe integration test is run then log-level-file is trace by default but can be changed with the --log-level-test-file option. Reduce buffer-size to 64k to save memory during testing and allow more processes to run in parallel. Update log replacement rules so that these options can change without affecting expect logs.
2019-12-17 22:23:07 +02:00
option-buffer-size=[BUFFER-SIZE]
Added the --checksum-page option. This option allows pgBackRest to validate page checksums in data files when checksums are enabled on PostgreSQL >= 9.3. Note that this functionality requires a C library which may not initially be available in OS packages. The option will automatically be enabled when the library is present and checksums are enabled on the cluster.
2016-12-13 01:54:07 +02:00
option-checksum-page=true
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
option-compress=false
Add additional options to backup.manifest for debugging purposes. Add the buffer-size, compress-level, compress-level-network, and process-max options to the backup:option section in backup.manifest to aid in debugging. It may also make sense to propagate these options up to backup.info so they can be displayed in the info command, but for now this is deemed sufficient. Contributed by blogh.
2019-03-10 11:03:52 +02:00
option-compress-level=3
option-compress-level-network=3
Add infrastructure for multiple compression type support. Add compress-type option and deprecate compress option. Since the compress option is boolean it won't work with multiple compression types. Add logic to cfgLoadUpdateOption() to update compress-type if it is not set directly. The compress option should no longer be referenced outside the cfgLoadUpdateOption() function. Add common/compress/helper module to contain interface functions that work with multiple compression types. Code outside this module should no longer call specific compression drivers, though it may be OK to reference a specific compression type using the new interface (e.g., saving backup history files in gz format). Unit tests only test compression using the gz format because other formats may not be available in all builds. It is the job of integration tests to exercise all compression types. Additional compression types will be added in future commits.
2020-03-06 21:41:03 +02:00
option-compress-type="none"
Automatically enable backup checksum delta when anomalies (e.g. timeline switch) are detected. There are a number of cases where a checksum delta is more appropriate than the default time-based delta: * Timeline has switched since the prior backup * File timestamp is older than recorded in the prior backup * File size changed but timestamp did not * File timestamp is in the future compared to the start of the backup * Online option has changed since the prior backup A practical example is that checksum delta will be enabled after a failover to standby due to the timeline switch. In this case, timestamps can't be trusted and our recommendation has been to run a full backup, which can impact the retention schedule and requires manual intervention. Now, a checksum delta will be performed if the backup type is incr/diff. This means more CPU will be used during the backup but the backup size will be smaller and the retention schedule will not be impacted. Contributed by Cynthia Shang.
2018-11-01 17:31:25 +02:00
option-delta=true
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
option-hardlink=false
Closed #167: Rename `--no-start-stop` option to `--no-online`.
2016-02-20 02:32:02 +02:00
option-online=false
Add additional options to backup.manifest for debugging purposes. Add the buffer-size, compress-level, compress-level-network, and process-max options to the backup:option section in backup.manifest to aid in debugging. It may also make sense to propagate these options up to backup.info so they can be displayed in the info command, but for now this is deemed sufficient. Contributed by blogh.
2019-03-10 11:03:52 +02:00
option-process-max=1
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
[backup:target]
Rename test paths for clarity. This was worked out as part of the test suite refactor [c8f806a] but not committed with it because of the large number of expect logs changes involved. Keeping them separate made it easier to audit the changes in the refactor.
2016-06-24 16:54:31 +02:00
pg_data={"path":"[TEST_PATH]/db-master/db/base-2","type":"path"}
pg_tblspc/2={"path":"[TEST_PATH]/db-master/db/tablespace/ts2-2","tablespace-id":"2","tablespace-name":"ts2","type":"link"}
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
[target:file]
Fixed a issue where tablespaces were copied from the master during standby backup.
2016-09-04 15:19:44 +02:00
pg_data/PG_VERSION={"checksum":"184473f470864e067ee3a22e64b47b0a1c356f29","master":true,"reference":"[BACKUP-FULL-2]","size":[SIZE],"timestamp":[TIMESTAMP-1]}
Remove integration tests that are now covered in the unit tests. Most of these tests are just checking that errors are thrown when required. These are well covered in various unit tests. The "cannot resume" tests are also well covered in the backup unit tests. Finally, config warnings are well covered in the config unit tests. There is more to be done here, but this accounts for the low-hanging fruit.
2019-12-18 03:14:45 +02:00
pg_data/badchecksum.txt={"checksum":"f927212cd08d11a42a666b2f04235398e9ceeb51","master":true,"reference":"[BACKUP-DIFF-1]","size":[SIZE],"timestamp":[TIMESTAMP-1]}
Added the --checksum-page option. This option allows pgBackRest to validate page checksums in data files when checksums are enabled on PostgreSQL >= 9.3. Note that this functionality requires a C library which may not initially be available in OS packages. The option will automatically be enabled when the library is present and checksums are enabled on the cluster.
2016-12-13 01:54:07 +02:00
pg_data/base/1/12000={"checksum":"22c98d248ff548311eda88559e4a8405ed77c003","checksum-page":true,"reference":"[BACKUP-FULL-2]","size":[SIZE],"timestamp":[TIMESTAMP-1]}
Database version refactoring. * Refactor db version constants into a separate module. * Update synthetic backup tests to PostgreSQL 9.4.
2016-08-12 04:35:24 +02:00
pg_data/base/1/PG_VERSION={"checksum":"184473f470864e067ee3a22e64b47b0a1c356f29","mode":"0660","reference":"[BACKUP-FULL-2]","size":[SIZE],"timestamp":[TIMESTAMP-1]}
Added the --checksum-page option. This option allows pgBackRest to validate page checksums in data files when checksums are enabled on PostgreSQL >= 9.3. Note that this functionality requires a C library which may not initially be available in OS packages. The option will automatically be enabled when the library is present and checksums are enabled on the cluster.
2016-12-13 01:54:07 +02:00
pg_data/base/16384/17000={"checksum":"9a53d532e27785e681766c98516a5e93f096a501","checksum-page":false,"size":[SIZE],"timestamp":[TIMESTAMP-1]}
Database version refactoring. * Refactor db version constants into a separate module. * Update synthetic backup tests to PostgreSQL 9.4.
2016-08-12 04:35:24 +02:00
pg_data/base/16384/PG_VERSION={"checksum":"184473f470864e067ee3a22e64b47b0a1c356f29","reference":"[BACKUP-FULL-2]","size":[SIZE],"timestamp":[TIMESTAMP-1]}
Fix incorrect offline upper bound for ignoring page checksum errors. For offline backups the upper bound was being set to 0x0000FFFF0000FFFF rather than UINT64_MAX. This meant that page checksum errors might be ignored for databases with a lot of past WAL in offline mode. Online mode is not affected since the upper bound is retrieved from pg_start_backup().
2019-07-11 15:13:56 +02:00
pg_data/base/32768/33000={"checksum":"7a16d165e4775f7c92e8cdf60c0af57313f0bf90","checksum-page":true,"reference":"[BACKUP-FULL-2]","size":[SIZE],"timestamp":[TIMESTAMP-1]}
pg_data/base/32768/33000.32767={"checksum":"6e99b589e550e68e934fd235ccba59fe5b592a9e","checksum-page":true,"reference":"[BACKUP-FULL-2]","size":[SIZE],"timestamp":[TIMESTAMP-1]}
Added the --checksum-page option. This option allows pgBackRest to validate page checksums in data files when checksums are enabled on PostgreSQL >= 9.3. Note that this functionality requires a C library which may not initially be available in OS packages. The option will automatically be enabled when the library is present and checksums are enabled on the cluster.
2016-12-13 01:54:07 +02:00
pg_data/base/32768/33001={"checksum":"6bf316f11d28c28914ea9be92c00de9bea6d9a6b","checksum-page":false,"checksum-page-error":[0,[3,5],7],"reference":"[BACKUP-FULL-2]","size":[SIZE],"timestamp":[TIMESTAMP-1]}
Fix incorrect offline upper bound for ignoring page checksum errors. For offline backups the upper bound was being set to 0x0000FFFF0000FFFF rather than UINT64_MAX. This meant that page checksum errors might be ignored for databases with a lot of past WAL in offline mode. Online mode is not affected since the upper bound is retrieved from pg_start_backup().
2019-07-11 15:13:56 +02:00
pg_data/base/32768/44000_init={"checksum":"7a16d165e4775f7c92e8cdf60c0af57313f0bf90","checksum-page":true,"reference":"[BACKUP-FULL-2]","size":[SIZE],"timestamp":[TIMESTAMP-1]}
Database version refactoring. * Refactor db version constants into a separate module. * Update synthetic backup tests to PostgreSQL 9.4.
2016-08-12 04:35:24 +02:00
pg_data/base/32768/PG_VERSION={"checksum":"184473f470864e067ee3a22e64b47b0a1c356f29","reference":"[BACKUP-FULL-2]","size":[SIZE],"timestamp":[TIMESTAMP-1]}
Update Perl page checksum expression. This expression determines which files contain page checksums but it was also including the directory above the relation directories. In a real PostgreSQL installation this not a problem because these directories don't contain any files. However, our tests place a file in `base` which the Perl code thought should have page checksums while the new C code says no. Update the expression to document the change and avoid churn in the expect logs later.
2019-11-25 14:37:09 +02:00
pg_data/base/base2.txt={"checksum":"09b5e31766be1dba1ec27de82f975c1b6eea2a92","reference":"[BACKUP-INCR-3]","size":[SIZE],"timestamp":[TIMESTAMP-1]}
Remove integration tests that are now covered in the unit tests. Most of these tests are just checking that errors are thrown when required. These are well covered in various unit tests. The "cannot resume" tests are also well covered in the backup unit tests. Finally, config warnings are well covered in the config unit tests. There is more to be done here, but this accounts for the low-hanging fruit.
2019-12-18 03:14:45 +02:00
pg_data/changecontent.txt={"checksum":"a094d94583e209556d03c3c5da33131a065f1689","master":true,"reference":"[BACKUP-DIFF-1]","size":[SIZE],"timestamp":[TIMESTAMP-1]}
Add checksum delta for incremental backups. Use checksums rather than timestamps to determine if files have changed. This is useful in cases where the timestamps may not be trustworthy, e.g. when performing an incremental after failing over to a standby. If checksum delta is enabled then checksums will be used for verification of resumed backups, even if they are full. Resumes have always used checksums to verify the files in the repository, enabling delta performs checksums on the database files as well. Note that the user must manually enable this feature in cases were it would be useful or just keep in enabled all the time. A future commit will address automatically enabling the feature in cases where it seems likely to be useful. Contributed by Cynthia Shang.
2018-09-19 17:12:45 +02:00
pg_data/changetime.txt={"checksum":"88087292ed82e26f3eb824d0bffc05ccf7a30f8d","master":true,"reference":"[BACKUP-FULL-2]","size":[SIZE],"timestamp":[TIMESTAMP-2]}
Support configurable WAL segment size. PostgreSQL 11 introduces configurable WAL segment sizes, from 1MB to 1GB. There are two areas that needed to be updated to support this: building the archive-get queue and checking that WAL has been archived after a backup. Both operations require the WAL segment size to properly build a list. Checking the archive after a backup is still implemented in Perl and has an active database connection, so just get the WAL segment size from the database. The archive-get command does not have a connection to the database, so get the WAL segment size from pg_control instead. This requires a deeper inspection of pg_control than has been done in the past, so it seemed best to copy the relevant data structures from each version of PostgreSQL and build a generic interface layer to address them. While this approach is a bit verbose, it has the advantage of being relatively simple, and can easily be updated for new versions of PostgreSQL. Since the integration tests generate pg_control files for testing, teach Perl how to generate files with the correct offsets for both 32-bit and 64-bit architectures.
2018-09-25 11:24:42 +02:00
pg_data/global/pg_control={"checksum":"4c77c900f7af0d9ab13fa9982051a42e0b637f6c","master":true,"reference":"[BACKUP-FULL-2]","size":[SIZE],"timestamp":[TIMESTAMP-2]}
Fixed a issue where tablespaces were copied from the master during standby backup.
2016-09-04 15:19:44 +02:00
pg_data/pg_stat/global.stat={"checksum":"e350d5ce0153f3e22d5db21cf2a4eff00f3ee877","master":true,"reference":"[BACKUP-FULL-2]","size":[SIZE],"timestamp":[TIMESTAMP-2]}
pg_data/postgresql.conf={"checksum":"6721d92c9fcdf4248acff1f9a1377127d9064807","master":true,"reference":"[BACKUP-FULL-2]","size":[SIZE],"timestamp":[TIMESTAMP-2]}
Use minio for integration testing. ScalityS3 has not received any maintenance in years and is slow to start which is bad for testing. Replace it with minio which starts quickly and ships as a single executable or a tiny container. Minio has stricter limits on allowable characters but should still provide enough coverage to show that our encoding is working correctly. This commit also includes the upgrade to openssl 1.1.1 in the Ubuntu 18.04 container.
2019-07-03 04:20:35 +02:00
pg_data/special-!_.*'()&!@;:+,?={"master":true,"reference":"[BACKUP-FULL-2]","size":[SIZE],"timestamp":[TIMESTAMP-1]}
Allow zero-size files in backup manifest to reference a prior manifest regardless of timestamp delta. Contributed by Cynthia Shang.
2018-08-24 22:50:33 +02:00
pg_data/zero_from_start={"master":true,"reference":"[BACKUP-FULL-2]","size":[SIZE],"timestamp":[TIMESTAMP-1]}
Remove integration tests that are now covered in the unit tests. Most of these tests are just checking that errors are thrown when required. These are well covered in various unit tests. The "cannot resume" tests are also well covered in the backup unit tests. Finally, config warnings are well covered in the config unit tests. There is more to be done here, but this accounts for the low-hanging fruit.
2019-12-18 03:14:45 +02:00
pg_data/zerosize.txt={"master":true,"reference":"[BACKUP-DIFF-1]","size":[SIZE],"timestamp":[TIMESTAMP-2]}
pg_tblspc/2/[TS_PATH-1]/32768/tablespace2.txt={"checksum":"dc7f76e43c46101b47acc55ae4d593a9e6983578","checksum-page":false,"reference":"[BACKUP-DIFF-1]","size":[SIZE],"timestamp":[TIMESTAMP-1]}
Added the --checksum-page option. This option allows pgBackRest to validate page checksums in data files when checksums are enabled on PostgreSQL >= 9.3. Note that this functionality requires a C library which may not initially be available in OS packages. The option will automatically be enabled when the library is present and checksums are enabled on the cluster.
2016-12-13 01:54:07 +02:00
pg_tblspc/2/[TS_PATH-1]/32768/tablespace2b.txt={"checksum":"e324463005236d83e6e54795dbddd20a74533bf3","checksum-page":false,"reference":"[BACKUP-INCR-3]","size":[SIZE],"timestamp":[TIMESTAMP-1]}
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
[target:file:default]
Continuous integration using travis-ci.
2017-02-21 15:59:23 +02:00
group="[GROUP-1]"
Fixed a issue where tablespaces were copied from the master during standby backup.
2016-09-04 15:19:44 +02:00
master=false
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
mode="0600"
Regression tests can now be run as any properly-configured user, not just vagrant.
2016-10-01 19:39:44 +02:00
user="[USER-1]"
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
[target:link]
Rename test paths for clarity. This was worked out as part of the test suite refactor [c8f806a] but not committed with it because of the large number of expect logs changes involved. Keeping them separate made it easier to audit the changes in the refactor.
2016-06-24 16:54:31 +02:00
pg_data/pg_tblspc/2={"destination":"[TEST_PATH]/db-master/db/tablespace/ts2-2"}
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
[target:link:default]
Continuous integration using travis-ci.
2017-02-21 15:59:23 +02:00
group="[GROUP-1]"
Regression tests can now be run as any properly-configured user, not just vagrant.
2016-10-01 19:39:44 +02:00
user="[USER-1]"
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
[target:path]
The restore command is implemented entirely in C. For the most part this is a direct migration of the Perl code into C. There is one important behavioral change with regard to how file permissions are handled. The Perl code tried to set ownership as it was in the manifest even when running as an unprivileged user. This usually just led to errors and frustration. The C code works like this: If a restore is run as a non-root user (the typical scenario) then all files restored will belong to the user/group executing pgBackRest. If existing files are not owned by the executing user/group then an error will result if the ownership cannot be updated to the executing user/group. In that case the file ownership will need to be updated by a privileged user before the restore can be retried. If a restore is run as the root user then pgBackRest will attempt to recreate the ownership recorded in the manifest when the backup was made. Only user/group names are stored in the manifest so the same names must exist on the restore host for this to work. If the user/group name cannot be found locally then the user/group of the PostgreSQL data directory will be used and finally root if the data directory user/group cannot be mapped to a name. Reviewed by Cynthia Shang.
2019-09-26 13:52:02 +02:00
pg_data={"mode":"0777"}
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
pg_data/base={}
Allow selective restore of databases from a cluster backup. This feature can result in major space and time savings when only specific databases are restored. Unrestored databases will not be accessible but must be manually dropped before they will be removed from the shared catalogue.
2016-05-11 15:21:39 +02:00
pg_data/base/1={}
pg_data/base/16384={}
pg_data/base/32768={}
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
pg_data/global={}
pg_data/pg_clog={}
Exclude directories during backup that are cleaned, recreated, or zeroed by PostgreSQL at startup. These include (depending on the version where they were introduced): pgsql_tmp, pg_dynshmem, pg_notify, pg_replslot, pg_serial, pg_snapshots, pg_stat_tmp, pg_subtrans. The postgresql.auto.conf.tmp file is now excluded in addition to files that were already excluded: backup_label.old, postmaster.opts, postmaster.pid, recovery.conf, recovery.done.
2016-08-16 15:35:16 +02:00
pg_data/pg_dynshmem={}
Allow arbitrary directories and/or files to be excluded from a backup. Misuse of this feature can lead to inconsistent backups so read the --exclude documentation carefully before using.
2018-08-27 21:51:05 +02:00
pg_data/pg_log={}
Exclude directories during backup that are cleaned, recreated, or zeroed by PostgreSQL at startup. These include (depending on the version where they were introduced): pgsql_tmp, pg_dynshmem, pg_notify, pg_replslot, pg_serial, pg_snapshots, pg_stat_tmp, pg_subtrans. The postgresql.auto.conf.tmp file is now excluded in addition to files that were already excluded: backup_label.old, postmaster.opts, postmaster.pid, recovery.conf, recovery.done.
2016-08-16 15:35:16 +02:00
pg_data/pg_notify={}
pg_data/pg_replslot={}
pg_data/pg_serial={}
pg_data/pg_snapshots={}
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
pg_data/pg_stat={}
Exclude directories during backup that are cleaned, recreated, or zeroed by PostgreSQL at startup. These include (depending on the version where they were introduced): pgsql_tmp, pg_dynshmem, pg_notify, pg_replslot, pg_serial, pg_snapshots, pg_stat_tmp, pg_subtrans. The postgresql.auto.conf.tmp file is now excluded in addition to files that were already excluded: backup_label.old, postmaster.opts, postmaster.pid, recovery.conf, recovery.done.
2016-08-16 15:35:16 +02:00
pg_data/pg_stat_tmp={}
pg_data/pg_subtrans={}
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
pg_data/pg_tblspc={}
The restore command is implemented entirely in C. For the most part this is a direct migration of the Perl code into C. There is one important behavioral change with regard to how file permissions are handled. The Perl code tried to set ownership as it was in the manifest even when running as an unprivileged user. This usually just led to errors and frustration. The C code works like this: If a restore is run as a non-root user (the typical scenario) then all files restored will belong to the user/group executing pgBackRest. If existing files are not owned by the executing user/group then an error will result if the ownership cannot be updated to the executing user/group. In that case the file ownership will need to be updated by a privileged user before the restore can be retried. If a restore is run as the root user then pgBackRest will attempt to recreate the ownership recorded in the manifest when the backup was made. Only user/group names are stored in the manifest so the same names must exist on the restore host for this to work. If the user/group name cannot be found locally then the user/group of the PostgreSQL data directory will be used and finally root if the data directory user/group cannot be mapped to a name. Reviewed by Cynthia Shang.
2019-09-26 13:52:02 +02:00
pg_tblspc={"mode":"0777"}
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
pg_tblspc/2={}
pg_tblspc/2/[TS_PATH-1]={}
Allow selective restore of databases from a cluster backup. This feature can result in major space and time savings when only specific databases are restored. Unrestored databases will not be accessible but must be manually dropped before they will be removed from the shared catalogue.
2016-05-11 15:21:39 +02:00
pg_tblspc/2/[TS_PATH-1]/32768={}
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
[target:path:default]
Continuous integration using travis-ci.
2017-02-21 15:59:23 +02:00
group="[GROUP-1]"
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
mode="0700"
Regression tests can now be run as any properly-configured user, not just vagrant.
2016-10-01 19:39:44 +02:00
user="[USER-1]"
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
Move info file checksum to the end of the file. Putting the checksum at the beginning of the file made it impossible to stream the file out when saving. The entire file had to be held in memory while it was checksummed so the checksum could be written at the beginning. Instead place the checksum at the end. This does not break the existing Perl or C code since the read is not order dependent. There are no plans to improve the Perl code to take advantage of this change, but it will make the C implementation more efficient. Reviewed by Cynthia Shang.
2019-08-22 01:45:48 +02:00
[backrest]
backrest-checksum="[CHECKSUM]"
Rename test paths for clarity. This was worked out as part of the test suite refactor [c8f806a] but not committed with it because of the large number of expect logs changes involved. Keeping them separate made it easier to audit the changes in the refactor.
2016-06-24 16:54:31 +02:00
+ supplemental file: [TEST_PATH]/db-master/repo/backup/db/backup.info
---------------------------------------------------------------------
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
[backrest]
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
backrest-format=5
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
backrest-version="[VERSION-1]"
[backup:current]
Set archive-check option in manifest correctly when offline. Archive check does not run when in offline backup mode but the option was set to true in the manifest. It's harmless since these options are informational only but it could cause confusion when debugging.
2019-11-28 15:27:21 +02:00
[BACKUP-FULL-2]={"backrest-format":5,"backrest-version":"[VERSION-1]","backup-archive-start":null,"backup-archive-stop":null,"backup-info-repo-size":[SIZE],"backup-info-repo-size-delta":[DELTA],"backup-info-size":[SIZE],"backup-info-size-delta":[DELTA],"backup-timestamp-start":[TIMESTAMP],"backup-timestamp-stop":[TIMESTAMP],"backup-type":"full","db-id":1,"option-archive-check":false,"option-archive-copy":true,"option-backup-standby":false,"option-checksum-page":true,"option-compress":false,"option-hardlink":false,"option-online":false}
Remove integration tests that are now covered in the unit tests. Most of these tests are just checking that errors are thrown when required. These are well covered in various unit tests. The "cannot resume" tests are also well covered in the backup unit tests. Finally, config warnings are well covered in the config unit tests. There is more to be done here, but this accounts for the low-hanging fruit.
2019-12-18 03:14:45 +02:00
[BACKUP-INCR-2]={"backrest-format":5,"backrest-version":"[VERSION-1]","backup-archive-start":null,"backup-archive-stop":null,"backup-info-repo-size":[SIZE],"backup-info-repo-size-delta":[DELTA],"backup-info-size":[SIZE],"backup-info-size-delta":[DELTA],"backup-prior":"[BACKUP-FULL-2]","backup-reference":["[BACKUP-FULL-2]"],"backup-timestamp-start":[TIMESTAMP],"backup-timestamp-stop":[TIMESTAMP],"backup-type":"incr","db-id":1,"option-archive-check":false,"option-archive-copy":true,"option-backup-standby":false,"option-checksum-page":true,"option-compress":false,"option-hardlink":false,"option-online":false}
[BACKUP-DIFF-1]={"backrest-format":5,"backrest-version":"[VERSION-1]","backup-archive-start":null,"backup-archive-stop":null,"backup-info-repo-size":[SIZE],"backup-info-repo-size-delta":[DELTA],"backup-info-size":[SIZE],"backup-info-size-delta":[DELTA],"backup-prior":"[BACKUP-FULL-2]","backup-reference":["[BACKUP-FULL-2]"],"backup-timestamp-start":[TIMESTAMP],"backup-timestamp-stop":[TIMESTAMP],"backup-type":"diff","db-id":1,"option-archive-check":false,"option-archive-copy":true,"option-backup-standby":false,"option-checksum-page":true,"option-compress":false,"option-hardlink":false,"option-online":false}
[BACKUP-INCR-3]={"backrest-format":5,"backrest-version":"[VERSION-1]","backup-archive-start":null,"backup-archive-stop":null,"backup-info-repo-size":[SIZE],"backup-info-repo-size-delta":[DELTA],"backup-info-size":[SIZE],"backup-info-size-delta":[DELTA],"backup-prior":"[BACKUP-DIFF-1]","backup-reference":["[BACKUP-FULL-2]","[BACKUP-DIFF-1]"],"backup-timestamp-start":[TIMESTAMP],"backup-timestamp-stop":[TIMESTAMP],"backup-type":"incr","db-id":1,"option-archive-check":false,"option-archive-copy":true,"option-backup-standby":false,"option-checksum-page":true,"option-compress":false,"option-hardlink":false,"option-online":false}
[BACKUP-INCR-4]={"backrest-format":5,"backrest-version":"[VERSION-1]","backup-archive-start":null,"backup-archive-stop":null,"backup-info-repo-size":[SIZE],"backup-info-repo-size-delta":[DELTA],"backup-info-size":[SIZE],"backup-info-size-delta":[DELTA],"backup-prior":"[BACKUP-INCR-3]","backup-reference":["[BACKUP-FULL-2]","[BACKUP-DIFF-1]","[BACKUP-INCR-3]"],"backup-timestamp-start":[TIMESTAMP],"backup-timestamp-stop":[TIMESTAMP],"backup-type":"incr","db-id":1,"option-archive-check":false,"option-archive-copy":true,"option-backup-standby":false,"option-checksum-page":true,"option-compress":false,"option-hardlink":false,"option-online":false}
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
[db]
Database version refactoring. * Refactor db version constants into a separate module. * Update synthetic backup tests to PostgreSQL 9.4.
2016-08-12 04:35:24 +02:00
db-catalog-version=201409291
db-control-version=942
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
db-id=1
Automate generation of WAL and pg_control test files. The existing static files would not work with 32-bit or big-endian systems so create functions to generate these files dynamically rather than creating a bunch of new static files.
2017-11-19 03:02:54 +02:00
db-system-id=1000000000000000094
Database version refactoring. * Refactor db version constants into a separate module. * Update synthetic backup tests to PostgreSQL 9.4.
2016-08-12 04:35:24 +02:00
db-version="9.4"
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
[db:history]
Automate generation of WAL and pg_control test files. The existing static files would not work with 32-bit or big-endian systems so create functions to generate these files dynamically rather than creating a bunch of new static files.
2017-11-19 03:02:54 +02:00
1={"db-catalog-version":201409291,"db-control-version":942,"db-system-id":1000000000000000094,"db-version":"9.4"}
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
Improve performance of info file load/save. Info files required three copies in memory to be loaded (the original string, an ini representation, and the final info object). Not only was this memory inefficient but the Ini object does sequential scans when searching for keys making large files very slow to load. This has not been an issue since archive.info and backup.info are very small, but it becomes a big deal when loading manifests with hundreds of thousands of files. Instead of holding copies of the data in memory, use a callback to deliver the ini data directly to the object when loading. Use a similar method for save to avoid having an intermediate copy. Save is a bit complex because sections/keys must be written in alpha order or older versions of pgBackRest will not calculate the correct checksum. Also move the load retry logic to helper functions rather than embedding it in the Info object. This allows for more flexibility in loading and ensures that stack traces will be available when developing unit tests. Reviewed by Cynthia Shang.
2019-09-06 19:48:28 +02:00
[backrest]
backrest-checksum="[CHECKSUM]"
Major refactor of the test suite. * Make the code more modular and object-oriented. * Multiple Docker containers can now be created for a single test to simulate more realistic environments.
2016-06-24 14:12:58 +02:00
diff backup - updates since last full (db-master host)
Reduce log level for all expect tests to detail. The C code is designed to be efficient rather than deterministic at the debug log level. As we move more testing from integration to unit tests it makes less sense to try and maintain the expect logs at this log level. Most of the expect logs have already been moved to detail level but mock/all still had tests at debug level. Change the logging defaults in the config file and remove as many references to log-level-console as possible.
2019-05-23 00:23:44 +02:00
> [CONTAINER-EXEC] db-master [BACKREST-BIN] --config=[TEST_PATH]/db-master/pgbackrest.conf --no-online --process-max=1 --delta --type=diff --stanza=db backup
Log testing can now be enabled for certain deterministic tests. This works by comparing the generated logs against a previous copy. Currently only enabled for the backup/synthetic tests.
2015-05-05 19:08:48 +02:00
------------------------------------------------------------------------------------------------------------------------------------
Add infrastructure for multiple compression type support. Add compress-type option and deprecate compress option. Since the compress option is boolean it won't work with multiple compression types. Add logic to cfgLoadUpdateOption() to update compress-type if it is not set directly. The compress option should no longer be referenced outside the cfgLoadUpdateOption() function. Add common/compress/helper module to contain interface functions that work with multiple compression types. Code outside this module should no longer call specific compression drivers, though it may be OK to reference a specific compression type using the new interface (e.g., saving backup history files in gz format). Unit tests only test compression using the gz format because other formats may not be available in all builds. It is the job of integration tests to exercise all compression types. Additional compression types will be added in future commits.
2020-03-06 21:41:03 +02:00
P00 INFO: backup command begin [BACKREST-VERSION]: --buffer-size=[BUFFER-SIZE] --compress-level=3 --compress-type=none --config=[TEST_PATH]/db-master/pgbackrest.conf --db-timeout=45 --delta --lock-path=[TEST_PATH]/db-master/lock --log-level-console=detail --log-level-file=[LOG-LEVEL-FILE] --log-level-stderr=off --log-path=[TEST_PATH]/db-master/log[] --no-log-timestamp --no-online --pg1-path=[TEST_PATH]/db-master/db/base-2 --process-max=1 --protocol-timeout=60 --repo1-path=[TEST_PATH]/db-master/repo --stanza=db --start-fast --type=diff
Add time-based retention for full backups. The --repo-retention-full-type option allows retention of full backups based on a time period, specified in days. The new option will default to 'count' and therefore will not affect current installations. Setting repo-retention-full-type to 'time' will allow the user to use a time period, in days, to indicate full backup retention. Using this method, a full backup can be expired only if the time the backup completed is older than the number of days set with repo-retention-full (calculated from the moment the 'expire' command is run) and at least one full backup meets the retention period. If archive retention has not been configured, then the default settings will expire archives that are prior to the oldest retained full backup. For example, if there are three full backups ending in times that are 25 days old (F1), 20 days old (F2) and 10 days old (F3), then if the full retention period is 15 days, then only F1 will be expired; F2 will be retained because F1 is not at least 15 days old.
2020-05-08 21:25:03 +02:00
P00 WARN: option 'repo1-retention-full' is not set for 'repo1-retention-full-type=count', the repository may run out of space
Rename retention-* options to repo-retention-*.
2018-02-19 22:07:24 +02:00
HINT: to retain full backups indefinitely (without warning), set option 'repo1-retention-full' to the maximum.
Show Process ID in log output instead of filtering it out with the timestamp.
2016-09-14 23:30:57 +02:00
P00 INFO: last backup label = [BACKUP-FULL-2], version = [VERSION-1]
Added the --checksum-page option. This option allows pgBackRest to validate page checksums in data files when checksums are enabled on PostgreSQL >= 9.3. Note that this functionality requires a C library which may not initially be available in OS packages. The option will automatically be enabled when the library is present and checksums are enabled on the cluster.
2016-12-13 01:54:07 +02:00
P00 WARN: diff backup cannot alter 'checksum-page' option to 'false', reset to 'true' from [BACKUP-FULL-2]
Add checksum delta for incremental backups. Use checksums rather than timestamps to determine if files have changed. This is useful in cases where the timestamps may not be trustworthy, e.g. when performing an incremental after failing over to a standby. If checksum delta is enabled then checksums will be used for verification of resumed backups, even if they are full. Resumes have always used checksums to verify the files in the repository, enabling delta performs checksums on the database files as well. Note that the user must manually enable this feature in cases were it would be useful or just keep in enabled all the time. A future commit will address automatically enabling the feature in cases where it seems likely to be useful. Contributed by Cynthia Shang.
2018-09-19 17:12:45 +02:00
P01 DETAIL: match file from prior backup [TEST_PATH]/db-master/db/base-2/base/32768/33001 (64KB, 36%) checksum 6bf316f11d28c28914ea9be92c00de9bea6d9a6b
Fix incorrect offline upper bound for ignoring page checksum errors. For offline backups the upper bound was being set to 0x0000FFFF0000FFFF rather than UINT64_MAX. This meant that page checksum errors might be ignored for databases with a lot of past WAL in offline mode. Online mode is not affected since the upper bound is retrieved from pg_start_backup().
2019-07-11 15:13:56 +02:00
P01 DETAIL: match file from prior backup [TEST_PATH]/db-master/db/base-2/base/32768/44000_init (32KB, 54%) checksum 7a16d165e4775f7c92e8cdf60c0af57313f0bf90
P01 DETAIL: match file from prior backup [TEST_PATH]/db-master/db/base-2/base/32768/33000.32767 (32KB, 72%) checksum 6e99b589e550e68e934fd235ccba59fe5b592a9e
P01 DETAIL: match file from prior backup [TEST_PATH]/db-master/db/base-2/base/32768/33000 (32KB, 90%) checksum 7a16d165e4775f7c92e8cdf60c0af57313f0bf90
Support configurable WAL segment size. PostgreSQL 11 introduces configurable WAL segment sizes, from 1MB to 1GB. There are two areas that needed to be updated to support this: building the archive-get queue and checking that WAL has been archived after a backup. Both operations require the WAL segment size to properly build a list. Checking the archive after a backup is still implemented in Perl and has an active database connection, so just get the WAL segment size from the database. The archive-get command does not have a connection to the database, so get the WAL segment size from pg_control instead. This requires a deeper inspection of pg_control than has been done in the past, so it seemed best to copy the relevant data structures from each version of PostgreSQL and build a generic interface layer to address them. While this approach is a bit verbose, it has the advantage of being relatively simple, and can easily be updated for new versions of PostgreSQL. Since the integration tests generate pg_control files for testing, teach Perl how to generate files with the correct offsets for both 32-bit and 64-bit architectures.
2018-09-25 11:24:42 +02:00
P01 DETAIL: match file from prior backup [TEST_PATH]/db-master/db/base-2/global/pg_control (8KB, 95%) checksum 4c77c900f7af0d9ab13fa9982051a42e0b637f6c
Add checksum delta for incremental backups. Use checksums rather than timestamps to determine if files have changed. This is useful in cases where the timestamps may not be trustworthy, e.g. when performing an incremental after failing over to a standby. If checksum delta is enabled then checksums will be used for verification of resumed backups, even if they are full. Resumes have always used checksums to verify the files in the repository, enabling delta performs checksums on the database files as well. Note that the user must manually enable this feature in cases were it would be useful or just keep in enabled all the time. A future commit will address automatically enabling the feature in cases where it seems likely to be useful. Contributed by Cynthia Shang.
2018-09-19 17:12:45 +02:00
P01 DETAIL: match file from prior backup [TEST_PATH]/db-master/db/base-2/base/1/12000 (8KB, 99%) checksum 22c98d248ff548311eda88559e4a8405ed77c003
P01 DETAIL: match file from prior backup [TEST_PATH]/db-master/db/base-2/postgresql.conf (21B, 99%) checksum 6721d92c9fcdf4248acff1f9a1377127d9064807
P01 INFO: backup file [TEST_PATH]/db-master/db/base-2/badchecksum.txt (11B, 99%) checksum f927212cd08d11a42a666b2f04235398e9ceeb51
P01 INFO: backup file [TEST_PATH]/db-master/db/base-2/base/16384/17000 (8B, 99%) checksum 9a53d532e27785e681766c98516a5e93f096a501
Added the --checksum-page option. This option allows pgBackRest to validate page checksums in data files when checksums are enabled on PostgreSQL >= 9.3. Note that this functionality requires a C library which may not initially be available in OS packages. The option will automatically be enabled when the library is present and checksums are enabled on the cluster.
2016-12-13 01:54:07 +02:00
P00 WARN: page misalignment in file [TEST_PATH]/db-master/db/base-2/base/16384/17000: file size 8 is not divisible by page size 8192
Add checksum delta for incremental backups. Use checksums rather than timestamps to determine if files have changed. This is useful in cases where the timestamps may not be trustworthy, e.g. when performing an incremental after failing over to a standby. If checksum delta is enabled then checksums will be used for verification of resumed backups, even if they are full. Resumes have always used checksums to verify the files in the repository, enabling delta performs checksums on the database files as well. Note that the user must manually enable this feature in cases were it would be useful or just keep in enabled all the time. A future commit will address automatically enabling the feature in cases where it seems likely to be useful. Contributed by Cynthia Shang.
2018-09-19 17:12:45 +02:00
P01 INFO: backup file [TEST_PATH]/db-master/db/base-2/changecontent.txt (7B, 99%) checksum a094d94583e209556d03c3c5da33131a065f1689
P01 DETAIL: match file from prior backup [TEST_PATH]/db-master/db/base-2/pg_stat/global.stat (5B, 99%) checksum e350d5ce0153f3e22d5db21cf2a4eff00f3ee877
P01 INFO: backup file [TEST_PATH]/db-master/db/base-2/base/base2.txt (5B, 99%) checksum 09b5e31766be1dba1ec27de82f975c1b6eea2a92
P01 DETAIL: match file from prior backup [TEST_PATH]/db-master/db/base-2/changetime.txt (4B, 99%) checksum 88087292ed82e26f3eb824d0bffc05ccf7a30f8d
P01 DETAIL: match file from prior backup [TEST_PATH]/db-master/db/base-2/base/32768/PG_VERSION (3B, 99%) checksum 184473f470864e067ee3a22e64b47b0a1c356f29
P01 DETAIL: match file from prior backup [TEST_PATH]/db-master/db/base-2/base/16384/PG_VERSION (3B, 99%) checksum 184473f470864e067ee3a22e64b47b0a1c356f29
P01 DETAIL: match file from prior backup [TEST_PATH]/db-master/db/base-2/base/1/PG_VERSION (3B, 99%) checksum 184473f470864e067ee3a22e64b47b0a1c356f29
P01 DETAIL: match file from prior backup [TEST_PATH]/db-master/db/base-2/PG_VERSION (3B, 99%) checksum 184473f470864e067ee3a22e64b47b0a1c356f29
P01 INFO: backup file [TEST_PATH]/db-master/db/base-2/zerosize.txt (0B, 99%)
P01 INFO: backup file [TEST_PATH]/db-master/db/base-2/pg_tblspc/2/[TS_PATH-1]/32768/tablespace2b.txt (8B, 99%) checksum e324463005236d83e6e54795dbddd20a74533bf3
Added the --checksum-page option. This option allows pgBackRest to validate page checksums in data files when checksums are enabled on PostgreSQL >= 9.3. Note that this functionality requires a C library which may not initially be available in OS packages. The option will automatically be enabled when the library is present and checksums are enabled on the cluster.
2016-12-13 01:54:07 +02:00
P00 WARN: page misalignment in file [TEST_PATH]/db-master/db/base-2/pg_tblspc/2/[TS_PATH-1]/32768/tablespace2b.txt: file size 8 is not divisible by page size 8192
Additional backup exclusions. * Exclude contents of pg_snapshots, pg_serial, pg_notify, and pg_dynshmem from backup since they are rebuilt on startup. * Exclude pg_internal.init files from backup since they are rebuilt on startup.
2017-09-04 14:26:57 +02:00
P01 INFO: backup file [TEST_PATH]/db-master/db/base-2/pg_tblspc/2/[TS_PATH-1]/32768/tablespace2.txt (7B, 100%) checksum dc7f76e43c46101b47acc55ae4d593a9e6983578
Added the --checksum-page option. This option allows pgBackRest to validate page checksums in data files when checksums are enabled on PostgreSQL >= 9.3. Note that this functionality requires a C library which may not initially be available in OS packages. The option will automatically be enabled when the library is present and checksums are enabled on the cluster.
2016-12-13 01:54:07 +02:00
P00 WARN: page misalignment in file [TEST_PATH]/db-master/db/base-2/pg_tblspc/2/[TS_PATH-1]/32768/tablespace2.txt: file size 7 is not divisible by page size 8192
The backup command is implemented entirely in C. For the most part this is a direct migration of the Perl code into C except as noted below. A backup can now be initiated from a linked directory. The link will not be stored in the manifest or recreated on restore. If a link or directory does not already exist in the restore location then a directory will be created. The logic for creating backup labels has been improved and it should no longer be possible to get a backup label earlier than the latest backup even with timezone changes or clock skew. This has never been an issue in the field that we know of, but we found it in testing. For online backups all times are fetched from the PostgreSQL primary host (before only copy start was). This doesn't affect backup integrity but it does prevent clock skew between hosts affecting backup duration reporting. Archive copy now works as expected when the archive and backup have different compression settings, i.e. when one is compressed and the other is not. This was a long-standing bug in the Perl code. Resume will now work even if hardlink settings have been changed. Reviewed by Cynthia Shang.
2019-12-14 00:14:26 +02:00
P00 DETAIL: reference pg_data/PG_VERSION to [BACKUP-FULL-2]
P00 DETAIL: reference pg_data/base/1/12000 to [BACKUP-FULL-2]
P00 DETAIL: reference pg_data/base/1/PG_VERSION to [BACKUP-FULL-2]
P00 DETAIL: reference pg_data/base/16384/PG_VERSION to [BACKUP-FULL-2]
P00 DETAIL: reference pg_data/base/32768/33000 to [BACKUP-FULL-2]
P00 DETAIL: reference pg_data/base/32768/33000.32767 to [BACKUP-FULL-2]
P00 DETAIL: reference pg_data/base/32768/33001 to [BACKUP-FULL-2]
P00 DETAIL: reference pg_data/base/32768/44000_init to [BACKUP-FULL-2]
P00 DETAIL: reference pg_data/base/32768/PG_VERSION to [BACKUP-FULL-2]
P00 DETAIL: reference pg_data/changetime.txt to [BACKUP-FULL-2]
P00 DETAIL: reference pg_data/global/pg_control to [BACKUP-FULL-2]
P00 DETAIL: reference pg_data/pg_stat/global.stat to [BACKUP-FULL-2]
P00 DETAIL: reference pg_data/postgresql.conf to [BACKUP-FULL-2]
P00 DETAIL: reference pg_data/special-!_.*'()&!@;:+,? to [BACKUP-FULL-2]
P00 DETAIL: reference pg_data/zero_from_start to [BACKUP-FULL-2]
Add checksum delta for incremental backups. Use checksums rather than timestamps to determine if files have changed. This is useful in cases where the timestamps may not be trustworthy, e.g. when performing an incremental after failing over to a standby. If checksum delta is enabled then checksums will be used for verification of resumed backups, even if they are full. Resumes have always used checksums to verify the files in the repository, enabling delta performs checksums on the database files as well. Note that the user must manually enable this feature in cases were it would be useful or just keep in enabled all the time. A future commit will address automatically enabling the feature in cases where it seems likely to be useful. Contributed by Cynthia Shang.
2018-09-19 17:12:45 +02:00
P00 INFO: diff backup size = 176KB
Remove integration tests that are now covered in the unit tests. Most of these tests are just checking that errors are thrown when required. These are well covered in various unit tests. The "cannot resume" tests are also well covered in the backup unit tests. Finally, config warnings are well covered in the config unit tests. There is more to be done here, but this accounts for the low-hanging fruit.
2019-12-18 03:14:45 +02:00
P00 INFO: new backup label = [BACKUP-DIFF-2]
Added a predictable log entry to signal that a command has completed successfully. For example a backup ends successfully with: INFO: backup command end: completed successfully.
2016-12-04 21:44:53 +02:00
P00 INFO: backup command end: completed successfully
Integration test improvements for disk and memory efficiency. Set log-level-file=off when more that one test will run. In this case is it impossible to see the logs anyway since they will be automatically cleaned up after the test. This improves performance pretty dramatically since trace-level logging is expensive. If a singe integration test is run then log-level-file is trace by default but can be changed with the --log-level-test-file option. Reduce buffer-size to 64k to save memory during testing and allow more processes to run in parallel. Update log replacement rules so that these options can change without affecting expect logs.
2019-12-17 22:23:07 +02:00
P00 INFO: expire command begin [BACKREST-VERSION]: --buffer-size=[BUFFER-SIZE] --config=[TEST_PATH]/db-master/pgbackrest.conf --lock-path=[TEST_PATH]/db-master/lock --log-level-console=detail --log-level-file=[LOG-LEVEL-FILE] --log-level-stderr=off --log-path=[TEST_PATH]/db-master/log[] --no-log-timestamp --repo1-path=[TEST_PATH]/db-master/repo --stanza=db
Rename retention-* options to repo-retention-*.
2018-02-19 22:07:24 +02:00
P00 INFO: option 'repo1-retention-archive' is not set - archive logs will not be expired
Added a predictable log entry to signal that a command has completed successfully. For example a backup ends successfully with: INFO: backup command end: completed successfully.
2016-12-04 21:44:53 +02:00
P00 INFO: expire command end: completed successfully
Log testing can now be enabled for certain deterministic tests. This works by comparing the generated logs against a previous copy. Currently only enabled for the backup/synthetic tests.
2015-05-05 19:08:48 +02:00
Rename test paths for clarity. This was worked out as part of the test suite refactor [c8f806a] but not committed with it because of the large number of expect logs changes involved. Keeping them separate made it easier to audit the changes in the refactor.
2016-06-24 16:54:31 +02:00
+ supplemental file: [TEST_PATH]/db-master/pgbackrest.conf
----------------------------------------------------------
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
[db]
Rename db-* options to pg-* and backup-* options to repo-* to improve consistency. * repo-* options are now indexed although only one is allowed. * List deprecated option names in documentation and command-line help.
2018-02-04 01:27:38 +02:00
pg1-path=[TEST_PATH]/db-master/db/base-2
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
New simpler configuration and consistent project/exe/path naming. * The repo-path option now always refers to the repository where backups and archive are stored, whether local or remote, so the repo-remote-path option has been removed. The new spool-path option can be used to define a location for queueing WAL segments when archiving asynchronously. Otherwise, a local repository is no longer required. * Implemented a new config format which should be far simpler to use. See the User Guide and Configuration Reference for details but for a simple configuration all options can now be placed in the stanza section. Options that are shared between stanzas can be placed in the [global] section. More complex configurations can still make use of command sections though this should be a rare use case. * The default configuration filename is now pgbackrest.conf instead of pg_backrest.conf. This was done for consistency with other naming changes but also to prevent old config files from being loaded accidentally. * The default repository name was changed from /var/lib/backup to /var/lib/pgbackrest. * Lock files are now stored in /tmp/pgbackrest by default. These days /run/pgbackrest would be the preferred location but that would require init scripts which are not part of this release. The lock-path option can be used to configure the lock directory. * Log files are now stored in /var/log/pgbackrest by default and no longer have the date appended so they can be managed with logrotate. The log-path option can be used to configure the lock directory. * Executable filename changed from pg_backrest to pgbackrest.
2016-04-14 15:30:54 +02:00
[db:restore]
Rename test paths for clarity. This was worked out as part of the test suite refactor [c8f806a] but not committed with it because of the large number of expect logs changes involved. Keeping them separate made it easier to audit the changes in the refactor.
2016-06-24 16:54:31 +02:00
tablespace-map=1=[TEST_PATH]/db-master/db/tablespace/ts1-2
tablespace-map=2=[TEST_PATH]/db-master/db/tablespace/ts2-2
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
New simpler configuration and consistent project/exe/path naming. * The repo-path option now always refers to the repository where backups and archive are stored, whether local or remote, so the repo-remote-path option has been removed. The new spool-path option can be used to define a location for queueing WAL segments when archiving asynchronously. Otherwise, a local repository is no longer required. * Implemented a new config format which should be far simpler to use. See the User Guide and Configuration Reference for details but for a simple configuration all options can now be placed in the stanza section. Options that are shared between stanzas can be placed in the [global] section. More complex configurations can still make use of command sections though this should be a rare use case. * The default configuration filename is now pgbackrest.conf instead of pg_backrest.conf. This was done for consistency with other naming changes but also to prevent old config files from being loaded accidentally. * The default repository name was changed from /var/lib/backup to /var/lib/pgbackrest. * Lock files are now stored in /tmp/pgbackrest by default. These days /run/pgbackrest would be the preferred location but that would require init scripts which are not part of this release. The lock-path option can be used to configure the lock directory. * Log files are now stored in /var/log/pgbackrest by default and no longer have the date appended so they can be managed with logrotate. The log-path option can be used to configure the lock directory. * Executable filename changed from pg_backrest to pgbackrest.
2016-04-14 15:30:54 +02:00
[global]
Integration test improvements for disk and memory efficiency. Set log-level-file=off when more that one test will run. In this case is it impossible to see the logs anyway since they will be automatically cleaned up after the test. This improves performance pretty dramatically since trace-level logging is expensive. If a singe integration test is run then log-level-file is trace by default but can be changed with the --log-level-test-file option. Reduce buffer-size to 64k to save memory during testing and allow more processes to run in parallel. Update log replacement rules so that these options can change without affecting expect logs.
2019-12-17 22:23:07 +02:00
buffer-size=[BUFFER-SIZE]
Fix custom settings for compress-level option being ignored. Reported by Jens Wilke.
2017-10-23 00:05:46 +02:00
compress-level=3
Add infrastructure for multiple compression type support. Add compress-type option and deprecate compress option. Since the compress option is boolean it won't work with multiple compression types. Add logic to cfgLoadUpdateOption() to update compress-type if it is not set directly. The compress option should no longer be referenced outside the cfgLoadUpdateOption() function. Add common/compress/helper module to contain interface functions that work with multiple compression types. Code outside this module should no longer call specific compression drivers, though it may be OK to reference a specific compression type using the new interface (e.g., saving backup history files in gz format). Unit tests only test compression using the gz format because other formats may not be available in all builds. It is the job of integration tests to exercise all compression types. Additional compression types will be added in future commits.
2020-03-06 21:41:03 +02:00
compress-type=none
Storage and IO layer refactor: Refactor storage layer to allow for new repository filesystems using drivers. (Reviewed by Cynthia Shang.) Refactor IO layer to allow for new compression formats, checksum types, and other capabilities using filters. (Reviewed by Cynthia Shang.)
2017-06-09 23:51:41 +02:00
db-timeout=45
lock-path=[TEST_PATH]/db-master/lock
Reduce log level for all expect tests to detail. The C code is designed to be efficient rather than deterministic at the debug log level. As we move more testing from integration to unit tests it makes less sense to try and maintain the expect logs at this log level. Most of the expect logs have already been moved to detail level but mock/all still had tests at debug level. Change the logging defaults in the config file and remove as many references to log-level-console as possible.
2019-05-23 00:23:44 +02:00
log-level-console=detail
Integration test improvements for disk and memory efficiency. Set log-level-file=off when more that one test will run. In this case is it impossible to see the logs anyway since they will be automatically cleaned up after the test. This improves performance pretty dramatically since trace-level logging is expensive. If a singe integration test is run then log-level-file is trace by default but can be changed with the --log-level-test-file option. Reduce buffer-size to 64k to save memory during testing and allow more processes to run in parallel. Update log replacement rules so that these options can change without affecting expect logs.
2019-12-17 22:23:07 +02:00
log-level-file=[LOG-LEVEL-FILE]
Added the log-level-stderr option. Controls whether console log messages are sent to stderr or stdout. By default this is set to warn which represents a change in behavior from previous versions, even though it may be more intuitive. Setting log-level-stderr=off will preserve the old behavior. Suggested by Sascha Biberhofer.
2016-10-05 15:09:30 +02:00
log-level-stderr=off
Storage and IO layer refactor: Refactor storage layer to allow for new repository filesystems using drivers. (Reviewed by Cynthia Shang.) Refactor IO layer to allow for new compression formats, checksum types, and other capabilities using filters. (Reviewed by Cynthia Shang.)
2017-06-09 23:51:41 +02:00
log-path=[TEST_PATH]/db-master/log
Integration test improvements for disk and memory efficiency. Set log-level-file=off when more that one test will run. In this case is it impossible to see the logs anyway since they will be automatically cleaned up after the test. This improves performance pretty dramatically since trace-level logging is expensive. If a singe integration test is run then log-level-file is trace by default but can be changed with the --log-level-test-file option. Reduce buffer-size to 64k to save memory during testing and allow more processes to run in parallel. Update log replacement rules so that these options can change without affecting expect logs.
2019-12-17 22:23:07 +02:00
log-subprocess=[LOG-SUBPROCESS]
Set log-timestamp=n for integration tests. This means less filtering of logs needs to be done and new timestamps can be added without adding new filters.
2018-05-11 17:24:38 +02:00
log-timestamp=n
Storage and IO layer refactor: Refactor storage layer to allow for new repository filesystems using drivers. (Reviewed by Cynthia Shang.) Refactor IO layer to allow for new compression formats, checksum types, and other capabilities using filters. (Reviewed by Cynthia Shang.)
2017-06-09 23:51:41 +02:00
protocol-timeout=60
Rename db-* options to pg-* and backup-* options to repo-* to improve consistency. * repo-* options are now indexed although only one is allowed. * List deprecated option names in documentation and command-line help.
2018-02-04 01:27:38 +02:00
repo1-path=[TEST_PATH]/db-master/repo
Consolidate archive-push tests into new archive test.
2017-06-21 14:38:53 +02:00
spool-path=[TEST_PATH]/db-master/spool
New simpler configuration and consistent project/exe/path naming. * The repo-path option now always refers to the repository where backups and archive are stored, whether local or remote, so the repo-remote-path option has been removed. The new spool-path option can be used to define a location for queueing WAL segments when archiving asynchronously. Otherwise, a local repository is no longer required. * Implemented a new config format which should be far simpler to use. See the User Guide and Configuration Reference for details but for a simple configuration all options can now be placed in the stanza section. Options that are shared between stanzas can be placed in the [global] section. More complex configurations can still make use of command sections though this should be a rare use case. * The default configuration filename is now pgbackrest.conf instead of pg_backrest.conf. This was done for consistency with other naming changes but also to prevent old config files from being loaded accidentally. * The default repository name was changed from /var/lib/backup to /var/lib/pgbackrest. * Lock files are now stored in /tmp/pgbackrest by default. These days /run/pgbackrest would be the preferred location but that would require init scripts which are not part of this release. The lock-path option can be used to configure the lock directory. * Log files are now stored in /var/log/pgbackrest by default and no longer have the date appended so they can be managed with logrotate. The log-path option can be used to configure the lock directory. * Executable filename changed from pg_backrest to pgbackrest.
2016-04-14 15:30:54 +02:00
[global:backup]
archive-copy=y
start-fast=y
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
Remove integration tests that are now covered in the unit tests. Most of these tests are just checking that errors are thrown when required. These are well covered in various unit tests. The "cannot resume" tests are also well covered in the backup unit tests. Finally, config warnings are well covered in the config unit tests. There is more to be done here, but this accounts for the low-hanging fruit.
2019-12-18 03:14:45 +02:00
+ supplemental file: [TEST_PATH]/db-master/repo/backup/db/[BACKUP-DIFF-2]/backup.manifest
Rename test paths for clarity. This was worked out as part of the test suite refactor [c8f806a] but not committed with it because of the large number of expect logs changes involved. Keeping them separate made it easier to audit the changes in the refactor.
2016-06-24 16:54:31 +02:00
-----------------------------------------------------------------------------------------
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
[backrest]
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
backrest-format=5
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
backrest-version="[VERSION-1]"
[backup]
Remove integration tests that are now covered in the unit tests. Most of these tests are just checking that errors are thrown when required. These are well covered in various unit tests. The "cannot resume" tests are also well covered in the backup unit tests. Finally, config warnings are well covered in the config unit tests. There is more to be done here, but this accounts for the low-hanging fruit.
2019-12-18 03:14:45 +02:00
backup-label="[BACKUP-DIFF-2]"
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
backup-prior="[BACKUP-FULL-2]"
backup-timestamp-copy-start=[TIMESTAMP]
backup-timestamp-start=[TIMESTAMP]
backup-timestamp-stop=[TIMESTAMP]
backup-type="diff"
[backup:db]
Database version refactoring. * Refactor db version constants into a separate module. * Update synthetic backup tests to PostgreSQL 9.4.
2016-08-12 04:35:24 +02:00
db-catalog-version=201409291
db-control-version=942
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
db-id=1
Automate generation of WAL and pg_control test files. The existing static files would not work with 32-bit or big-endian systems so create functions to generate these files dynamically rather than creating a bunch of new static files.
2017-11-19 03:02:54 +02:00
db-system-id=1000000000000000094
Database version refactoring. * Refactor db version constants into a separate module. * Update synthetic backup tests to PostgreSQL 9.4.
2016-08-12 04:35:24 +02:00
db-version="9.4"
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
[backup:option]
Set archive-check option in manifest correctly when offline. Archive check does not run when in offline backup mode but the option was set to true in the manifest. It's harmless since these options are informational only but it could cause confusion when debugging.
2019-11-28 15:27:21 +02:00
option-archive-check=false
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
option-archive-copy=true
Backup from a standby cluster. A connection to the primary cluster is still required to start/stop the backup and copy files that are not replicated, but the vast majority of files are copied from the standby in order to reduce load on the master.
2016-08-25 17:25:46 +02:00
option-backup-standby=false
Integration test improvements for disk and memory efficiency. Set log-level-file=off when more that one test will run. In this case is it impossible to see the logs anyway since they will be automatically cleaned up after the test. This improves performance pretty dramatically since trace-level logging is expensive. If a singe integration test is run then log-level-file is trace by default but can be changed with the --log-level-test-file option. Reduce buffer-size to 64k to save memory during testing and allow more processes to run in parallel. Update log replacement rules so that these options can change without affecting expect logs.
2019-12-17 22:23:07 +02:00
option-buffer-size=[BUFFER-SIZE]
Added the --checksum-page option. This option allows pgBackRest to validate page checksums in data files when checksums are enabled on PostgreSQL >= 9.3. Note that this functionality requires a C library which may not initially be available in OS packages. The option will automatically be enabled when the library is present and checksums are enabled on the cluster.
2016-12-13 01:54:07 +02:00
option-checksum-page=true
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
option-compress=false
Add additional options to backup.manifest for debugging purposes. Add the buffer-size, compress-level, compress-level-network, and process-max options to the backup:option section in backup.manifest to aid in debugging. It may also make sense to propagate these options up to backup.info so they can be displayed in the info command, but for now this is deemed sufficient. Contributed by blogh.
2019-03-10 11:03:52 +02:00
option-compress-level=3
option-compress-level-network=3
Add infrastructure for multiple compression type support. Add compress-type option and deprecate compress option. Since the compress option is boolean it won't work with multiple compression types. Add logic to cfgLoadUpdateOption() to update compress-type if it is not set directly. The compress option should no longer be referenced outside the cfgLoadUpdateOption() function. Add common/compress/helper module to contain interface functions that work with multiple compression types. Code outside this module should no longer call specific compression drivers, though it may be OK to reference a specific compression type using the new interface (e.g., saving backup history files in gz format). Unit tests only test compression using the gz format because other formats may not be available in all builds. It is the job of integration tests to exercise all compression types. Additional compression types will be added in future commits.
2020-03-06 21:41:03 +02:00
option-compress-type="none"
Add checksum delta for incremental backups. Use checksums rather than timestamps to determine if files have changed. This is useful in cases where the timestamps may not be trustworthy, e.g. when performing an incremental after failing over to a standby. If checksum delta is enabled then checksums will be used for verification of resumed backups, even if they are full. Resumes have always used checksums to verify the files in the repository, enabling delta performs checksums on the database files as well. Note that the user must manually enable this feature in cases were it would be useful or just keep in enabled all the time. A future commit will address automatically enabling the feature in cases where it seems likely to be useful. Contributed by Cynthia Shang.
2018-09-19 17:12:45 +02:00
option-delta=true
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
option-hardlink=false
Closed #167: Rename `--no-start-stop` option to `--no-online`.
2016-02-20 02:32:02 +02:00
option-online=false
Add additional options to backup.manifest for debugging purposes. Add the buffer-size, compress-level, compress-level-network, and process-max options to the backup:option section in backup.manifest to aid in debugging. It may also make sense to propagate these options up to backup.info so they can be displayed in the info command, but for now this is deemed sufficient. Contributed by blogh.
2019-03-10 11:03:52 +02:00
option-process-max=1
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
[backup:target]
Rename test paths for clarity. This was worked out as part of the test suite refactor [c8f806a] but not committed with it because of the large number of expect logs changes involved. Keeping them separate made it easier to audit the changes in the refactor.
2016-06-24 16:54:31 +02:00
pg_data={"path":"[TEST_PATH]/db-master/db/base-2","type":"path"}
pg_tblspc/2={"path":"[TEST_PATH]/db-master/db/tablespace/ts2-2","tablespace-id":"2","tablespace-name":"ts2","type":"link"}
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
[target:file]
Fixed a issue where tablespaces were copied from the master during standby backup.
2016-09-04 15:19:44 +02:00
pg_data/PG_VERSION={"checksum":"184473f470864e067ee3a22e64b47b0a1c356f29","master":true,"reference":"[BACKUP-FULL-2]","size":[SIZE],"timestamp":[TIMESTAMP-1]}
pg_data/badchecksum.txt={"checksum":"f927212cd08d11a42a666b2f04235398e9ceeb51","master":true,"size":[SIZE],"timestamp":[TIMESTAMP-1]}
Added the --checksum-page option. This option allows pgBackRest to validate page checksums in data files when checksums are enabled on PostgreSQL >= 9.3. Note that this functionality requires a C library which may not initially be available in OS packages. The option will automatically be enabled when the library is present and checksums are enabled on the cluster.
2016-12-13 01:54:07 +02:00
pg_data/base/1/12000={"checksum":"22c98d248ff548311eda88559e4a8405ed77c003","checksum-page":true,"reference":"[BACKUP-FULL-2]","size":[SIZE],"timestamp":[TIMESTAMP-1]}
Database version refactoring. * Refactor db version constants into a separate module. * Update synthetic backup tests to PostgreSQL 9.4.
2016-08-12 04:35:24 +02:00
pg_data/base/1/PG_VERSION={"checksum":"184473f470864e067ee3a22e64b47b0a1c356f29","mode":"0660","reference":"[BACKUP-FULL-2]","size":[SIZE],"timestamp":[TIMESTAMP-1]}
Added the --checksum-page option. This option allows pgBackRest to validate page checksums in data files when checksums are enabled on PostgreSQL >= 9.3. Note that this functionality requires a C library which may not initially be available in OS packages. The option will automatically be enabled when the library is present and checksums are enabled on the cluster.
2016-12-13 01:54:07 +02:00
pg_data/base/16384/17000={"checksum":"9a53d532e27785e681766c98516a5e93f096a501","checksum-page":false,"size":[SIZE],"timestamp":[TIMESTAMP-1]}
Database version refactoring. * Refactor db version constants into a separate module. * Update synthetic backup tests to PostgreSQL 9.4.
2016-08-12 04:35:24 +02:00
pg_data/base/16384/PG_VERSION={"checksum":"184473f470864e067ee3a22e64b47b0a1c356f29","reference":"[BACKUP-FULL-2]","size":[SIZE],"timestamp":[TIMESTAMP-1]}
Fix incorrect offline upper bound for ignoring page checksum errors. For offline backups the upper bound was being set to 0x0000FFFF0000FFFF rather than UINT64_MAX. This meant that page checksum errors might be ignored for databases with a lot of past WAL in offline mode. Online mode is not affected since the upper bound is retrieved from pg_start_backup().
2019-07-11 15:13:56 +02:00
pg_data/base/32768/33000={"checksum":"7a16d165e4775f7c92e8cdf60c0af57313f0bf90","checksum-page":true,"reference":"[BACKUP-FULL-2]","size":[SIZE],"timestamp":[TIMESTAMP-1]}
pg_data/base/32768/33000.32767={"checksum":"6e99b589e550e68e934fd235ccba59fe5b592a9e","checksum-page":true,"reference":"[BACKUP-FULL-2]","size":[SIZE],"timestamp":[TIMESTAMP-1]}
Added the --checksum-page option. This option allows pgBackRest to validate page checksums in data files when checksums are enabled on PostgreSQL >= 9.3. Note that this functionality requires a C library which may not initially be available in OS packages. The option will automatically be enabled when the library is present and checksums are enabled on the cluster.
2016-12-13 01:54:07 +02:00
pg_data/base/32768/33001={"checksum":"6bf316f11d28c28914ea9be92c00de9bea6d9a6b","checksum-page":false,"checksum-page-error":[0,[3,5],7],"reference":"[BACKUP-FULL-2]","size":[SIZE],"timestamp":[TIMESTAMP-1]}
Fix incorrect offline upper bound for ignoring page checksum errors. For offline backups the upper bound was being set to 0x0000FFFF0000FFFF rather than UINT64_MAX. This meant that page checksum errors might be ignored for databases with a lot of past WAL in offline mode. Online mode is not affected since the upper bound is retrieved from pg_start_backup().
2019-07-11 15:13:56 +02:00
pg_data/base/32768/44000_init={"checksum":"7a16d165e4775f7c92e8cdf60c0af57313f0bf90","checksum-page":true,"reference":"[BACKUP-FULL-2]","size":[SIZE],"timestamp":[TIMESTAMP-1]}
Database version refactoring. * Refactor db version constants into a separate module. * Update synthetic backup tests to PostgreSQL 9.4.
2016-08-12 04:35:24 +02:00
pg_data/base/32768/PG_VERSION={"checksum":"184473f470864e067ee3a22e64b47b0a1c356f29","reference":"[BACKUP-FULL-2]","size":[SIZE],"timestamp":[TIMESTAMP-1]}
Update Perl page checksum expression. This expression determines which files contain page checksums but it was also including the directory above the relation directories. In a real PostgreSQL installation this not a problem because these directories don't contain any files. However, our tests place a file in `base` which the Perl code thought should have page checksums while the new C code says no. Update the expression to document the change and avoid churn in the expect logs later.
2019-11-25 14:37:09 +02:00
pg_data/base/base2.txt={"checksum":"09b5e31766be1dba1ec27de82f975c1b6eea2a92","size":[SIZE],"timestamp":[TIMESTAMP-1]}
Add checksum delta for incremental backups. Use checksums rather than timestamps to determine if files have changed. This is useful in cases where the timestamps may not be trustworthy, e.g. when performing an incremental after failing over to a standby. If checksum delta is enabled then checksums will be used for verification of resumed backups, even if they are full. Resumes have always used checksums to verify the files in the repository, enabling delta performs checksums on the database files as well. Note that the user must manually enable this feature in cases were it would be useful or just keep in enabled all the time. A future commit will address automatically enabling the feature in cases where it seems likely to be useful. Contributed by Cynthia Shang.
2018-09-19 17:12:45 +02:00
pg_data/changecontent.txt={"checksum":"a094d94583e209556d03c3c5da33131a065f1689","master":true,"size":[SIZE],"timestamp":[TIMESTAMP-1]}
pg_data/changetime.txt={"checksum":"88087292ed82e26f3eb824d0bffc05ccf7a30f8d","master":true,"reference":"[BACKUP-FULL-2]","size":[SIZE],"timestamp":[TIMESTAMP-2]}
Support configurable WAL segment size. PostgreSQL 11 introduces configurable WAL segment sizes, from 1MB to 1GB. There are two areas that needed to be updated to support this: building the archive-get queue and checking that WAL has been archived after a backup. Both operations require the WAL segment size to properly build a list. Checking the archive after a backup is still implemented in Perl and has an active database connection, so just get the WAL segment size from the database. The archive-get command does not have a connection to the database, so get the WAL segment size from pg_control instead. This requires a deeper inspection of pg_control than has been done in the past, so it seemed best to copy the relevant data structures from each version of PostgreSQL and build a generic interface layer to address them. While this approach is a bit verbose, it has the advantage of being relatively simple, and can easily be updated for new versions of PostgreSQL. Since the integration tests generate pg_control files for testing, teach Perl how to generate files with the correct offsets for both 32-bit and 64-bit architectures.
2018-09-25 11:24:42 +02:00
pg_data/global/pg_control={"checksum":"4c77c900f7af0d9ab13fa9982051a42e0b637f6c","master":true,"reference":"[BACKUP-FULL-2]","size":[SIZE],"timestamp":[TIMESTAMP-2]}
Fixed a issue where tablespaces were copied from the master during standby backup.
2016-09-04 15:19:44 +02:00
pg_data/pg_stat/global.stat={"checksum":"e350d5ce0153f3e22d5db21cf2a4eff00f3ee877","master":true,"reference":"[BACKUP-FULL-2]","size":[SIZE],"timestamp":[TIMESTAMP-2]}
pg_data/postgresql.conf={"checksum":"6721d92c9fcdf4248acff1f9a1377127d9064807","master":true,"reference":"[BACKUP-FULL-2]","size":[SIZE],"timestamp":[TIMESTAMP-2]}
Use minio for integration testing. ScalityS3 has not received any maintenance in years and is slow to start which is bad for testing. Replace it with minio which starts quickly and ships as a single executable or a tiny container. Minio has stricter limits on allowable characters but should still provide enough coverage to show that our encoding is working correctly. This commit also includes the upgrade to openssl 1.1.1 in the Ubuntu 18.04 container.
2019-07-03 04:20:35 +02:00
pg_data/special-!_.*'()&!@;:+,?={"master":true,"reference":"[BACKUP-FULL-2]","size":[SIZE],"timestamp":[TIMESTAMP-1]}
Allow zero-size files in backup manifest to reference a prior manifest regardless of timestamp delta. Contributed by Cynthia Shang.
2018-08-24 22:50:33 +02:00
pg_data/zero_from_start={"master":true,"reference":"[BACKUP-FULL-2]","size":[SIZE],"timestamp":[TIMESTAMP-1]}
Add zero-length file to mock/all test.
2018-07-05 21:40:50 +02:00
pg_data/zerosize.txt={"master":true,"size":[SIZE],"timestamp":[TIMESTAMP-2]}
Added the --checksum-page option. This option allows pgBackRest to validate page checksums in data files when checksums are enabled on PostgreSQL >= 9.3. Note that this functionality requires a C library which may not initially be available in OS packages. The option will automatically be enabled when the library is present and checksums are enabled on the cluster.
2016-12-13 01:54:07 +02:00
pg_tblspc/2/[TS_PATH-1]/32768/tablespace2.txt={"checksum":"dc7f76e43c46101b47acc55ae4d593a9e6983578","checksum-page":false,"size":[SIZE],"timestamp":[TIMESTAMP-1]}
pg_tblspc/2/[TS_PATH-1]/32768/tablespace2b.txt={"checksum":"e324463005236d83e6e54795dbddd20a74533bf3","checksum-page":false,"size":[SIZE],"timestamp":[TIMESTAMP-1]}
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
[target:file:default]
Continuous integration using travis-ci.
2017-02-21 15:59:23 +02:00
group="[GROUP-1]"
Fixed a issue where tablespaces were copied from the master during standby backup.
2016-09-04 15:19:44 +02:00
master=false
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
mode="0600"
Regression tests can now be run as any properly-configured user, not just vagrant.
2016-10-01 19:39:44 +02:00
user="[USER-1]"
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
[target:link]
Rename test paths for clarity. This was worked out as part of the test suite refactor [c8f806a] but not committed with it because of the large number of expect logs changes involved. Keeping them separate made it easier to audit the changes in the refactor.
2016-06-24 16:54:31 +02:00
pg_data/pg_tblspc/2={"destination":"[TEST_PATH]/db-master/db/tablespace/ts2-2"}
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
[target:link:default]
Continuous integration using travis-ci.
2017-02-21 15:59:23 +02:00
group="[GROUP-1]"
Regression tests can now be run as any properly-configured user, not just vagrant.
2016-10-01 19:39:44 +02:00
user="[USER-1]"
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
[target:path]
The restore command is implemented entirely in C. For the most part this is a direct migration of the Perl code into C. There is one important behavioral change with regard to how file permissions are handled. The Perl code tried to set ownership as it was in the manifest even when running as an unprivileged user. This usually just led to errors and frustration. The C code works like this: If a restore is run as a non-root user (the typical scenario) then all files restored will belong to the user/group executing pgBackRest. If existing files are not owned by the executing user/group then an error will result if the ownership cannot be updated to the executing user/group. In that case the file ownership will need to be updated by a privileged user before the restore can be retried. If a restore is run as the root user then pgBackRest will attempt to recreate the ownership recorded in the manifest when the backup was made. Only user/group names are stored in the manifest so the same names must exist on the restore host for this to work. If the user/group name cannot be found locally then the user/group of the PostgreSQL data directory will be used and finally root if the data directory user/group cannot be mapped to a name. Reviewed by Cynthia Shang.
2019-09-26 13:52:02 +02:00
pg_data={"mode":"0777"}
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
pg_data/base={}
Allow selective restore of databases from a cluster backup. This feature can result in major space and time savings when only specific databases are restored. Unrestored databases will not be accessible but must be manually dropped before they will be removed from the shared catalogue.
2016-05-11 15:21:39 +02:00
pg_data/base/1={}
pg_data/base/16384={}
pg_data/base/32768={}
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
pg_data/global={}
pg_data/pg_clog={}
Exclude directories during backup that are cleaned, recreated, or zeroed by PostgreSQL at startup. These include (depending on the version where they were introduced): pgsql_tmp, pg_dynshmem, pg_notify, pg_replslot, pg_serial, pg_snapshots, pg_stat_tmp, pg_subtrans. The postgresql.auto.conf.tmp file is now excluded in addition to files that were already excluded: backup_label.old, postmaster.opts, postmaster.pid, recovery.conf, recovery.done.
2016-08-16 15:35:16 +02:00
pg_data/pg_dynshmem={}
Allow arbitrary directories and/or files to be excluded from a backup. Misuse of this feature can lead to inconsistent backups so read the --exclude documentation carefully before using.
2018-08-27 21:51:05 +02:00
pg_data/pg_log={}
Exclude directories during backup that are cleaned, recreated, or zeroed by PostgreSQL at startup. These include (depending on the version where they were introduced): pgsql_tmp, pg_dynshmem, pg_notify, pg_replslot, pg_serial, pg_snapshots, pg_stat_tmp, pg_subtrans. The postgresql.auto.conf.tmp file is now excluded in addition to files that were already excluded: backup_label.old, postmaster.opts, postmaster.pid, recovery.conf, recovery.done.
2016-08-16 15:35:16 +02:00
pg_data/pg_notify={}
pg_data/pg_replslot={}
pg_data/pg_serial={}
pg_data/pg_snapshots={}
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
pg_data/pg_stat={}
Exclude directories during backup that are cleaned, recreated, or zeroed by PostgreSQL at startup. These include (depending on the version where they were introduced): pgsql_tmp, pg_dynshmem, pg_notify, pg_replslot, pg_serial, pg_snapshots, pg_stat_tmp, pg_subtrans. The postgresql.auto.conf.tmp file is now excluded in addition to files that were already excluded: backup_label.old, postmaster.opts, postmaster.pid, recovery.conf, recovery.done.
2016-08-16 15:35:16 +02:00
pg_data/pg_stat_tmp={}
pg_data/pg_subtrans={}
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
pg_data/pg_tblspc={}
The restore command is implemented entirely in C. For the most part this is a direct migration of the Perl code into C. There is one important behavioral change with regard to how file permissions are handled. The Perl code tried to set ownership as it was in the manifest even when running as an unprivileged user. This usually just led to errors and frustration. The C code works like this: If a restore is run as a non-root user (the typical scenario) then all files restored will belong to the user/group executing pgBackRest. If existing files are not owned by the executing user/group then an error will result if the ownership cannot be updated to the executing user/group. In that case the file ownership will need to be updated by a privileged user before the restore can be retried. If a restore is run as the root user then pgBackRest will attempt to recreate the ownership recorded in the manifest when the backup was made. Only user/group names are stored in the manifest so the same names must exist on the restore host for this to work. If the user/group name cannot be found locally then the user/group of the PostgreSQL data directory will be used and finally root if the data directory user/group cannot be mapped to a name. Reviewed by Cynthia Shang.
2019-09-26 13:52:02 +02:00
pg_tblspc={"mode":"0777"}
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
pg_tblspc/2={}
pg_tblspc/2/[TS_PATH-1]={}
Allow selective restore of databases from a cluster backup. This feature can result in major space and time savings when only specific databases are restored. Unrestored databases will not be accessible but must be manually dropped before they will be removed from the shared catalogue.
2016-05-11 15:21:39 +02:00
pg_tblspc/2/[TS_PATH-1]/32768={}
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
[target:path:default]
Continuous integration using travis-ci.
2017-02-21 15:59:23 +02:00
group="[GROUP-1]"
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
mode="0700"
Regression tests can now be run as any properly-configured user, not just vagrant.
2016-10-01 19:39:44 +02:00
user="[USER-1]"
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
Move info file checksum to the end of the file. Putting the checksum at the beginning of the file made it impossible to stream the file out when saving. The entire file had to be held in memory while it was checksummed so the checksum could be written at the beginning. Instead place the checksum at the end. This does not break the existing Perl or C code since the read is not order dependent. There are no plans to improve the Perl code to take advantage of this change, but it will make the C implementation more efficient. Reviewed by Cynthia Shang.
2019-08-22 01:45:48 +02:00
[backrest]
backrest-checksum="[CHECKSUM]"
Rename test paths for clarity. This was worked out as part of the test suite refactor [c8f806a] but not committed with it because of the large number of expect logs changes involved. Keeping them separate made it easier to audit the changes in the refactor.
2016-06-24 16:54:31 +02:00
+ supplemental file: [TEST_PATH]/db-master/repo/backup/db/backup.info
---------------------------------------------------------------------
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
[backrest]
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
backrest-format=5
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
backrest-version="[VERSION-1]"
[backup:current]
Set archive-check option in manifest correctly when offline. Archive check does not run when in offline backup mode but the option was set to true in the manifest. It's harmless since these options are informational only but it could cause confusion when debugging.
2019-11-28 15:27:21 +02:00
[BACKUP-FULL-2]={"backrest-format":5,"backrest-version":"[VERSION-1]","backup-archive-start":null,"backup-archive-stop":null,"backup-info-repo-size":[SIZE],"backup-info-repo-size-delta":[DELTA],"backup-info-size":[SIZE],"backup-info-size-delta":[DELTA],"backup-timestamp-start":[TIMESTAMP],"backup-timestamp-stop":[TIMESTAMP],"backup-type":"full","db-id":1,"option-archive-check":false,"option-archive-copy":true,"option-backup-standby":false,"option-checksum-page":true,"option-compress":false,"option-hardlink":false,"option-online":false}
Remove integration tests that are now covered in the unit tests. Most of these tests are just checking that errors are thrown when required. These are well covered in various unit tests. The "cannot resume" tests are also well covered in the backup unit tests. Finally, config warnings are well covered in the config unit tests. There is more to be done here, but this accounts for the low-hanging fruit.
2019-12-18 03:14:45 +02:00
[BACKUP-INCR-2]={"backrest-format":5,"backrest-version":"[VERSION-1]","backup-archive-start":null,"backup-archive-stop":null,"backup-info-repo-size":[SIZE],"backup-info-repo-size-delta":[DELTA],"backup-info-size":[SIZE],"backup-info-size-delta":[DELTA],"backup-prior":"[BACKUP-FULL-2]","backup-reference":["[BACKUP-FULL-2]"],"backup-timestamp-start":[TIMESTAMP],"backup-timestamp-stop":[TIMESTAMP],"backup-type":"incr","db-id":1,"option-archive-check":false,"option-archive-copy":true,"option-backup-standby":false,"option-checksum-page":true,"option-compress":false,"option-hardlink":false,"option-online":false}
[BACKUP-DIFF-1]={"backrest-format":5,"backrest-version":"[VERSION-1]","backup-archive-start":null,"backup-archive-stop":null,"backup-info-repo-size":[SIZE],"backup-info-repo-size-delta":[DELTA],"backup-info-size":[SIZE],"backup-info-size-delta":[DELTA],"backup-prior":"[BACKUP-FULL-2]","backup-reference":["[BACKUP-FULL-2]"],"backup-timestamp-start":[TIMESTAMP],"backup-timestamp-stop":[TIMESTAMP],"backup-type":"diff","db-id":1,"option-archive-check":false,"option-archive-copy":true,"option-backup-standby":false,"option-checksum-page":true,"option-compress":false,"option-hardlink":false,"option-online":false}
[BACKUP-INCR-3]={"backrest-format":5,"backrest-version":"[VERSION-1]","backup-archive-start":null,"backup-archive-stop":null,"backup-info-repo-size":[SIZE],"backup-info-repo-size-delta":[DELTA],"backup-info-size":[SIZE],"backup-info-size-delta":[DELTA],"backup-prior":"[BACKUP-DIFF-1]","backup-reference":["[BACKUP-FULL-2]","[BACKUP-DIFF-1]"],"backup-timestamp-start":[TIMESTAMP],"backup-timestamp-stop":[TIMESTAMP],"backup-type":"incr","db-id":1,"option-archive-check":false,"option-archive-copy":true,"option-backup-standby":false,"option-checksum-page":true,"option-compress":false,"option-hardlink":false,"option-online":false}
[BACKUP-INCR-4]={"backrest-format":5,"backrest-version":"[VERSION-1]","backup-archive-start":null,"backup-archive-stop":null,"backup-info-repo-size":[SIZE],"backup-info-repo-size-delta":[DELTA],"backup-info-size":[SIZE],"backup-info-size-delta":[DELTA],"backup-prior":"[BACKUP-INCR-3]","backup-reference":["[BACKUP-FULL-2]","[BACKUP-DIFF-1]","[BACKUP-INCR-3]"],"backup-timestamp-start":[TIMESTAMP],"backup-timestamp-stop":[TIMESTAMP],"backup-type":"incr","db-id":1,"option-archive-check":false,"option-archive-copy":true,"option-backup-standby":false,"option-checksum-page":true,"option-compress":false,"option-hardlink":false,"option-online":false}
Set archive-check option in manifest correctly when offline. Archive check does not run when in offline backup mode but the option was set to true in the manifest. It's harmless since these options are informational only but it could cause confusion when debugging.
2019-11-28 15:27:21 +02:00
[BACKUP-DIFF-2]={"backrest-format":5,"backrest-version":"[VERSION-1]","backup-archive-start":null,"backup-archive-stop":null,"backup-info-repo-size":[SIZE],"backup-info-repo-size-delta":[DELTA],"backup-info-size":[SIZE],"backup-info-size-delta":[DELTA],"backup-prior":"[BACKUP-FULL-2]","backup-reference":["[BACKUP-FULL-2]"],"backup-timestamp-start":[TIMESTAMP],"backup-timestamp-stop":[TIMESTAMP],"backup-type":"diff","db-id":1,"option-archive-check":false,"option-archive-copy":true,"option-backup-standby":false,"option-checksum-page":true,"option-compress":false,"option-hardlink":false,"option-online":false}
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
[db]
Database version refactoring. * Refactor db version constants into a separate module. * Update synthetic backup tests to PostgreSQL 9.4.
2016-08-12 04:35:24 +02:00
db-catalog-version=201409291
db-control-version=942
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
db-id=1
Automate generation of WAL and pg_control test files. The existing static files would not work with 32-bit or big-endian systems so create functions to generate these files dynamically rather than creating a bunch of new static files.
2017-11-19 03:02:54 +02:00
db-system-id=1000000000000000094
Database version refactoring. * Refactor db version constants into a separate module. * Update synthetic backup tests to PostgreSQL 9.4.
2016-08-12 04:35:24 +02:00
db-version="9.4"
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
[db:history]
Automate generation of WAL and pg_control test files. The existing static files would not work with 32-bit or big-endian systems so create functions to generate these files dynamically rather than creating a bunch of new static files.
2017-11-19 03:02:54 +02:00
1={"db-catalog-version":201409291,"db-control-version":942,"db-system-id":1000000000000000094,"db-version":"9.4"}
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
Improve performance of info file load/save. Info files required three copies in memory to be loaded (the original string, an ini representation, and the final info object). Not only was this memory inefficient but the Ini object does sequential scans when searching for keys making large files very slow to load. This has not been an issue since archive.info and backup.info are very small, but it becomes a big deal when loading manifests with hundreds of thousands of files. Instead of holding copies of the data in memory, use a callback to deliver the ini data directly to the object when loading. Use a similar method for save to avoid having an intermediate copy. Save is a bit complex because sections/keys must be written in alpha order or older versions of pgBackRest will not calculate the correct checksum. Also move the load retry logic to helper functions rather than embedding it in the Info object. This allows for more flexibility in loading and ensures that stack traces will be available when developing unit tests. Reviewed by Cynthia Shang.
2019-09-06 19:48:28 +02:00
[backrest]
backrest-checksum="[CHECKSUM]"
Remove integration tests that are now covered in the unit tests. Most of these tests are just checking that errors are thrown when required. These are well covered in various unit tests. The "cannot resume" tests are also well covered in the backup unit tests. Finally, config warnings are well covered in the config unit tests. There is more to be done here, but this accounts for the low-hanging fruit.
2019-12-18 03:14:45 +02:00
diff backup - remove files (db-master host)
Update integration tests in mock/all that use test points. Test points will not be available in the C code so update these tests as best as possible without using them. This represents a loss of coverage for the Perl code (soon to be removed) which will be made up in the C code with unit tests.
2019-11-25 14:48:52 +02:00
> [CONTAINER-EXEC] db-master [BACKREST-BIN] --config=[TEST_PATH]/db-master/pgbackrest.conf --no-online --process-max=1 --delta --type=diff --stanza=db backup
Log testing can now be enabled for certain deterministic tests. This works by comparing the generated logs against a previous copy. Currently only enabled for the backup/synthetic tests.
2015-05-05 19:08:48 +02:00
------------------------------------------------------------------------------------------------------------------------------------
Add Zstandard compression support. Zstandard is a fast lossless compression algorithm targeting real-time compression scenarios at zlib-level and better compression ratios. It's backed by a very fast entropy stage, provided by Huff0 and FSE library. Zstandard version >= 1.0 is required, which is generally only available on newer distributions.
2020-05-04 21:25:27 +02:00
P00 INFO: backup command begin [BACKREST-VERSION]: --buffer-size=[BUFFER-SIZE] --compress-level=3 --compress-type=zst --config=[TEST_PATH]/db-master/pgbackrest.conf --db-timeout=45 --delta --lock-path=[TEST_PATH]/db-master/lock --log-level-console=detail --log-level-file=[LOG-LEVEL-FILE] --log-level-stderr=off --log-path=[TEST_PATH]/db-master/log[] --no-log-timestamp --no-online --pg1-path=[TEST_PATH]/db-master/db/base-2 --process-max=1 --protocol-timeout=60 --repo1-hardlink --repo1-path=[TEST_PATH]/db-master/repo --stanza=db --start-fast --type=diff
Add time-based retention for full backups. The --repo-retention-full-type option allows retention of full backups based on a time period, specified in days. The new option will default to 'count' and therefore will not affect current installations. Setting repo-retention-full-type to 'time' will allow the user to use a time period, in days, to indicate full backup retention. Using this method, a full backup can be expired only if the time the backup completed is older than the number of days set with repo-retention-full (calculated from the moment the 'expire' command is run) and at least one full backup meets the retention period. If archive retention has not been configured, then the default settings will expire archives that are prior to the oldest retained full backup. For example, if there are three full backups ending in times that are 25 days old (F1), 20 days old (F2) and 10 days old (F3), then if the full retention period is 15 days, then only F1 will be expired; F2 will be retained because F1 is not at least 15 days old.
2020-05-08 21:25:03 +02:00
P00 WARN: option 'repo1-retention-full' is not set for 'repo1-retention-full-type=count', the repository may run out of space
Rename retention-* options to repo-retention-*.
2018-02-19 22:07:24 +02:00
HINT: to retain full backups indefinitely (without warning), set option 'repo1-retention-full' to the maximum.
Show Process ID in log output instead of filtering it out with the timestamp.
2016-09-14 23:30:57 +02:00
P00 INFO: last backup label = [BACKUP-FULL-2], version = [VERSION-1]
Add Zstandard compression support. Zstandard is a fast lossless compression algorithm targeting real-time compression scenarios at zlib-level and better compression ratios. It's backed by a very fast entropy stage, provided by Huff0 and FSE library. Zstandard version >= 1.0 is required, which is generally only available on newer distributions.
2020-05-04 21:25:27 +02:00
P00 WARN: diff backup cannot alter compress-type option to 'zst', reset to value in [BACKUP-FULL-2]
Full/Synthetic test refactor. * Combine hardlink and non/compressed in synthetic tests to reduce test time and improve coverage. * Change log level of hardlink logging to detail. * Cast size in S3 manifest to integer.
2017-06-15 21:32:10 +02:00
P00 WARN: diff backup cannot alter hardlink option to 'true', reset to value in [BACKUP-FULL-2]
Added the --checksum-page option. This option allows pgBackRest to validate page checksums in data files when checksums are enabled on PostgreSQL >= 9.3. Note that this functionality requires a C library which may not initially be available in OS packages. The option will automatically be enabled when the library is present and checksums are enabled on the cluster.
2016-12-13 01:54:07 +02:00
P00 WARN: diff backup cannot alter 'checksum-page' option to 'false', reset to 'true' from [BACKUP-FULL-2]
Add checksum delta for incremental backups. Use checksums rather than timestamps to determine if files have changed. This is useful in cases where the timestamps may not be trustworthy, e.g. when performing an incremental after failing over to a standby. If checksum delta is enabled then checksums will be used for verification of resumed backups, even if they are full. Resumes have always used checksums to verify the files in the repository, enabling delta performs checksums on the database files as well. Note that the user must manually enable this feature in cases were it would be useful or just keep in enabled all the time. A future commit will address automatically enabling the feature in cases where it seems likely to be useful. Contributed by Cynthia Shang.
2018-09-19 17:12:45 +02:00
P01 DETAIL: match file from prior backup [TEST_PATH]/db-master/db/base-2/base/32768/33001 (64KB, 36%) checksum 6bf316f11d28c28914ea9be92c00de9bea6d9a6b
Fix incorrect offline upper bound for ignoring page checksum errors. For offline backups the upper bound was being set to 0x0000FFFF0000FFFF rather than UINT64_MAX. This meant that page checksum errors might be ignored for databases with a lot of past WAL in offline mode. Online mode is not affected since the upper bound is retrieved from pg_start_backup().
2019-07-11 15:13:56 +02:00
P01 DETAIL: match file from prior backup [TEST_PATH]/db-master/db/base-2/base/32768/44000_init (32KB, 54%) checksum 7a16d165e4775f7c92e8cdf60c0af57313f0bf90
P01 DETAIL: match file from prior backup [TEST_PATH]/db-master/db/base-2/base/32768/33000.32767 (32KB, 72%) checksum 6e99b589e550e68e934fd235ccba59fe5b592a9e
P01 DETAIL: match file from prior backup [TEST_PATH]/db-master/db/base-2/base/32768/33000 (32KB, 90%) checksum 7a16d165e4775f7c92e8cdf60c0af57313f0bf90
Support configurable WAL segment size. PostgreSQL 11 introduces configurable WAL segment sizes, from 1MB to 1GB. There are two areas that needed to be updated to support this: building the archive-get queue and checking that WAL has been archived after a backup. Both operations require the WAL segment size to properly build a list. Checking the archive after a backup is still implemented in Perl and has an active database connection, so just get the WAL segment size from the database. The archive-get command does not have a connection to the database, so get the WAL segment size from pg_control instead. This requires a deeper inspection of pg_control than has been done in the past, so it seemed best to copy the relevant data structures from each version of PostgreSQL and build a generic interface layer to address them. While this approach is a bit verbose, it has the advantage of being relatively simple, and can easily be updated for new versions of PostgreSQL. Since the integration tests generate pg_control files for testing, teach Perl how to generate files with the correct offsets for both 32-bit and 64-bit architectures.
2018-09-25 11:24:42 +02:00
P01 DETAIL: match file from prior backup [TEST_PATH]/db-master/db/base-2/global/pg_control (8KB, 95%) checksum 4c77c900f7af0d9ab13fa9982051a42e0b637f6c
Add checksum delta for incremental backups. Use checksums rather than timestamps to determine if files have changed. This is useful in cases where the timestamps may not be trustworthy, e.g. when performing an incremental after failing over to a standby. If checksum delta is enabled then checksums will be used for verification of resumed backups, even if they are full. Resumes have always used checksums to verify the files in the repository, enabling delta performs checksums on the database files as well. Note that the user must manually enable this feature in cases were it would be useful or just keep in enabled all the time. A future commit will address automatically enabling the feature in cases where it seems likely to be useful. Contributed by Cynthia Shang.
2018-09-19 17:12:45 +02:00
P01 DETAIL: match file from prior backup [TEST_PATH]/db-master/db/base-2/base/1/12000 (8KB, 99%) checksum 22c98d248ff548311eda88559e4a8405ed77c003
P01 DETAIL: match file from prior backup [TEST_PATH]/db-master/db/base-2/postgresql.conf (21B, 99%) checksum 6721d92c9fcdf4248acff1f9a1377127d9064807
P01 INFO: backup file [TEST_PATH]/db-master/db/base-2/badchecksum.txt (11B, 99%) checksum f927212cd08d11a42a666b2f04235398e9ceeb51
P01 INFO: backup file [TEST_PATH]/db-master/db/base-2/changecontent.txt (7B, 99%) checksum a094d94583e209556d03c3c5da33131a065f1689
P01 DETAIL: match file from prior backup [TEST_PATH]/db-master/db/base-2/pg_stat/global.stat (5B, 99%) checksum e350d5ce0153f3e22d5db21cf2a4eff00f3ee877
P01 DETAIL: match file from prior backup [TEST_PATH]/db-master/db/base-2/changetime.txt (4B, 99%) checksum 88087292ed82e26f3eb824d0bffc05ccf7a30f8d
P01 DETAIL: match file from prior backup [TEST_PATH]/db-master/db/base-2/base/32768/PG_VERSION (3B, 99%) checksum 184473f470864e067ee3a22e64b47b0a1c356f29
P01 DETAIL: match file from prior backup [TEST_PATH]/db-master/db/base-2/base/16384/PG_VERSION (3B, 99%) checksum 184473f470864e067ee3a22e64b47b0a1c356f29
P01 DETAIL: match file from prior backup [TEST_PATH]/db-master/db/base-2/base/1/PG_VERSION (3B, 99%) checksum 184473f470864e067ee3a22e64b47b0a1c356f29
P01 DETAIL: match file from prior backup [TEST_PATH]/db-master/db/base-2/PG_VERSION (3B, 99%) checksum 184473f470864e067ee3a22e64b47b0a1c356f29
P01 INFO: backup file [TEST_PATH]/db-master/db/base-2/zerosize.txt (0B, 99%)
P01 INFO: backup file [TEST_PATH]/db-master/db/base-2/pg_tblspc/2/[TS_PATH-1]/32768/tablespace2c.txt (12B, 99%) checksum dfcb8679956b734706cf87259d50c88f83e80e66
Added the --checksum-page option. This option allows pgBackRest to validate page checksums in data files when checksums are enabled on PostgreSQL >= 9.3. Note that this functionality requires a C library which may not initially be available in OS packages. The option will automatically be enabled when the library is present and checksums are enabled on the cluster.
2016-12-13 01:54:07 +02:00
P00 WARN: page misalignment in file [TEST_PATH]/db-master/db/base-2/pg_tblspc/2/[TS_PATH-1]/32768/tablespace2c.txt: file size 12 is not divisible by page size 8192
Additional backup exclusions. * Exclude contents of pg_snapshots, pg_serial, pg_notify, and pg_dynshmem from backup since they are rebuilt on startup. * Exclude pg_internal.init files from backup since they are rebuilt on startup.
2017-09-04 14:26:57 +02:00
P01 INFO: backup file [TEST_PATH]/db-master/db/base-2/pg_tblspc/2/[TS_PATH-1]/32768/tablespace2.txt (7B, 100%) checksum dc7f76e43c46101b47acc55ae4d593a9e6983578
Added the --checksum-page option. This option allows pgBackRest to validate page checksums in data files when checksums are enabled on PostgreSQL >= 9.3. Note that this functionality requires a C library which may not initially be available in OS packages. The option will automatically be enabled when the library is present and checksums are enabled on the cluster.
2016-12-13 01:54:07 +02:00
P00 WARN: page misalignment in file [TEST_PATH]/db-master/db/base-2/pg_tblspc/2/[TS_PATH-1]/32768/tablespace2.txt: file size 7 is not divisible by page size 8192
The backup command is implemented entirely in C. For the most part this is a direct migration of the Perl code into C except as noted below. A backup can now be initiated from a linked directory. The link will not be stored in the manifest or recreated on restore. If a link or directory does not already exist in the restore location then a directory will be created. The logic for creating backup labels has been improved and it should no longer be possible to get a backup label earlier than the latest backup even with timezone changes or clock skew. This has never been an issue in the field that we know of, but we found it in testing. For online backups all times are fetched from the PostgreSQL primary host (before only copy start was). This doesn't affect backup integrity but it does prevent clock skew between hosts affecting backup duration reporting. Archive copy now works as expected when the archive and backup have different compression settings, i.e. when one is compressed and the other is not. This was a long-standing bug in the Perl code. Resume will now work even if hardlink settings have been changed. Reviewed by Cynthia Shang.
2019-12-14 00:14:26 +02:00
P00 DETAIL: reference pg_data/PG_VERSION to [BACKUP-FULL-2]
P00 DETAIL: reference pg_data/base/1/12000 to [BACKUP-FULL-2]
P00 DETAIL: reference pg_data/base/1/PG_VERSION to [BACKUP-FULL-2]
P00 DETAIL: reference pg_data/base/16384/PG_VERSION to [BACKUP-FULL-2]
P00 DETAIL: reference pg_data/base/32768/33000 to [BACKUP-FULL-2]
P00 DETAIL: reference pg_data/base/32768/33000.32767 to [BACKUP-FULL-2]
P00 DETAIL: reference pg_data/base/32768/33001 to [BACKUP-FULL-2]
P00 DETAIL: reference pg_data/base/32768/44000_init to [BACKUP-FULL-2]
P00 DETAIL: reference pg_data/base/32768/PG_VERSION to [BACKUP-FULL-2]
P00 DETAIL: reference pg_data/changetime.txt to [BACKUP-FULL-2]
P00 DETAIL: reference pg_data/global/pg_control to [BACKUP-FULL-2]
P00 DETAIL: reference pg_data/pg_stat/global.stat to [BACKUP-FULL-2]
P00 DETAIL: reference pg_data/postgresql.conf to [BACKUP-FULL-2]
P00 DETAIL: reference pg_data/special-!_.*'()&!@;:+,? to [BACKUP-FULL-2]
P00 DETAIL: reference pg_data/zero_from_start to [BACKUP-FULL-2]
Add checksum delta for incremental backups. Use checksums rather than timestamps to determine if files have changed. This is useful in cases where the timestamps may not be trustworthy, e.g. when performing an incremental after failing over to a standby. If checksum delta is enabled then checksums will be used for verification of resumed backups, even if they are full. Resumes have always used checksums to verify the files in the repository, enabling delta performs checksums on the database files as well. Note that the user must manually enable this feature in cases were it would be useful or just keep in enabled all the time. A future commit will address automatically enabling the feature in cases where it seems likely to be useful. Contributed by Cynthia Shang.
2018-09-19 17:12:45 +02:00
P00 INFO: diff backup size = 176KB
Remove integration tests that are now covered in the unit tests. Most of these tests are just checking that errors are thrown when required. These are well covered in various unit tests. The "cannot resume" tests are also well covered in the backup unit tests. Finally, config warnings are well covered in the config unit tests. There is more to be done here, but this accounts for the low-hanging fruit.
2019-12-18 03:14:45 +02:00
P00 INFO: new backup label = [BACKUP-DIFF-3]
Added a predictable log entry to signal that a command has completed successfully. For example a backup ends successfully with: INFO: backup command end: completed successfully.
2016-12-04 21:44:53 +02:00
P00 INFO: backup command end: completed successfully
Integration test improvements for disk and memory efficiency. Set log-level-file=off when more that one test will run. In this case is it impossible to see the logs anyway since they will be automatically cleaned up after the test. This improves performance pretty dramatically since trace-level logging is expensive. If a singe integration test is run then log-level-file is trace by default but can be changed with the --log-level-test-file option. Reduce buffer-size to 64k to save memory during testing and allow more processes to run in parallel. Update log replacement rules so that these options can change without affecting expect logs.
2019-12-17 22:23:07 +02:00
P00 INFO: expire command begin [BACKREST-VERSION]: --buffer-size=[BUFFER-SIZE] --config=[TEST_PATH]/db-master/pgbackrest.conf --lock-path=[TEST_PATH]/db-master/lock --log-level-console=detail --log-level-file=[LOG-LEVEL-FILE] --log-level-stderr=off --log-path=[TEST_PATH]/db-master/log[] --no-log-timestamp --repo1-path=[TEST_PATH]/db-master/repo --stanza=db
Rename retention-* options to repo-retention-*.
2018-02-19 22:07:24 +02:00
P00 INFO: option 'repo1-retention-archive' is not set - archive logs will not be expired
Added a predictable log entry to signal that a command has completed successfully. For example a backup ends successfully with: INFO: backup command end: completed successfully.
2016-12-04 21:44:53 +02:00
P00 INFO: expire command end: completed successfully
Log testing can now be enabled for certain deterministic tests. This works by comparing the generated logs against a previous copy. Currently only enabled for the backup/synthetic tests.
2015-05-05 19:08:48 +02:00
Rename test paths for clarity. This was worked out as part of the test suite refactor [c8f806a] but not committed with it because of the large number of expect logs changes involved. Keeping them separate made it easier to audit the changes in the refactor.
2016-06-24 16:54:31 +02:00
+ supplemental file: [TEST_PATH]/db-master/pgbackrest.conf
----------------------------------------------------------
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
[db]
Rename db-* options to pg-* and backup-* options to repo-* to improve consistency. * repo-* options are now indexed although only one is allowed. * List deprecated option names in documentation and command-line help.
2018-02-04 01:27:38 +02:00
pg1-path=[TEST_PATH]/db-master/db/base-2
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
New simpler configuration and consistent project/exe/path naming. * The repo-path option now always refers to the repository where backups and archive are stored, whether local or remote, so the repo-remote-path option has been removed. The new spool-path option can be used to define a location for queueing WAL segments when archiving asynchronously. Otherwise, a local repository is no longer required. * Implemented a new config format which should be far simpler to use. See the User Guide and Configuration Reference for details but for a simple configuration all options can now be placed in the stanza section. Options that are shared between stanzas can be placed in the [global] section. More complex configurations can still make use of command sections though this should be a rare use case. * The default configuration filename is now pgbackrest.conf instead of pg_backrest.conf. This was done for consistency with other naming changes but also to prevent old config files from being loaded accidentally. * The default repository name was changed from /var/lib/backup to /var/lib/pgbackrest. * Lock files are now stored in /tmp/pgbackrest by default. These days /run/pgbackrest would be the preferred location but that would require init scripts which are not part of this release. The lock-path option can be used to configure the lock directory. * Log files are now stored in /var/log/pgbackrest by default and no longer have the date appended so they can be managed with logrotate. The log-path option can be used to configure the lock directory. * Executable filename changed from pg_backrest to pgbackrest.
2016-04-14 15:30:54 +02:00
[db:restore]
Rename test paths for clarity. This was worked out as part of the test suite refactor [c8f806a] but not committed with it because of the large number of expect logs changes involved. Keeping them separate made it easier to audit the changes in the refactor.
2016-06-24 16:54:31 +02:00
tablespace-map=1=[TEST_PATH]/db-master/db/tablespace/ts1-2
tablespace-map=2=[TEST_PATH]/db-master/db/tablespace/ts2-2
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
New simpler configuration and consistent project/exe/path naming. * The repo-path option now always refers to the repository where backups and archive are stored, whether local or remote, so the repo-remote-path option has been removed. The new spool-path option can be used to define a location for queueing WAL segments when archiving asynchronously. Otherwise, a local repository is no longer required. * Implemented a new config format which should be far simpler to use. See the User Guide and Configuration Reference for details but for a simple configuration all options can now be placed in the stanza section. Options that are shared between stanzas can be placed in the [global] section. More complex configurations can still make use of command sections though this should be a rare use case. * The default configuration filename is now pgbackrest.conf instead of pg_backrest.conf. This was done for consistency with other naming changes but also to prevent old config files from being loaded accidentally. * The default repository name was changed from /var/lib/backup to /var/lib/pgbackrest. * Lock files are now stored in /tmp/pgbackrest by default. These days /run/pgbackrest would be the preferred location but that would require init scripts which are not part of this release. The lock-path option can be used to configure the lock directory. * Log files are now stored in /var/log/pgbackrest by default and no longer have the date appended so they can be managed with logrotate. The log-path option can be used to configure the lock directory. * Executable filename changed from pg_backrest to pgbackrest.
2016-04-14 15:30:54 +02:00
[global]
Integration test improvements for disk and memory efficiency. Set log-level-file=off when more that one test will run. In this case is it impossible to see the logs anyway since they will be automatically cleaned up after the test. This improves performance pretty dramatically since trace-level logging is expensive. If a singe integration test is run then log-level-file is trace by default but can be changed with the --log-level-test-file option. Reduce buffer-size to 64k to save memory during testing and allow more processes to run in parallel. Update log replacement rules so that these options can change without affecting expect logs.
2019-12-17 22:23:07 +02:00
buffer-size=[BUFFER-SIZE]
Fix custom settings for compress-level option being ignored. Reported by Jens Wilke.
2017-10-23 00:05:46 +02:00
compress-level=3
Add Zstandard compression support. Zstandard is a fast lossless compression algorithm targeting real-time compression scenarios at zlib-level and better compression ratios. It's backed by a very fast entropy stage, provided by Huff0 and FSE library. Zstandard version >= 1.0 is required, which is generally only available on newer distributions.
2020-05-04 21:25:27 +02:00
compress-type=zst
Storage and IO layer refactor: Refactor storage layer to allow for new repository filesystems using drivers. (Reviewed by Cynthia Shang.) Refactor IO layer to allow for new compression formats, checksum types, and other capabilities using filters. (Reviewed by Cynthia Shang.)
2017-06-09 23:51:41 +02:00
db-timeout=45
lock-path=[TEST_PATH]/db-master/lock
Reduce log level for all expect tests to detail. The C code is designed to be efficient rather than deterministic at the debug log level. As we move more testing from integration to unit tests it makes less sense to try and maintain the expect logs at this log level. Most of the expect logs have already been moved to detail level but mock/all still had tests at debug level. Change the logging defaults in the config file and remove as many references to log-level-console as possible.
2019-05-23 00:23:44 +02:00
log-level-console=detail
Integration test improvements for disk and memory efficiency. Set log-level-file=off when more that one test will run. In this case is it impossible to see the logs anyway since they will be automatically cleaned up after the test. This improves performance pretty dramatically since trace-level logging is expensive. If a singe integration test is run then log-level-file is trace by default but can be changed with the --log-level-test-file option. Reduce buffer-size to 64k to save memory during testing and allow more processes to run in parallel. Update log replacement rules so that these options can change without affecting expect logs.
2019-12-17 22:23:07 +02:00
log-level-file=[LOG-LEVEL-FILE]
Added the log-level-stderr option. Controls whether console log messages are sent to stderr or stdout. By default this is set to warn which represents a change in behavior from previous versions, even though it may be more intuitive. Setting log-level-stderr=off will preserve the old behavior. Suggested by Sascha Biberhofer.
2016-10-05 15:09:30 +02:00
log-level-stderr=off
Storage and IO layer refactor: Refactor storage layer to allow for new repository filesystems using drivers. (Reviewed by Cynthia Shang.) Refactor IO layer to allow for new compression formats, checksum types, and other capabilities using filters. (Reviewed by Cynthia Shang.)
2017-06-09 23:51:41 +02:00
log-path=[TEST_PATH]/db-master/log
Integration test improvements for disk and memory efficiency. Set log-level-file=off when more that one test will run. In this case is it impossible to see the logs anyway since they will be automatically cleaned up after the test. This improves performance pretty dramatically since trace-level logging is expensive. If a singe integration test is run then log-level-file is trace by default but can be changed with the --log-level-test-file option. Reduce buffer-size to 64k to save memory during testing and allow more processes to run in parallel. Update log replacement rules so that these options can change without affecting expect logs.
2019-12-17 22:23:07 +02:00
log-subprocess=[LOG-SUBPROCESS]
Set log-timestamp=n for integration tests. This means less filtering of logs needs to be done and new timestamps can be added without adding new filters.
2018-05-11 17:24:38 +02:00
log-timestamp=n
Storage and IO layer refactor: Refactor storage layer to allow for new repository filesystems using drivers. (Reviewed by Cynthia Shang.) Refactor IO layer to allow for new compression formats, checksum types, and other capabilities using filters. (Reviewed by Cynthia Shang.)
2017-06-09 23:51:41 +02:00
protocol-timeout=60
Rename db-* options to pg-* and backup-* options to repo-* to improve consistency. * repo-* options are now indexed although only one is allowed. * List deprecated option names in documentation and command-line help.
2018-02-04 01:27:38 +02:00
repo1-hardlink=y
repo1-path=[TEST_PATH]/db-master/repo
Consolidate archive-push tests into new archive test.
2017-06-21 14:38:53 +02:00
spool-path=[TEST_PATH]/db-master/spool
New simpler configuration and consistent project/exe/path naming. * The repo-path option now always refers to the repository where backups and archive are stored, whether local or remote, so the repo-remote-path option has been removed. The new spool-path option can be used to define a location for queueing WAL segments when archiving asynchronously. Otherwise, a local repository is no longer required. * Implemented a new config format which should be far simpler to use. See the User Guide and Configuration Reference for details but for a simple configuration all options can now be placed in the stanza section. Options that are shared between stanzas can be placed in the [global] section. More complex configurations can still make use of command sections though this should be a rare use case. * The default configuration filename is now pgbackrest.conf instead of pg_backrest.conf. This was done for consistency with other naming changes but also to prevent old config files from being loaded accidentally. * The default repository name was changed from /var/lib/backup to /var/lib/pgbackrest. * Lock files are now stored in /tmp/pgbackrest by default. These days /run/pgbackrest would be the preferred location but that would require init scripts which are not part of this release. The lock-path option can be used to configure the lock directory. * Log files are now stored in /var/log/pgbackrest by default and no longer have the date appended so they can be managed with logrotate. The log-path option can be used to configure the lock directory. * Executable filename changed from pg_backrest to pgbackrest.
2016-04-14 15:30:54 +02:00
[global:backup]
archive-copy=y
start-fast=y
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
Remove integration tests that are now covered in the unit tests. Most of these tests are just checking that errors are thrown when required. These are well covered in various unit tests. The "cannot resume" tests are also well covered in the backup unit tests. Finally, config warnings are well covered in the config unit tests. There is more to be done here, but this accounts for the low-hanging fruit.
2019-12-18 03:14:45 +02:00
+ supplemental file: [TEST_PATH]/db-master/repo/backup/db/[BACKUP-DIFF-3]/backup.manifest
Rename test paths for clarity. This was worked out as part of the test suite refactor [c8f806a] but not committed with it because of the large number of expect logs changes involved. Keeping them separate made it easier to audit the changes in the refactor.
2016-06-24 16:54:31 +02:00
-----------------------------------------------------------------------------------------
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
[backrest]
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
backrest-format=5
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
backrest-version="[VERSION-1]"
[backup]
Remove integration tests that are now covered in the unit tests. Most of these tests are just checking that errors are thrown when required. These are well covered in various unit tests. The "cannot resume" tests are also well covered in the backup unit tests. Finally, config warnings are well covered in the config unit tests. There is more to be done here, but this accounts for the low-hanging fruit.
2019-12-18 03:14:45 +02:00
backup-label="[BACKUP-DIFF-3]"
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
backup-prior="[BACKUP-FULL-2]"
backup-timestamp-copy-start=[TIMESTAMP]
backup-timestamp-start=[TIMESTAMP]
backup-timestamp-stop=[TIMESTAMP]
backup-type="diff"
[backup:db]
Database version refactoring. * Refactor db version constants into a separate module. * Update synthetic backup tests to PostgreSQL 9.4.
2016-08-12 04:35:24 +02:00
db-catalog-version=201409291
db-control-version=942
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
db-id=1
Automate generation of WAL and pg_control test files. The existing static files would not work with 32-bit or big-endian systems so create functions to generate these files dynamically rather than creating a bunch of new static files.
2017-11-19 03:02:54 +02:00
db-system-id=1000000000000000094
Database version refactoring. * Refactor db version constants into a separate module. * Update synthetic backup tests to PostgreSQL 9.4.
2016-08-12 04:35:24 +02:00
db-version="9.4"
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
[backup:option]
Set archive-check option in manifest correctly when offline. Archive check does not run when in offline backup mode but the option was set to true in the manifest. It's harmless since these options are informational only but it could cause confusion when debugging.
2019-11-28 15:27:21 +02:00
option-archive-check=false
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
option-archive-copy=true
Backup from a standby cluster. A connection to the primary cluster is still required to start/stop the backup and copy files that are not replicated, but the vast majority of files are copied from the standby in order to reduce load on the master.
2016-08-25 17:25:46 +02:00
option-backup-standby=false
Integration test improvements for disk and memory efficiency. Set log-level-file=off when more that one test will run. In this case is it impossible to see the logs anyway since they will be automatically cleaned up after the test. This improves performance pretty dramatically since trace-level logging is expensive. If a singe integration test is run then log-level-file is trace by default but can be changed with the --log-level-test-file option. Reduce buffer-size to 64k to save memory during testing and allow more processes to run in parallel. Update log replacement rules so that these options can change without affecting expect logs.
2019-12-17 22:23:07 +02:00
option-buffer-size=[BUFFER-SIZE]
Added the --checksum-page option. This option allows pgBackRest to validate page checksums in data files when checksums are enabled on PostgreSQL >= 9.3. Note that this functionality requires a C library which may not initially be available in OS packages. The option will automatically be enabled when the library is present and checksums are enabled on the cluster.
2016-12-13 01:54:07 +02:00
option-checksum-page=true
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
option-compress=false
Add additional options to backup.manifest for debugging purposes. Add the buffer-size, compress-level, compress-level-network, and process-max options to the backup:option section in backup.manifest to aid in debugging. It may also make sense to propagate these options up to backup.info so they can be displayed in the info command, but for now this is deemed sufficient. Contributed by blogh.
2019-03-10 11:03:52 +02:00
option-compress-level=3
option-compress-level-network=3
Add infrastructure for multiple compression type support. Add compress-type option and deprecate compress option. Since the compress option is boolean it won't work with multiple compression types. Add logic to cfgLoadUpdateOption() to update compress-type if it is not set directly. The compress option should no longer be referenced outside the cfgLoadUpdateOption() function. Add common/compress/helper module to contain interface functions that work with multiple compression types. Code outside this module should no longer call specific compression drivers, though it may be OK to reference a specific compression type using the new interface (e.g., saving backup history files in gz format). Unit tests only test compression using the gz format because other formats may not be available in all builds. It is the job of integration tests to exercise all compression types. Additional compression types will be added in future commits.
2020-03-06 21:41:03 +02:00
option-compress-type="none"
Add checksum delta for incremental backups. Use checksums rather than timestamps to determine if files have changed. This is useful in cases where the timestamps may not be trustworthy, e.g. when performing an incremental after failing over to a standby. If checksum delta is enabled then checksums will be used for verification of resumed backups, even if they are full. Resumes have always used checksums to verify the files in the repository, enabling delta performs checksums on the database files as well. Note that the user must manually enable this feature in cases were it would be useful or just keep in enabled all the time. A future commit will address automatically enabling the feature in cases where it seems likely to be useful. Contributed by Cynthia Shang.
2018-09-19 17:12:45 +02:00
option-delta=true
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
option-hardlink=false
Closed #167: Rename `--no-start-stop` option to `--no-online`.
2016-02-20 02:32:02 +02:00
option-online=false
Add additional options to backup.manifest for debugging purposes. Add the buffer-size, compress-level, compress-level-network, and process-max options to the backup:option section in backup.manifest to aid in debugging. It may also make sense to propagate these options up to backup.info so they can be displayed in the info command, but for now this is deemed sufficient. Contributed by blogh.
2019-03-10 11:03:52 +02:00
option-process-max=1
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
[backup:target]
Rename test paths for clarity. This was worked out as part of the test suite refactor [c8f806a] but not committed with it because of the large number of expect logs changes involved. Keeping them separate made it easier to audit the changes in the refactor.
2016-06-24 16:54:31 +02:00
pg_data={"path":"[TEST_PATH]/db-master/db/base-2","type":"path"}
pg_tblspc/2={"path":"[TEST_PATH]/db-master/db/tablespace/ts2-2","tablespace-id":"2","tablespace-name":"ts2","type":"link"}
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
[target:file]
Fixed a issue where tablespaces were copied from the master during standby backup.
2016-09-04 15:19:44 +02:00
pg_data/PG_VERSION={"checksum":"184473f470864e067ee3a22e64b47b0a1c356f29","master":true,"reference":"[BACKUP-FULL-2]","size":[SIZE],"timestamp":[TIMESTAMP-1]}
pg_data/badchecksum.txt={"checksum":"f927212cd08d11a42a666b2f04235398e9ceeb51","master":true,"size":[SIZE],"timestamp":[TIMESTAMP-1]}
Added the --checksum-page option. This option allows pgBackRest to validate page checksums in data files when checksums are enabled on PostgreSQL >= 9.3. Note that this functionality requires a C library which may not initially be available in OS packages. The option will automatically be enabled when the library is present and checksums are enabled on the cluster.
2016-12-13 01:54:07 +02:00
pg_data/base/1/12000={"checksum":"22c98d248ff548311eda88559e4a8405ed77c003","checksum-page":true,"reference":"[BACKUP-FULL-2]","size":[SIZE],"timestamp":[TIMESTAMP-1]}
Database version refactoring. * Refactor db version constants into a separate module. * Update synthetic backup tests to PostgreSQL 9.4.
2016-08-12 04:35:24 +02:00
pg_data/base/1/PG_VERSION={"checksum":"184473f470864e067ee3a22e64b47b0a1c356f29","mode":"0660","reference":"[BACKUP-FULL-2]","size":[SIZE],"timestamp":[TIMESTAMP-1]}
pg_data/base/16384/PG_VERSION={"checksum":"184473f470864e067ee3a22e64b47b0a1c356f29","reference":"[BACKUP-FULL-2]","size":[SIZE],"timestamp":[TIMESTAMP-1]}
Fix incorrect offline upper bound for ignoring page checksum errors. For offline backups the upper bound was being set to 0x0000FFFF0000FFFF rather than UINT64_MAX. This meant that page checksum errors might be ignored for databases with a lot of past WAL in offline mode. Online mode is not affected since the upper bound is retrieved from pg_start_backup().
2019-07-11 15:13:56 +02:00
pg_data/base/32768/33000={"checksum":"7a16d165e4775f7c92e8cdf60c0af57313f0bf90","checksum-page":true,"reference":"[BACKUP-FULL-2]","size":[SIZE],"timestamp":[TIMESTAMP-1]}
pg_data/base/32768/33000.32767={"checksum":"6e99b589e550e68e934fd235ccba59fe5b592a9e","checksum-page":true,"reference":"[BACKUP-FULL-2]","size":[SIZE],"timestamp":[TIMESTAMP-1]}
Added the --checksum-page option. This option allows pgBackRest to validate page checksums in data files when checksums are enabled on PostgreSQL >= 9.3. Note that this functionality requires a C library which may not initially be available in OS packages. The option will automatically be enabled when the library is present and checksums are enabled on the cluster.
2016-12-13 01:54:07 +02:00
pg_data/base/32768/33001={"checksum":"6bf316f11d28c28914ea9be92c00de9bea6d9a6b","checksum-page":false,"checksum-page-error":[0,[3,5],7],"reference":"[BACKUP-FULL-2]","size":[SIZE],"timestamp":[TIMESTAMP-1]}
Fix incorrect offline upper bound for ignoring page checksum errors. For offline backups the upper bound was being set to 0x0000FFFF0000FFFF rather than UINT64_MAX. This meant that page checksum errors might be ignored for databases with a lot of past WAL in offline mode. Online mode is not affected since the upper bound is retrieved from pg_start_backup().
2019-07-11 15:13:56 +02:00
pg_data/base/32768/44000_init={"checksum":"7a16d165e4775f7c92e8cdf60c0af57313f0bf90","checksum-page":true,"reference":"[BACKUP-FULL-2]","size":[SIZE],"timestamp":[TIMESTAMP-1]}
Database version refactoring. * Refactor db version constants into a separate module. * Update synthetic backup tests to PostgreSQL 9.4.
2016-08-12 04:35:24 +02:00
pg_data/base/32768/PG_VERSION={"checksum":"184473f470864e067ee3a22e64b47b0a1c356f29","reference":"[BACKUP-FULL-2]","size":[SIZE],"timestamp":[TIMESTAMP-1]}
Add checksum delta for incremental backups. Use checksums rather than timestamps to determine if files have changed. This is useful in cases where the timestamps may not be trustworthy, e.g. when performing an incremental after failing over to a standby. If checksum delta is enabled then checksums will be used for verification of resumed backups, even if they are full. Resumes have always used checksums to verify the files in the repository, enabling delta performs checksums on the database files as well. Note that the user must manually enable this feature in cases were it would be useful or just keep in enabled all the time. A future commit will address automatically enabling the feature in cases where it seems likely to be useful. Contributed by Cynthia Shang.
2018-09-19 17:12:45 +02:00
pg_data/changecontent.txt={"checksum":"a094d94583e209556d03c3c5da33131a065f1689","master":true,"size":[SIZE],"timestamp":[TIMESTAMP-1]}
pg_data/changetime.txt={"checksum":"88087292ed82e26f3eb824d0bffc05ccf7a30f8d","master":true,"reference":"[BACKUP-FULL-2]","size":[SIZE],"timestamp":[TIMESTAMP-2]}
Support configurable WAL segment size. PostgreSQL 11 introduces configurable WAL segment sizes, from 1MB to 1GB. There are two areas that needed to be updated to support this: building the archive-get queue and checking that WAL has been archived after a backup. Both operations require the WAL segment size to properly build a list. Checking the archive after a backup is still implemented in Perl and has an active database connection, so just get the WAL segment size from the database. The archive-get command does not have a connection to the database, so get the WAL segment size from pg_control instead. This requires a deeper inspection of pg_control than has been done in the past, so it seemed best to copy the relevant data structures from each version of PostgreSQL and build a generic interface layer to address them. While this approach is a bit verbose, it has the advantage of being relatively simple, and can easily be updated for new versions of PostgreSQL. Since the integration tests generate pg_control files for testing, teach Perl how to generate files with the correct offsets for both 32-bit and 64-bit architectures.
2018-09-25 11:24:42 +02:00
pg_data/global/pg_control={"checksum":"4c77c900f7af0d9ab13fa9982051a42e0b637f6c","master":true,"reference":"[BACKUP-FULL-2]","size":[SIZE],"timestamp":[TIMESTAMP-2]}
Fixed a issue where tablespaces were copied from the master during standby backup.
2016-09-04 15:19:44 +02:00
pg_data/pg_stat/global.stat={"checksum":"e350d5ce0153f3e22d5db21cf2a4eff00f3ee877","master":true,"reference":"[BACKUP-FULL-2]","size":[SIZE],"timestamp":[TIMESTAMP-2]}
pg_data/postgresql.conf={"checksum":"6721d92c9fcdf4248acff1f9a1377127d9064807","master":true,"reference":"[BACKUP-FULL-2]","size":[SIZE],"timestamp":[TIMESTAMP-2]}
Use minio for integration testing. ScalityS3 has not received any maintenance in years and is slow to start which is bad for testing. Replace it with minio which starts quickly and ships as a single executable or a tiny container. Minio has stricter limits on allowable characters but should still provide enough coverage to show that our encoding is working correctly. This commit also includes the upgrade to openssl 1.1.1 in the Ubuntu 18.04 container.
2019-07-03 04:20:35 +02:00
pg_data/special-!_.*'()&!@;:+,?={"master":true,"reference":"[BACKUP-FULL-2]","size":[SIZE],"timestamp":[TIMESTAMP-1]}
Allow zero-size files in backup manifest to reference a prior manifest regardless of timestamp delta. Contributed by Cynthia Shang.
2018-08-24 22:50:33 +02:00
pg_data/zero_from_start={"master":true,"reference":"[BACKUP-FULL-2]","size":[SIZE],"timestamp":[TIMESTAMP-1]}
Add zero-length file to mock/all test.
2018-07-05 21:40:50 +02:00
pg_data/zerosize.txt={"master":true,"size":[SIZE],"timestamp":[TIMESTAMP-2]}
Added the --checksum-page option. This option allows pgBackRest to validate page checksums in data files when checksums are enabled on PostgreSQL >= 9.3. Note that this functionality requires a C library which may not initially be available in OS packages. The option will automatically be enabled when the library is present and checksums are enabled on the cluster.
2016-12-13 01:54:07 +02:00
pg_tblspc/2/[TS_PATH-1]/32768/tablespace2.txt={"checksum":"dc7f76e43c46101b47acc55ae4d593a9e6983578","checksum-page":false,"size":[SIZE],"timestamp":[TIMESTAMP-1]}
pg_tblspc/2/[TS_PATH-1]/32768/tablespace2c.txt={"checksum":"dfcb8679956b734706cf87259d50c88f83e80e66","checksum-page":false,"size":[SIZE],"timestamp":[TIMESTAMP-1]}
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
[target:file:default]
Continuous integration using travis-ci.
2017-02-21 15:59:23 +02:00
group="[GROUP-1]"
Fixed a issue where tablespaces were copied from the master during standby backup.
2016-09-04 15:19:44 +02:00
master=false
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
mode="0600"
Regression tests can now be run as any properly-configured user, not just vagrant.
2016-10-01 19:39:44 +02:00
user="[USER-1]"
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
[target:link]
Rename test paths for clarity. This was worked out as part of the test suite refactor [c8f806a] but not committed with it because of the large number of expect logs changes involved. Keeping them separate made it easier to audit the changes in the refactor.
2016-06-24 16:54:31 +02:00
pg_data/pg_tblspc/2={"destination":"[TEST_PATH]/db-master/db/tablespace/ts2-2"}
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
[target:link:default]
Continuous integration using travis-ci.
2017-02-21 15:59:23 +02:00
group="[GROUP-1]"
Regression tests can now be run as any properly-configured user, not just vagrant.
2016-10-01 19:39:44 +02:00
user="[USER-1]"
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
[target:path]
The restore command is implemented entirely in C. For the most part this is a direct migration of the Perl code into C. There is one important behavioral change with regard to how file permissions are handled. The Perl code tried to set ownership as it was in the manifest even when running as an unprivileged user. This usually just led to errors and frustration. The C code works like this: If a restore is run as a non-root user (the typical scenario) then all files restored will belong to the user/group executing pgBackRest. If existing files are not owned by the executing user/group then an error will result if the ownership cannot be updated to the executing user/group. In that case the file ownership will need to be updated by a privileged user before the restore can be retried. If a restore is run as the root user then pgBackRest will attempt to recreate the ownership recorded in the manifest when the backup was made. Only user/group names are stored in the manifest so the same names must exist on the restore host for this to work. If the user/group name cannot be found locally then the user/group of the PostgreSQL data directory will be used and finally root if the data directory user/group cannot be mapped to a name. Reviewed by Cynthia Shang.
2019-09-26 13:52:02 +02:00
pg_data={"mode":"0777"}
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
pg_data/base={}
Allow selective restore of databases from a cluster backup. This feature can result in major space and time savings when only specific databases are restored. Unrestored databases will not be accessible but must be manually dropped before they will be removed from the shared catalogue.
2016-05-11 15:21:39 +02:00
pg_data/base/1={}
pg_data/base/16384={}
pg_data/base/32768={}
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
pg_data/global={}
pg_data/pg_clog={}
Exclude directories during backup that are cleaned, recreated, or zeroed by PostgreSQL at startup. These include (depending on the version where they were introduced): pgsql_tmp, pg_dynshmem, pg_notify, pg_replslot, pg_serial, pg_snapshots, pg_stat_tmp, pg_subtrans. The postgresql.auto.conf.tmp file is now excluded in addition to files that were already excluded: backup_label.old, postmaster.opts, postmaster.pid, recovery.conf, recovery.done.
2016-08-16 15:35:16 +02:00
pg_data/pg_dynshmem={}
Allow arbitrary directories and/or files to be excluded from a backup. Misuse of this feature can lead to inconsistent backups so read the --exclude documentation carefully before using.
2018-08-27 21:51:05 +02:00
pg_data/pg_log={}
Exclude directories during backup that are cleaned, recreated, or zeroed by PostgreSQL at startup. These include (depending on the version where they were introduced): pgsql_tmp, pg_dynshmem, pg_notify, pg_replslot, pg_serial, pg_snapshots, pg_stat_tmp, pg_subtrans. The postgresql.auto.conf.tmp file is now excluded in addition to files that were already excluded: backup_label.old, postmaster.opts, postmaster.pid, recovery.conf, recovery.done.
2016-08-16 15:35:16 +02:00
pg_data/pg_notify={}
pg_data/pg_replslot={}
pg_data/pg_serial={}
pg_data/pg_snapshots={}
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
pg_data/pg_stat={}
Exclude directories during backup that are cleaned, recreated, or zeroed by PostgreSQL at startup. These include (depending on the version where they were introduced): pgsql_tmp, pg_dynshmem, pg_notify, pg_replslot, pg_serial, pg_snapshots, pg_stat_tmp, pg_subtrans. The postgresql.auto.conf.tmp file is now excluded in addition to files that were already excluded: backup_label.old, postmaster.opts, postmaster.pid, recovery.conf, recovery.done.
2016-08-16 15:35:16 +02:00
pg_data/pg_stat_tmp={}
pg_data/pg_subtrans={}
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
pg_data/pg_tblspc={}
The restore command is implemented entirely in C. For the most part this is a direct migration of the Perl code into C. There is one important behavioral change with regard to how file permissions are handled. The Perl code tried to set ownership as it was in the manifest even when running as an unprivileged user. This usually just led to errors and frustration. The C code works like this: If a restore is run as a non-root user (the typical scenario) then all files restored will belong to the user/group executing pgBackRest. If existing files are not owned by the executing user/group then an error will result if the ownership cannot be updated to the executing user/group. In that case the file ownership will need to be updated by a privileged user before the restore can be retried. If a restore is run as the root user then pgBackRest will attempt to recreate the ownership recorded in the manifest when the backup was made. Only user/group names are stored in the manifest so the same names must exist on the restore host for this to work. If the user/group name cannot be found locally then the user/group of the PostgreSQL data directory will be used and finally root if the data directory user/group cannot be mapped to a name. Reviewed by Cynthia Shang.
2019-09-26 13:52:02 +02:00
pg_tblspc={"mode":"0777"}
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
pg_tblspc/2={}
pg_tblspc/2/[TS_PATH-1]={}
Allow selective restore of databases from a cluster backup. This feature can result in major space and time savings when only specific databases are restored. Unrestored databases will not be accessible but must be manually dropped before they will be removed from the shared catalogue.
2016-05-11 15:21:39 +02:00
pg_tblspc/2/[TS_PATH-1]/32768={}
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
[target:path:default]
Continuous integration using travis-ci.
2017-02-21 15:59:23 +02:00
group="[GROUP-1]"
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
mode="0700"
Regression tests can now be run as any properly-configured user, not just vagrant.
2016-10-01 19:39:44 +02:00
user="[USER-1]"
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
Move info file checksum to the end of the file. Putting the checksum at the beginning of the file made it impossible to stream the file out when saving. The entire file had to be held in memory while it was checksummed so the checksum could be written at the beginning. Instead place the checksum at the end. This does not break the existing Perl or C code since the read is not order dependent. There are no plans to improve the Perl code to take advantage of this change, but it will make the C implementation more efficient. Reviewed by Cynthia Shang.
2019-08-22 01:45:48 +02:00
[backrest]
backrest-checksum="[CHECKSUM]"
Rename test paths for clarity. This was worked out as part of the test suite refactor [c8f806a] but not committed with it because of the large number of expect logs changes involved. Keeping them separate made it easier to audit the changes in the refactor.
2016-06-24 16:54:31 +02:00
+ supplemental file: [TEST_PATH]/db-master/repo/backup/db/backup.info
---------------------------------------------------------------------
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
[backrest]
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
backrest-format=5
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
backrest-version="[VERSION-1]"
[backup:current]
Set archive-check option in manifest correctly when offline. Archive check does not run when in offline backup mode but the option was set to true in the manifest. It's harmless since these options are informational only but it could cause confusion when debugging.
2019-11-28 15:27:21 +02:00
[BACKUP-FULL-2]={"backrest-format":5,"backrest-version":"[VERSION-1]","backup-archive-start":null,"backup-archive-stop":null,"backup-info-repo-size":[SIZE],"backup-info-repo-size-delta":[DELTA],"backup-info-size":[SIZE],"backup-info-size-delta":[DELTA],"backup-timestamp-start":[TIMESTAMP],"backup-timestamp-stop":[TIMESTAMP],"backup-type":"full","db-id":1,"option-archive-check":false,"option-archive-copy":true,"option-backup-standby":false,"option-checksum-page":true,"option-compress":false,"option-hardlink":false,"option-online":false}
Remove integration tests that are now covered in the unit tests. Most of these tests are just checking that errors are thrown when required. These are well covered in various unit tests. The "cannot resume" tests are also well covered in the backup unit tests. Finally, config warnings are well covered in the config unit tests. There is more to be done here, but this accounts for the low-hanging fruit.
2019-12-18 03:14:45 +02:00
[BACKUP-INCR-2]={"backrest-format":5,"backrest-version":"[VERSION-1]","backup-archive-start":null,"backup-archive-stop":null,"backup-info-repo-size":[SIZE],"backup-info-repo-size-delta":[DELTA],"backup-info-size":[SIZE],"backup-info-size-delta":[DELTA],"backup-prior":"[BACKUP-FULL-2]","backup-reference":["[BACKUP-FULL-2]"],"backup-timestamp-start":[TIMESTAMP],"backup-timestamp-stop":[TIMESTAMP],"backup-type":"incr","db-id":1,"option-archive-check":false,"option-archive-copy":true,"option-backup-standby":false,"option-checksum-page":true,"option-compress":false,"option-hardlink":false,"option-online":false}
[BACKUP-DIFF-1]={"backrest-format":5,"backrest-version":"[VERSION-1]","backup-archive-start":null,"backup-archive-stop":null,"backup-info-repo-size":[SIZE],"backup-info-repo-size-delta":[DELTA],"backup-info-size":[SIZE],"backup-info-size-delta":[DELTA],"backup-prior":"[BACKUP-FULL-2]","backup-reference":["[BACKUP-FULL-2]"],"backup-timestamp-start":[TIMESTAMP],"backup-timestamp-stop":[TIMESTAMP],"backup-type":"diff","db-id":1,"option-archive-check":false,"option-archive-copy":true,"option-backup-standby":false,"option-checksum-page":true,"option-compress":false,"option-hardlink":false,"option-online":false}
[BACKUP-INCR-3]={"backrest-format":5,"backrest-version":"[VERSION-1]","backup-archive-start":null,"backup-archive-stop":null,"backup-info-repo-size":[SIZE],"backup-info-repo-size-delta":[DELTA],"backup-info-size":[SIZE],"backup-info-size-delta":[DELTA],"backup-prior":"[BACKUP-DIFF-1]","backup-reference":["[BACKUP-FULL-2]","[BACKUP-DIFF-1]"],"backup-timestamp-start":[TIMESTAMP],"backup-timestamp-stop":[TIMESTAMP],"backup-type":"incr","db-id":1,"option-archive-check":false,"option-archive-copy":true,"option-backup-standby":false,"option-checksum-page":true,"option-compress":false,"option-hardlink":false,"option-online":false}
[BACKUP-INCR-4]={"backrest-format":5,"backrest-version":"[VERSION-1]","backup-archive-start":null,"backup-archive-stop":null,"backup-info-repo-size":[SIZE],"backup-info-repo-size-delta":[DELTA],"backup-info-size":[SIZE],"backup-info-size-delta":[DELTA],"backup-prior":"[BACKUP-INCR-3]","backup-reference":["[BACKUP-FULL-2]","[BACKUP-DIFF-1]","[BACKUP-INCR-3]"],"backup-timestamp-start":[TIMESTAMP],"backup-timestamp-stop":[TIMESTAMP],"backup-type":"incr","db-id":1,"option-archive-check":false,"option-archive-copy":true,"option-backup-standby":false,"option-checksum-page":true,"option-compress":false,"option-hardlink":false,"option-online":false}
Set archive-check option in manifest correctly when offline. Archive check does not run when in offline backup mode but the option was set to true in the manifest. It's harmless since these options are informational only but it could cause confusion when debugging.
2019-11-28 15:27:21 +02:00
[BACKUP-DIFF-2]={"backrest-format":5,"backrest-version":"[VERSION-1]","backup-archive-start":null,"backup-archive-stop":null,"backup-info-repo-size":[SIZE],"backup-info-repo-size-delta":[DELTA],"backup-info-size":[SIZE],"backup-info-size-delta":[DELTA],"backup-prior":"[BACKUP-FULL-2]","backup-reference":["[BACKUP-FULL-2]"],"backup-timestamp-start":[TIMESTAMP],"backup-timestamp-stop":[TIMESTAMP],"backup-type":"diff","db-id":1,"option-archive-check":false,"option-archive-copy":true,"option-backup-standby":false,"option-checksum-page":true,"option-compress":false,"option-hardlink":false,"option-online":false}
[BACKUP-DIFF-3]={"backrest-format":5,"backrest-version":"[VERSION-1]","backup-archive-start":null,"backup-archive-stop":null,"backup-info-repo-size":[SIZE],"backup-info-repo-size-delta":[DELTA],"backup-info-size":[SIZE],"backup-info-size-delta":[DELTA],"backup-prior":"[BACKUP-FULL-2]","backup-reference":["[BACKUP-FULL-2]"],"backup-timestamp-start":[TIMESTAMP],"backup-timestamp-stop":[TIMESTAMP],"backup-type":"diff","db-id":1,"option-archive-check":false,"option-archive-copy":true,"option-backup-standby":false,"option-checksum-page":true,"option-compress":false,"option-hardlink":false,"option-online":false}
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
[db]
Database version refactoring. * Refactor db version constants into a separate module. * Update synthetic backup tests to PostgreSQL 9.4.
2016-08-12 04:35:24 +02:00
db-catalog-version=201409291
db-control-version=942
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
db-id=1
Automate generation of WAL and pg_control test files. The existing static files would not work with 32-bit or big-endian systems so create functions to generate these files dynamically rather than creating a bunch of new static files.
2017-11-19 03:02:54 +02:00
db-system-id=1000000000000000094
Database version refactoring. * Refactor db version constants into a separate module. * Update synthetic backup tests to PostgreSQL 9.4.
2016-08-12 04:35:24 +02:00
db-version="9.4"
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
[db:history]
Automate generation of WAL and pg_control test files. The existing static files would not work with 32-bit or big-endian systems so create functions to generate these files dynamically rather than creating a bunch of new static files.
2017-11-19 03:02:54 +02:00
1={"db-catalog-version":201409291,"db-control-version":942,"db-system-id":1000000000000000094,"db-version":"9.4"}
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
Improve performance of info file load/save. Info files required three copies in memory to be loaded (the original string, an ini representation, and the final info object). Not only was this memory inefficient but the Ini object does sequential scans when searching for keys making large files very slow to load. This has not been an issue since archive.info and backup.info are very small, but it becomes a big deal when loading manifests with hundreds of thousands of files. Instead of holding copies of the data in memory, use a callback to deliver the ini data directly to the object when loading. Use a similar method for save to avoid having an intermediate copy. Save is a bit complex because sections/keys must be written in alpha order or older versions of pgBackRest will not calculate the correct checksum. Also move the load retry logic to helper functions rather than embedding it in the Info object. This allows for more flexibility in loading and ensures that stack traces will be available when developing unit tests. Reviewed by Cynthia Shang.
2019-09-06 19:48:28 +02:00
[backrest]
backrest-checksum="[CHECKSUM]"
Major refactor of the test suite. * Make the code more modular and object-oriented. * Multiple Docker containers can now be created for a single test to simulate more realistic environments.
2016-06-24 14:12:58 +02:00
full backup - update file (db-master host)
Reduce log level for all expect tests to detail. The C code is designed to be efficient rather than deterministic at the debug log level. As we move more testing from integration to unit tests it makes less sense to try and maintain the expect logs at this log level. Most of the expect logs have already been moved to detail level but mock/all still had tests at debug level. Change the logging defaults in the config file and remove as many references to log-level-console as possible.
2019-05-23 00:23:44 +02:00
> [CONTAINER-EXEC] db-master [BACKREST-BIN] --config=[TEST_PATH]/db-master/pgbackrest.conf --no-online --type=full --stanza=db backup
Log testing can now be enabled for certain deterministic tests. This works by comparing the generated logs against a previous copy. Currently only enabled for the backup/synthetic tests.
2015-05-05 19:08:48 +02:00
------------------------------------------------------------------------------------------------------------------------------------
Add Zstandard compression support. Zstandard is a fast lossless compression algorithm targeting real-time compression scenarios at zlib-level and better compression ratios. It's backed by a very fast entropy stage, provided by Huff0 and FSE library. Zstandard version >= 1.0 is required, which is generally only available on newer distributions.
2020-05-04 21:25:27 +02:00
P00 INFO: backup command begin [BACKREST-VERSION]: --buffer-size=[BUFFER-SIZE] --compress-level=3 --compress-type=zst --config=[TEST_PATH]/db-master/pgbackrest.conf --db-timeout=45 --lock-path=[TEST_PATH]/db-master/lock --log-level-console=detail --log-level-file=[LOG-LEVEL-FILE] --log-level-stderr=off --log-path=[TEST_PATH]/db-master/log[] --no-log-timestamp --no-online --pg1-path=[TEST_PATH]/db-master/db/base-2 --protocol-timeout=60 --repo1-hardlink --repo1-path=[TEST_PATH]/db-master/repo --stanza=db --start-fast --type=full
Add time-based retention for full backups. The --repo-retention-full-type option allows retention of full backups based on a time period, specified in days. The new option will default to 'count' and therefore will not affect current installations. Setting repo-retention-full-type to 'time' will allow the user to use a time period, in days, to indicate full backup retention. Using this method, a full backup can be expired only if the time the backup completed is older than the number of days set with repo-retention-full (calculated from the moment the 'expire' command is run) and at least one full backup meets the retention period. If archive retention has not been configured, then the default settings will expire archives that are prior to the oldest retained full backup. For example, if there are three full backups ending in times that are 25 days old (F1), 20 days old (F2) and 10 days old (F3), then if the full retention period is 15 days, then only F1 will be expired; F2 will be retained because F1 is not at least 15 days old.
2020-05-08 21:25:03 +02:00
P00 WARN: option 'repo1-retention-full' is not set for 'repo1-retention-full-type=count', the repository may run out of space
Rename retention-* options to repo-retention-*.
2018-02-19 22:07:24 +02:00
HINT: to retain full backups indefinitely (without warning), set option 'repo1-retention-full' to the maximum.
Exclude temporary and unlogged relation (table/index) files from backup. Implemented using the same logic as the patches adding this feature to PostgreSQL, 8694cc96 and 920a5e50. Temporary relation exclusion is enabled in PostgreSQL ≥ 9.0. Unlogged relation exclusion is enabled in PostgreSQL ≥ 9.1, where the feature was introduced. Contributed by Cynthia Shang.
2018-07-31 00:53:34 +02:00
P01 INFO: backup file [TEST_PATH]/db-master/db/base-2/base/32768/33001 (64KB, 36%) checksum 6bf316f11d28c28914ea9be92c00de9bea6d9a6b
Fix incorrect offline upper bound for ignoring page checksum errors. For offline backups the upper bound was being set to 0x0000FFFF0000FFFF rather than UINT64_MAX. This meant that page checksum errors might be ignored for databases with a lot of past WAL in offline mode. Online mode is not affected since the upper bound is retrieved from pg_start_backup().
2019-07-11 15:13:56 +02:00
P01 INFO: backup file [TEST_PATH]/db-master/db/base-2/base/32768/44000_init (32KB, 54%) checksum 7a16d165e4775f7c92e8cdf60c0af57313f0bf90
P01 INFO: backup file [TEST_PATH]/db-master/db/base-2/base/32768/33000.32767 (32KB, 72%) checksum 6e99b589e550e68e934fd235ccba59fe5b592a9e
P01 INFO: backup file [TEST_PATH]/db-master/db/base-2/base/32768/33000 (32KB, 90%) checksum 7a16d165e4775f7c92e8cdf60c0af57313f0bf90
Support configurable WAL segment size. PostgreSQL 11 introduces configurable WAL segment sizes, from 1MB to 1GB. There are two areas that needed to be updated to support this: building the archive-get queue and checking that WAL has been archived after a backup. Both operations require the WAL segment size to properly build a list. Checking the archive after a backup is still implemented in Perl and has an active database connection, so just get the WAL segment size from the database. The archive-get command does not have a connection to the database, so get the WAL segment size from pg_control instead. This requires a deeper inspection of pg_control than has been done in the past, so it seemed best to copy the relevant data structures from each version of PostgreSQL and build a generic interface layer to address them. While this approach is a bit verbose, it has the advantage of being relatively simple, and can easily be updated for new versions of PostgreSQL. Since the integration tests generate pg_control files for testing, teach Perl how to generate files with the correct offsets for both 32-bit and 64-bit architectures.
2018-09-25 11:24:42 +02:00
P01 INFO: backup file [TEST_PATH]/db-master/db/base-2/global/pg_control (8KB, 95%) checksum 4c77c900f7af0d9ab13fa9982051a42e0b637f6c
Add more realistic data files to synthetic backup and restore tests.
2016-12-10 16:13:55 +02:00
P01 INFO: backup file [TEST_PATH]/db-master/db/base-2/base/1/12000 (8KB, 99%) checksum 22c98d248ff548311eda88559e4a8405ed77c003
For simplicity, the pg_control file is now copied with the rest of the files instead of by itself of at the end of the process. The backup command does not require this behavior and the restore copies to a temporary file which is renamed at the end of the restore.
2016-11-30 21:23:05 +02:00
P01 INFO: backup file [TEST_PATH]/db-master/db/base-2/postgresql.conf (21B, 99%) checksum 6721d92c9fcdf4248acff1f9a1377127d9064807
P01 INFO: backup file [TEST_PATH]/db-master/db/base-2/badchecksum.txt (11B, 99%) checksum f927212cd08d11a42a666b2f04235398e9ceeb51
P01 INFO: backup file [TEST_PATH]/db-master/db/base-2/base/16384/17000 (9B, 99%) checksum 7579ada0808d7f98087a0a586d0df9de009cdc33
Add checksum delta for incremental backups. Use checksums rather than timestamps to determine if files have changed. This is useful in cases where the timestamps may not be trustworthy, e.g. when performing an incremental after failing over to a standby. If checksum delta is enabled then checksums will be used for verification of resumed backups, even if they are full. Resumes have always used checksums to verify the files in the repository, enabling delta performs checksums on the database files as well. Note that the user must manually enable this feature in cases were it would be useful or just keep in enabled all the time. A future commit will address automatically enabling the feature in cases where it seems likely to be useful. Contributed by Cynthia Shang.
2018-09-19 17:12:45 +02:00
P01 INFO: backup file [TEST_PATH]/db-master/db/base-2/changecontent.txt (7B, 99%) checksum a094d94583e209556d03c3c5da33131a065f1689
For simplicity, the pg_control file is now copied with the rest of the files instead of by itself of at the end of the process. The backup command does not require this behavior and the restore copies to a temporary file which is renamed at the end of the restore.
2016-11-30 21:23:05 +02:00
P01 INFO: backup file [TEST_PATH]/db-master/db/base-2/pg_stat/global.stat (5B, 99%) checksum e350d5ce0153f3e22d5db21cf2a4eff00f3ee877
Add checksum delta for incremental backups. Use checksums rather than timestamps to determine if files have changed. This is useful in cases where the timestamps may not be trustworthy, e.g. when performing an incremental after failing over to a standby. If checksum delta is enabled then checksums will be used for verification of resumed backups, even if they are full. Resumes have always used checksums to verify the files in the repository, enabling delta performs checksums on the database files as well. Note that the user must manually enable this feature in cases were it would be useful or just keep in enabled all the time. A future commit will address automatically enabling the feature in cases where it seems likely to be useful. Contributed by Cynthia Shang.
2018-09-19 17:12:45 +02:00
P01 INFO: backup file [TEST_PATH]/db-master/db/base-2/changetime.txt (4B, 99%) checksum 88087292ed82e26f3eb824d0bffc05ccf7a30f8d
For simplicity, the pg_control file is now copied with the rest of the files instead of by itself of at the end of the process. The backup command does not require this behavior and the restore copies to a temporary file which is renamed at the end of the restore.
2016-11-30 21:23:05 +02:00
P01 INFO: backup file [TEST_PATH]/db-master/db/base-2/base/32768/PG_VERSION (3B, 99%) checksum 184473f470864e067ee3a22e64b47b0a1c356f29
P01 INFO: backup file [TEST_PATH]/db-master/db/base-2/base/16384/PG_VERSION (3B, 99%) checksum 184473f470864e067ee3a22e64b47b0a1c356f29
P01 INFO: backup file [TEST_PATH]/db-master/db/base-2/base/1/PG_VERSION (3B, 99%) checksum 184473f470864e067ee3a22e64b47b0a1c356f29
P01 INFO: backup file [TEST_PATH]/db-master/db/base-2/PG_VERSION (3B, 99%) checksum 184473f470864e067ee3a22e64b47b0a1c356f29
Add zero-length file to mock/all test.
2018-07-05 21:40:50 +02:00
P01 INFO: backup file [TEST_PATH]/db-master/db/base-2/zerosize.txt (0B, 99%)
Allow zero-size files in backup manifest to reference a prior manifest regardless of timestamp delta. Contributed by Cynthia Shang.
2018-08-24 22:50:33 +02:00
P01 INFO: backup file [TEST_PATH]/db-master/db/base-2/zero_from_start (0B, 99%)
Use minio for integration testing. ScalityS3 has not received any maintenance in years and is slow to start which is bad for testing. Replace it with minio which starts quickly and ships as a single executable or a tiny container. Minio has stricter limits on allowable characters but should still provide enough coverage to show that our encoding is working correctly. This commit also includes the upgrade to openssl 1.1.1 in the Ubuntu 18.04 container.
2019-07-03 04:20:35 +02:00
P01 INFO: backup file [TEST_PATH]/db-master/db/base-2/special-!_.*'()&!@;:+,? (0B, 99%)
For simplicity, the pg_control file is now copied with the rest of the files instead of by itself of at the end of the process. The backup command does not require this behavior and the restore copies to a temporary file which is renamed at the end of the restore.
2016-11-30 21:23:05 +02:00
P01 INFO: backup file [TEST_PATH]/db-master/db/base-2/pg_tblspc/2/[TS_PATH-1]/32768/tablespace2c.txt (12B, 99%) checksum dfcb8679956b734706cf87259d50c88f83e80e66
Additional backup exclusions. * Exclude contents of pg_snapshots, pg_serial, pg_notify, and pg_dynshmem from backup since they are rebuilt on startup. * Exclude pg_internal.init files from backup since they are rebuilt on startup.
2017-09-04 14:26:57 +02:00
P01 INFO: backup file [TEST_PATH]/db-master/db/base-2/pg_tblspc/2/[TS_PATH-1]/32768/tablespace2.txt (7B, 100%) checksum dc7f76e43c46101b47acc55ae4d593a9e6983578
Exclude temporary and unlogged relation (table/index) files from backup. Implemented using the same logic as the patches adding this feature to PostgreSQL, 8694cc96 and 920a5e50. Temporary relation exclusion is enabled in PostgreSQL ≥ 9.0. Unlogged relation exclusion is enabled in PostgreSQL ≥ 9.1, where the feature was introduced. Contributed by Cynthia Shang.
2018-07-31 00:53:34 +02:00
P00 INFO: full backup size = 176KB
Show Process ID in log output instead of filtering it out with the timestamp.
2016-09-14 23:30:57 +02:00
P00 INFO: new backup label = [BACKUP-FULL-3]
Added a predictable log entry to signal that a command has completed successfully. For example a backup ends successfully with: INFO: backup command end: completed successfully.
2016-12-04 21:44:53 +02:00
P00 INFO: backup command end: completed successfully
Integration test improvements for disk and memory efficiency. Set log-level-file=off when more that one test will run. In this case is it impossible to see the logs anyway since they will be automatically cleaned up after the test. This improves performance pretty dramatically since trace-level logging is expensive. If a singe integration test is run then log-level-file is trace by default but can be changed with the --log-level-test-file option. Reduce buffer-size to 64k to save memory during testing and allow more processes to run in parallel. Update log replacement rules so that these options can change without affecting expect logs.
2019-12-17 22:23:07 +02:00
P00 INFO: expire command begin [BACKREST-VERSION]: --buffer-size=[BUFFER-SIZE] --config=[TEST_PATH]/db-master/pgbackrest.conf --lock-path=[TEST_PATH]/db-master/lock --log-level-console=detail --log-level-file=[LOG-LEVEL-FILE] --log-level-stderr=off --log-path=[TEST_PATH]/db-master/log[] --no-log-timestamp --repo1-path=[TEST_PATH]/db-master/repo --stanza=db
Rename retention-* options to repo-retention-*.
2018-02-19 22:07:24 +02:00
P00 INFO: option 'repo1-retention-archive' is not set - archive logs will not be expired
Added a predictable log entry to signal that a command has completed successfully. For example a backup ends successfully with: INFO: backup command end: completed successfully.
2016-12-04 21:44:53 +02:00
P00 INFO: expire command end: completed successfully
Log testing can now be enabled for certain deterministic tests. This works by comparing the generated logs against a previous copy. Currently only enabled for the backup/synthetic tests.
2015-05-05 19:08:48 +02:00
Rename test paths for clarity. This was worked out as part of the test suite refactor [c8f806a] but not committed with it because of the large number of expect logs changes involved. Keeping them separate made it easier to audit the changes in the refactor.
2016-06-24 16:54:31 +02:00
+ supplemental file: [TEST_PATH]/db-master/pgbackrest.conf
----------------------------------------------------------
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
[db]
Rename db-* options to pg-* and backup-* options to repo-* to improve consistency. * repo-* options are now indexed although only one is allowed. * List deprecated option names in documentation and command-line help.
2018-02-04 01:27:38 +02:00
pg1-path=[TEST_PATH]/db-master/db/base-2
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
New simpler configuration and consistent project/exe/path naming. * The repo-path option now always refers to the repository where backups and archive are stored, whether local or remote, so the repo-remote-path option has been removed. The new spool-path option can be used to define a location for queueing WAL segments when archiving asynchronously. Otherwise, a local repository is no longer required. * Implemented a new config format which should be far simpler to use. See the User Guide and Configuration Reference for details but for a simple configuration all options can now be placed in the stanza section. Options that are shared between stanzas can be placed in the [global] section. More complex configurations can still make use of command sections though this should be a rare use case. * The default configuration filename is now pgbackrest.conf instead of pg_backrest.conf. This was done for consistency with other naming changes but also to prevent old config files from being loaded accidentally. * The default repository name was changed from /var/lib/backup to /var/lib/pgbackrest. * Lock files are now stored in /tmp/pgbackrest by default. These days /run/pgbackrest would be the preferred location but that would require init scripts which are not part of this release. The lock-path option can be used to configure the lock directory. * Log files are now stored in /var/log/pgbackrest by default and no longer have the date appended so they can be managed with logrotate. The log-path option can be used to configure the lock directory. * Executable filename changed from pg_backrest to pgbackrest.
2016-04-14 15:30:54 +02:00
[db:restore]
Rename test paths for clarity. This was worked out as part of the test suite refactor [c8f806a] but not committed with it because of the large number of expect logs changes involved. Keeping them separate made it easier to audit the changes in the refactor.
2016-06-24 16:54:31 +02:00
tablespace-map=1=[TEST_PATH]/db-master/db/tablespace/ts1-2
tablespace-map=2=[TEST_PATH]/db-master/db/tablespace/ts2-2
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
New simpler configuration and consistent project/exe/path naming. * The repo-path option now always refers to the repository where backups and archive are stored, whether local or remote, so the repo-remote-path option has been removed. The new spool-path option can be used to define a location for queueing WAL segments when archiving asynchronously. Otherwise, a local repository is no longer required. * Implemented a new config format which should be far simpler to use. See the User Guide and Configuration Reference for details but for a simple configuration all options can now be placed in the stanza section. Options that are shared between stanzas can be placed in the [global] section. More complex configurations can still make use of command sections though this should be a rare use case. * The default configuration filename is now pgbackrest.conf instead of pg_backrest.conf. This was done for consistency with other naming changes but also to prevent old config files from being loaded accidentally. * The default repository name was changed from /var/lib/backup to /var/lib/pgbackrest. * Lock files are now stored in /tmp/pgbackrest by default. These days /run/pgbackrest would be the preferred location but that would require init scripts which are not part of this release. The lock-path option can be used to configure the lock directory. * Log files are now stored in /var/log/pgbackrest by default and no longer have the date appended so they can be managed with logrotate. The log-path option can be used to configure the lock directory. * Executable filename changed from pg_backrest to pgbackrest.
2016-04-14 15:30:54 +02:00
[global]
Integration test improvements for disk and memory efficiency. Set log-level-file=off when more that one test will run. In this case is it impossible to see the logs anyway since they will be automatically cleaned up after the test. This improves performance pretty dramatically since trace-level logging is expensive. If a singe integration test is run then log-level-file is trace by default but can be changed with the --log-level-test-file option. Reduce buffer-size to 64k to save memory during testing and allow more processes to run in parallel. Update log replacement rules so that these options can change without affecting expect logs.
2019-12-17 22:23:07 +02:00
buffer-size=[BUFFER-SIZE]
Fix custom settings for compress-level option being ignored. Reported by Jens Wilke.
2017-10-23 00:05:46 +02:00
compress-level=3
Add Zstandard compression support. Zstandard is a fast lossless compression algorithm targeting real-time compression scenarios at zlib-level and better compression ratios. It's backed by a very fast entropy stage, provided by Huff0 and FSE library. Zstandard version >= 1.0 is required, which is generally only available on newer distributions.
2020-05-04 21:25:27 +02:00
compress-type=zst
Storage and IO layer refactor: Refactor storage layer to allow for new repository filesystems using drivers. (Reviewed by Cynthia Shang.) Refactor IO layer to allow for new compression formats, checksum types, and other capabilities using filters. (Reviewed by Cynthia Shang.)
2017-06-09 23:51:41 +02:00
db-timeout=45
lock-path=[TEST_PATH]/db-master/lock
Reduce log level for all expect tests to detail. The C code is designed to be efficient rather than deterministic at the debug log level. As we move more testing from integration to unit tests it makes less sense to try and maintain the expect logs at this log level. Most of the expect logs have already been moved to detail level but mock/all still had tests at debug level. Change the logging defaults in the config file and remove as many references to log-level-console as possible.
2019-05-23 00:23:44 +02:00
log-level-console=detail
Integration test improvements for disk and memory efficiency. Set log-level-file=off when more that one test will run. In this case is it impossible to see the logs anyway since they will be automatically cleaned up after the test. This improves performance pretty dramatically since trace-level logging is expensive. If a singe integration test is run then log-level-file is trace by default but can be changed with the --log-level-test-file option. Reduce buffer-size to 64k to save memory during testing and allow more processes to run in parallel. Update log replacement rules so that these options can change without affecting expect logs.
2019-12-17 22:23:07 +02:00
log-level-file=[LOG-LEVEL-FILE]
Added the log-level-stderr option. Controls whether console log messages are sent to stderr or stdout. By default this is set to warn which represents a change in behavior from previous versions, even though it may be more intuitive. Setting log-level-stderr=off will preserve the old behavior. Suggested by Sascha Biberhofer.
2016-10-05 15:09:30 +02:00
log-level-stderr=off
Storage and IO layer refactor: Refactor storage layer to allow for new repository filesystems using drivers. (Reviewed by Cynthia Shang.) Refactor IO layer to allow for new compression formats, checksum types, and other capabilities using filters. (Reviewed by Cynthia Shang.)
2017-06-09 23:51:41 +02:00
log-path=[TEST_PATH]/db-master/log
Integration test improvements for disk and memory efficiency. Set log-level-file=off when more that one test will run. In this case is it impossible to see the logs anyway since they will be automatically cleaned up after the test. This improves performance pretty dramatically since trace-level logging is expensive. If a singe integration test is run then log-level-file is trace by default but can be changed with the --log-level-test-file option. Reduce buffer-size to 64k to save memory during testing and allow more processes to run in parallel. Update log replacement rules so that these options can change without affecting expect logs.
2019-12-17 22:23:07 +02:00
log-subprocess=[LOG-SUBPROCESS]
Set log-timestamp=n for integration tests. This means less filtering of logs needs to be done and new timestamps can be added without adding new filters.
2018-05-11 17:24:38 +02:00
log-timestamp=n
Storage and IO layer refactor: Refactor storage layer to allow for new repository filesystems using drivers. (Reviewed by Cynthia Shang.) Refactor IO layer to allow for new compression formats, checksum types, and other capabilities using filters. (Reviewed by Cynthia Shang.)
2017-06-09 23:51:41 +02:00
protocol-timeout=60
Rename db-* options to pg-* and backup-* options to repo-* to improve consistency. * repo-* options are now indexed although only one is allowed. * List deprecated option names in documentation and command-line help.
2018-02-04 01:27:38 +02:00
repo1-hardlink=y
repo1-path=[TEST_PATH]/db-master/repo
Consolidate archive-push tests into new archive test.
2017-06-21 14:38:53 +02:00
spool-path=[TEST_PATH]/db-master/spool
New simpler configuration and consistent project/exe/path naming. * The repo-path option now always refers to the repository where backups and archive are stored, whether local or remote, so the repo-remote-path option has been removed. The new spool-path option can be used to define a location for queueing WAL segments when archiving asynchronously. Otherwise, a local repository is no longer required. * Implemented a new config format which should be far simpler to use. See the User Guide and Configuration Reference for details but for a simple configuration all options can now be placed in the stanza section. Options that are shared between stanzas can be placed in the [global] section. More complex configurations can still make use of command sections though this should be a rare use case. * The default configuration filename is now pgbackrest.conf instead of pg_backrest.conf. This was done for consistency with other naming changes but also to prevent old config files from being loaded accidentally. * The default repository name was changed from /var/lib/backup to /var/lib/pgbackrest. * Lock files are now stored in /tmp/pgbackrest by default. These days /run/pgbackrest would be the preferred location but that would require init scripts which are not part of this release. The lock-path option can be used to configure the lock directory. * Log files are now stored in /var/log/pgbackrest by default and no longer have the date appended so they can be managed with logrotate. The log-path option can be used to configure the lock directory. * Executable filename changed from pg_backrest to pgbackrest.
2016-04-14 15:30:54 +02:00
[global:backup]
archive-copy=y
start-fast=y
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
Rename test paths for clarity. This was worked out as part of the test suite refactor [c8f806a] but not committed with it because of the large number of expect logs changes involved. Keeping them separate made it easier to audit the changes in the refactor.
2016-06-24 16:54:31 +02:00
+ supplemental file: [TEST_PATH]/db-master/repo/backup/db/[BACKUP-FULL-3]/backup.manifest
-----------------------------------------------------------------------------------------
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
[backrest]
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
backrest-format=5
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
backrest-version="[VERSION-1]"
[backup]
backup-label="[BACKUP-FULL-3]"
backup-timestamp-copy-start=[TIMESTAMP]
backup-timestamp-start=[TIMESTAMP]
backup-timestamp-stop=[TIMESTAMP]
backup-type="full"
[backup:db]
Database version refactoring. * Refactor db version constants into a separate module. * Update synthetic backup tests to PostgreSQL 9.4.
2016-08-12 04:35:24 +02:00
db-catalog-version=201409291
db-control-version=942
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
db-id=1
Automate generation of WAL and pg_control test files. The existing static files would not work with 32-bit or big-endian systems so create functions to generate these files dynamically rather than creating a bunch of new static files.
2017-11-19 03:02:54 +02:00
db-system-id=1000000000000000094
Database version refactoring. * Refactor db version constants into a separate module. * Update synthetic backup tests to PostgreSQL 9.4.
2016-08-12 04:35:24 +02:00
db-version="9.4"
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
[backup:option]
Set archive-check option in manifest correctly when offline. Archive check does not run when in offline backup mode but the option was set to true in the manifest. It's harmless since these options are informational only but it could cause confusion when debugging.
2019-11-28 15:27:21 +02:00
option-archive-check=false
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
option-archive-copy=true
Backup from a standby cluster. A connection to the primary cluster is still required to start/stop the backup and copy files that are not replicated, but the vast majority of files are copied from the standby in order to reduce load on the master.
2016-08-25 17:25:46 +02:00
option-backup-standby=false
Integration test improvements for disk and memory efficiency. Set log-level-file=off when more that one test will run. In this case is it impossible to see the logs anyway since they will be automatically cleaned up after the test. This improves performance pretty dramatically since trace-level logging is expensive. If a singe integration test is run then log-level-file is trace by default but can be changed with the --log-level-test-file option. Reduce buffer-size to 64k to save memory during testing and allow more processes to run in parallel. Update log replacement rules so that these options can change without affecting expect logs.
2019-12-17 22:23:07 +02:00
option-buffer-size=[BUFFER-SIZE]
Added the --checksum-page option. This option allows pgBackRest to validate page checksums in data files when checksums are enabled on PostgreSQL >= 9.3. Note that this functionality requires a C library which may not initially be available in OS packages. The option will automatically be enabled when the library is present and checksums are enabled on the cluster.
2016-12-13 01:54:07 +02:00
option-checksum-page=false
Full/Synthetic test refactor. * Combine hardlink and non/compressed in synthetic tests to reduce test time and improve coverage. * Change log level of hardlink logging to detail. * Cast size in S3 manifest to integer.
2017-06-15 21:32:10 +02:00
option-compress=true
Add additional options to backup.manifest for debugging purposes. Add the buffer-size, compress-level, compress-level-network, and process-max options to the backup:option section in backup.manifest to aid in debugging. It may also make sense to propagate these options up to backup.info so they can be displayed in the info command, but for now this is deemed sufficient. Contributed by blogh.
2019-03-10 11:03:52 +02:00
option-compress-level=3
option-compress-level-network=3
Add Zstandard compression support. Zstandard is a fast lossless compression algorithm targeting real-time compression scenarios at zlib-level and better compression ratios. It's backed by a very fast entropy stage, provided by Huff0 and FSE library. Zstandard version >= 1.0 is required, which is generally only available on newer distributions.
2020-05-04 21:25:27 +02:00
option-compress-type="zst"
Add checksum delta for incremental backups. Use checksums rather than timestamps to determine if files have changed. This is useful in cases where the timestamps may not be trustworthy, e.g. when performing an incremental after failing over to a standby. If checksum delta is enabled then checksums will be used for verification of resumed backups, even if they are full. Resumes have always used checksums to verify the files in the repository, enabling delta performs checksums on the database files as well. Note that the user must manually enable this feature in cases were it would be useful or just keep in enabled all the time. A future commit will address automatically enabling the feature in cases where it seems likely to be useful. Contributed by Cynthia Shang.
2018-09-19 17:12:45 +02:00
option-delta=false
Full/Synthetic test refactor. * Combine hardlink and non/compressed in synthetic tests to reduce test time and improve coverage. * Change log level of hardlink logging to detail. * Cast size in S3 manifest to integer.
2017-06-15 21:32:10 +02:00
option-hardlink=true
Closed #167: Rename `--no-start-stop` option to `--no-online`.
2016-02-20 02:32:02 +02:00
option-online=false
Add additional options to backup.manifest for debugging purposes. Add the buffer-size, compress-level, compress-level-network, and process-max options to the backup:option section in backup.manifest to aid in debugging. It may also make sense to propagate these options up to backup.info so they can be displayed in the info command, but for now this is deemed sufficient. Contributed by blogh.
2019-03-10 11:03:52 +02:00
option-process-max=1
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
[backup:target]
Rename test paths for clarity. This was worked out as part of the test suite refactor [c8f806a] but not committed with it because of the large number of expect logs changes involved. Keeping them separate made it easier to audit the changes in the refactor.
2016-06-24 16:54:31 +02:00
pg_data={"path":"[TEST_PATH]/db-master/db/base-2","type":"path"}
pg_tblspc/2={"path":"[TEST_PATH]/db-master/db/tablespace/ts2-2","tablespace-id":"2","tablespace-name":"ts2","type":"link"}
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
[target:file]
Full/Synthetic test refactor. * Combine hardlink and non/compressed in synthetic tests to reduce test time and improve coverage. * Change log level of hardlink logging to detail. * Cast size in S3 manifest to integer.
2017-06-15 21:32:10 +02:00
pg_data/PG_VERSION={"checksum":"184473f470864e067ee3a22e64b47b0a1c356f29","master":true,"repo-size":[SIZE],"size":[SIZE],"timestamp":[TIMESTAMP-1]}
pg_data/badchecksum.txt={"checksum":"f927212cd08d11a42a666b2f04235398e9ceeb51","master":true,"repo-size":[SIZE],"size":[SIZE],"timestamp":[TIMESTAMP-1]}
pg_data/base/1/12000={"checksum":"22c98d248ff548311eda88559e4a8405ed77c003","repo-size":[SIZE],"size":[SIZE],"timestamp":[TIMESTAMP-1]}
pg_data/base/1/PG_VERSION={"checksum":"184473f470864e067ee3a22e64b47b0a1c356f29","mode":"0660","repo-size":[SIZE],"size":[SIZE],"timestamp":[TIMESTAMP-1]}
pg_data/base/16384/17000={"checksum":"7579ada0808d7f98087a0a586d0df9de009cdc33","repo-size":[SIZE],"size":[SIZE],"timestamp":[TIMESTAMP-1]}
pg_data/base/16384/PG_VERSION={"checksum":"184473f470864e067ee3a22e64b47b0a1c356f29","repo-size":[SIZE],"size":[SIZE],"timestamp":[TIMESTAMP-1]}
Fix incorrect offline upper bound for ignoring page checksum errors. For offline backups the upper bound was being set to 0x0000FFFF0000FFFF rather than UINT64_MAX. This meant that page checksum errors might be ignored for databases with a lot of past WAL in offline mode. Online mode is not affected since the upper bound is retrieved from pg_start_backup().
2019-07-11 15:13:56 +02:00
pg_data/base/32768/33000={"checksum":"7a16d165e4775f7c92e8cdf60c0af57313f0bf90","repo-size":[SIZE],"size":[SIZE],"timestamp":[TIMESTAMP-1]}
pg_data/base/32768/33000.32767={"checksum":"6e99b589e550e68e934fd235ccba59fe5b592a9e","repo-size":[SIZE],"size":[SIZE],"timestamp":[TIMESTAMP-1]}
Full/Synthetic test refactor. * Combine hardlink and non/compressed in synthetic tests to reduce test time and improve coverage. * Change log level of hardlink logging to detail. * Cast size in S3 manifest to integer.
2017-06-15 21:32:10 +02:00
pg_data/base/32768/33001={"checksum":"6bf316f11d28c28914ea9be92c00de9bea6d9a6b","repo-size":[SIZE],"size":[SIZE],"timestamp":[TIMESTAMP-1]}
Fix incorrect offline upper bound for ignoring page checksum errors. For offline backups the upper bound was being set to 0x0000FFFF0000FFFF rather than UINT64_MAX. This meant that page checksum errors might be ignored for databases with a lot of past WAL in offline mode. Online mode is not affected since the upper bound is retrieved from pg_start_backup().
2019-07-11 15:13:56 +02:00
pg_data/base/32768/44000_init={"checksum":"7a16d165e4775f7c92e8cdf60c0af57313f0bf90","repo-size":[SIZE],"size":[SIZE],"timestamp":[TIMESTAMP-1]}
Full/Synthetic test refactor. * Combine hardlink and non/compressed in synthetic tests to reduce test time and improve coverage. * Change log level of hardlink logging to detail. * Cast size in S3 manifest to integer.
2017-06-15 21:32:10 +02:00
pg_data/base/32768/PG_VERSION={"checksum":"184473f470864e067ee3a22e64b47b0a1c356f29","repo-size":[SIZE],"size":[SIZE],"timestamp":[TIMESTAMP-1]}
Add checksum delta for incremental backups. Use checksums rather than timestamps to determine if files have changed. This is useful in cases where the timestamps may not be trustworthy, e.g. when performing an incremental after failing over to a standby. If checksum delta is enabled then checksums will be used for verification of resumed backups, even if they are full. Resumes have always used checksums to verify the files in the repository, enabling delta performs checksums on the database files as well. Note that the user must manually enable this feature in cases were it would be useful or just keep in enabled all the time. A future commit will address automatically enabling the feature in cases where it seems likely to be useful. Contributed by Cynthia Shang.
2018-09-19 17:12:45 +02:00
pg_data/changecontent.txt={"checksum":"a094d94583e209556d03c3c5da33131a065f1689","master":true,"repo-size":[SIZE],"size":[SIZE],"timestamp":[TIMESTAMP-1]}
pg_data/changetime.txt={"checksum":"88087292ed82e26f3eb824d0bffc05ccf7a30f8d","master":true,"repo-size":[SIZE],"size":[SIZE],"timestamp":[TIMESTAMP-2]}
Support configurable WAL segment size. PostgreSQL 11 introduces configurable WAL segment sizes, from 1MB to 1GB. There are two areas that needed to be updated to support this: building the archive-get queue and checking that WAL has been archived after a backup. Both operations require the WAL segment size to properly build a list. Checking the archive after a backup is still implemented in Perl and has an active database connection, so just get the WAL segment size from the database. The archive-get command does not have a connection to the database, so get the WAL segment size from pg_control instead. This requires a deeper inspection of pg_control than has been done in the past, so it seemed best to copy the relevant data structures from each version of PostgreSQL and build a generic interface layer to address them. While this approach is a bit verbose, it has the advantage of being relatively simple, and can easily be updated for new versions of PostgreSQL. Since the integration tests generate pg_control files for testing, teach Perl how to generate files with the correct offsets for both 32-bit and 64-bit architectures.
2018-09-25 11:24:42 +02:00
pg_data/global/pg_control={"checksum":"4c77c900f7af0d9ab13fa9982051a42e0b637f6c","master":true,"repo-size":[SIZE],"size":[SIZE],"timestamp":[TIMESTAMP-2]}
Full/Synthetic test refactor. * Combine hardlink and non/compressed in synthetic tests to reduce test time and improve coverage. * Change log level of hardlink logging to detail. * Cast size in S3 manifest to integer.
2017-06-15 21:32:10 +02:00
pg_data/pg_stat/global.stat={"checksum":"e350d5ce0153f3e22d5db21cf2a4eff00f3ee877","master":true,"repo-size":[SIZE],"size":[SIZE],"timestamp":[TIMESTAMP-2]}
pg_data/postgresql.conf={"checksum":"6721d92c9fcdf4248acff1f9a1377127d9064807","master":true,"repo-size":[SIZE],"size":[SIZE],"timestamp":[TIMESTAMP-2]}
Use minio for integration testing. ScalityS3 has not received any maintenance in years and is slow to start which is bad for testing. Replace it with minio which starts quickly and ships as a single executable or a tiny container. Minio has stricter limits on allowable characters but should still provide enough coverage to show that our encoding is working correctly. This commit also includes the upgrade to openssl 1.1.1 in the Ubuntu 18.04 container.
2019-07-03 04:20:35 +02:00
pg_data/special-!_.*'()&!@;:+,?={"master":true,"repo-size":[SIZE],"size":[SIZE],"timestamp":[TIMESTAMP-1]}
Allow zero-size files in backup manifest to reference a prior manifest regardless of timestamp delta. Contributed by Cynthia Shang.
2018-08-24 22:50:33 +02:00
pg_data/zero_from_start={"master":true,"repo-size":[SIZE],"size":[SIZE],"timestamp":[TIMESTAMP-1]}
Add zero-length file to mock/all test.
2018-07-05 21:40:50 +02:00
pg_data/zerosize.txt={"master":true,"repo-size":[SIZE],"size":[SIZE],"timestamp":[TIMESTAMP-2]}
Full/Synthetic test refactor. * Combine hardlink and non/compressed in synthetic tests to reduce test time and improve coverage. * Change log level of hardlink logging to detail. * Cast size in S3 manifest to integer.
2017-06-15 21:32:10 +02:00
pg_tblspc/2/[TS_PATH-1]/32768/tablespace2.txt={"checksum":"dc7f76e43c46101b47acc55ae4d593a9e6983578","repo-size":[SIZE],"size":[SIZE],"timestamp":[TIMESTAMP-1]}
pg_tblspc/2/[TS_PATH-1]/32768/tablespace2c.txt={"checksum":"dfcb8679956b734706cf87259d50c88f83e80e66","repo-size":[SIZE],"size":[SIZE],"timestamp":[TIMESTAMP-1]}
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
[target:file:default]
Continuous integration using travis-ci.
2017-02-21 15:59:23 +02:00
group="[GROUP-1]"
Fixed a issue where tablespaces were copied from the master during standby backup.
2016-09-04 15:19:44 +02:00
master=false
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
mode="0600"
Regression tests can now be run as any properly-configured user, not just vagrant.
2016-10-01 19:39:44 +02:00
user="[USER-1]"
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
[target:link]
Rename test paths for clarity. This was worked out as part of the test suite refactor [c8f806a] but not committed with it because of the large number of expect logs changes involved. Keeping them separate made it easier to audit the changes in the refactor.
2016-06-24 16:54:31 +02:00
pg_data/pg_tblspc/2={"destination":"[TEST_PATH]/db-master/db/tablespace/ts2-2"}
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
[target:link:default]
Continuous integration using travis-ci.
2017-02-21 15:59:23 +02:00
group="[GROUP-1]"
Regression tests can now be run as any properly-configured user, not just vagrant.
2016-10-01 19:39:44 +02:00
user="[USER-1]"
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
[target:path]
The restore command is implemented entirely in C. For the most part this is a direct migration of the Perl code into C. There is one important behavioral change with regard to how file permissions are handled. The Perl code tried to set ownership as it was in the manifest even when running as an unprivileged user. This usually just led to errors and frustration. The C code works like this: If a restore is run as a non-root user (the typical scenario) then all files restored will belong to the user/group executing pgBackRest. If existing files are not owned by the executing user/group then an error will result if the ownership cannot be updated to the executing user/group. In that case the file ownership will need to be updated by a privileged user before the restore can be retried. If a restore is run as the root user then pgBackRest will attempt to recreate the ownership recorded in the manifest when the backup was made. Only user/group names are stored in the manifest so the same names must exist on the restore host for this to work. If the user/group name cannot be found locally then the user/group of the PostgreSQL data directory will be used and finally root if the data directory user/group cannot be mapped to a name. Reviewed by Cynthia Shang.
2019-09-26 13:52:02 +02:00
pg_data={"mode":"0777"}
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
pg_data/base={}
Allow selective restore of databases from a cluster backup. This feature can result in major space and time savings when only specific databases are restored. Unrestored databases will not be accessible but must be manually dropped before they will be removed from the shared catalogue.
2016-05-11 15:21:39 +02:00
pg_data/base/1={}
pg_data/base/16384={}
pg_data/base/32768={}
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
pg_data/global={}
pg_data/pg_clog={}
Exclude directories during backup that are cleaned, recreated, or zeroed by PostgreSQL at startup. These include (depending on the version where they were introduced): pgsql_tmp, pg_dynshmem, pg_notify, pg_replslot, pg_serial, pg_snapshots, pg_stat_tmp, pg_subtrans. The postgresql.auto.conf.tmp file is now excluded in addition to files that were already excluded: backup_label.old, postmaster.opts, postmaster.pid, recovery.conf, recovery.done.
2016-08-16 15:35:16 +02:00
pg_data/pg_dynshmem={}
Allow arbitrary directories and/or files to be excluded from a backup. Misuse of this feature can lead to inconsistent backups so read the --exclude documentation carefully before using.
2018-08-27 21:51:05 +02:00
pg_data/pg_log={}
Exclude directories during backup that are cleaned, recreated, or zeroed by PostgreSQL at startup. These include (depending on the version where they were introduced): pgsql_tmp, pg_dynshmem, pg_notify, pg_replslot, pg_serial, pg_snapshots, pg_stat_tmp, pg_subtrans. The postgresql.auto.conf.tmp file is now excluded in addition to files that were already excluded: backup_label.old, postmaster.opts, postmaster.pid, recovery.conf, recovery.done.
2016-08-16 15:35:16 +02:00
pg_data/pg_notify={}
pg_data/pg_replslot={}
pg_data/pg_serial={}
pg_data/pg_snapshots={}
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
pg_data/pg_stat={}
Exclude directories during backup that are cleaned, recreated, or zeroed by PostgreSQL at startup. These include (depending on the version where they were introduced): pgsql_tmp, pg_dynshmem, pg_notify, pg_replslot, pg_serial, pg_snapshots, pg_stat_tmp, pg_subtrans. The postgresql.auto.conf.tmp file is now excluded in addition to files that were already excluded: backup_label.old, postmaster.opts, postmaster.pid, recovery.conf, recovery.done.
2016-08-16 15:35:16 +02:00
pg_data/pg_stat_tmp={}
pg_data/pg_subtrans={}
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
pg_data/pg_tblspc={}
The restore command is implemented entirely in C. For the most part this is a direct migration of the Perl code into C. There is one important behavioral change with regard to how file permissions are handled. The Perl code tried to set ownership as it was in the manifest even when running as an unprivileged user. This usually just led to errors and frustration. The C code works like this: If a restore is run as a non-root user (the typical scenario) then all files restored will belong to the user/group executing pgBackRest. If existing files are not owned by the executing user/group then an error will result if the ownership cannot be updated to the executing user/group. In that case the file ownership will need to be updated by a privileged user before the restore can be retried. If a restore is run as the root user then pgBackRest will attempt to recreate the ownership recorded in the manifest when the backup was made. Only user/group names are stored in the manifest so the same names must exist on the restore host for this to work. If the user/group name cannot be found locally then the user/group of the PostgreSQL data directory will be used and finally root if the data directory user/group cannot be mapped to a name. Reviewed by Cynthia Shang.
2019-09-26 13:52:02 +02:00
pg_tblspc={"mode":"0777"}
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
pg_tblspc/2={}
pg_tblspc/2/[TS_PATH-1]={}
Allow selective restore of databases from a cluster backup. This feature can result in major space and time savings when only specific databases are restored. Unrestored databases will not be accessible but must be manually dropped before they will be removed from the shared catalogue.
2016-05-11 15:21:39 +02:00
pg_tblspc/2/[TS_PATH-1]/32768={}
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
[target:path:default]
Continuous integration using travis-ci.
2017-02-21 15:59:23 +02:00
group="[GROUP-1]"
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
mode="0700"
Regression tests can now be run as any properly-configured user, not just vagrant.
2016-10-01 19:39:44 +02:00
user="[USER-1]"
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
Move info file checksum to the end of the file. Putting the checksum at the beginning of the file made it impossible to stream the file out when saving. The entire file had to be held in memory while it was checksummed so the checksum could be written at the beginning. Instead place the checksum at the end. This does not break the existing Perl or C code since the read is not order dependent. There are no plans to improve the Perl code to take advantage of this change, but it will make the C implementation more efficient. Reviewed by Cynthia Shang.
2019-08-22 01:45:48 +02:00
[backrest]
backrest-checksum="[CHECKSUM]"
Rename test paths for clarity. This was worked out as part of the test suite refactor [c8f806a] but not committed with it because of the large number of expect logs changes involved. Keeping them separate made it easier to audit the changes in the refactor.
2016-06-24 16:54:31 +02:00
+ supplemental file: [TEST_PATH]/db-master/repo/backup/db/backup.info
---------------------------------------------------------------------
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
[backrest]
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
backrest-format=5
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
backrest-version="[VERSION-1]"
[backup:current]
Set archive-check option in manifest correctly when offline. Archive check does not run when in offline backup mode but the option was set to true in the manifest. It's harmless since these options are informational only but it could cause confusion when debugging.
2019-11-28 15:27:21 +02:00
[BACKUP-FULL-2]={"backrest-format":5,"backrest-version":"[VERSION-1]","backup-archive-start":null,"backup-archive-stop":null,"backup-info-repo-size":[SIZE],"backup-info-repo-size-delta":[DELTA],"backup-info-size":[SIZE],"backup-info-size-delta":[DELTA],"backup-timestamp-start":[TIMESTAMP],"backup-timestamp-stop":[TIMESTAMP],"backup-type":"full","db-id":1,"option-archive-check":false,"option-archive-copy":true,"option-backup-standby":false,"option-checksum-page":true,"option-compress":false,"option-hardlink":false,"option-online":false}
Remove integration tests that are now covered in the unit tests. Most of these tests are just checking that errors are thrown when required. These are well covered in various unit tests. The "cannot resume" tests are also well covered in the backup unit tests. Finally, config warnings are well covered in the config unit tests. There is more to be done here, but this accounts for the low-hanging fruit.
2019-12-18 03:14:45 +02:00
[BACKUP-INCR-2]={"backrest-format":5,"backrest-version":"[VERSION-1]","backup-archive-start":null,"backup-archive-stop":null,"backup-info-repo-size":[SIZE],"backup-info-repo-size-delta":[DELTA],"backup-info-size":[SIZE],"backup-info-size-delta":[DELTA],"backup-prior":"[BACKUP-FULL-2]","backup-reference":["[BACKUP-FULL-2]"],"backup-timestamp-start":[TIMESTAMP],"backup-timestamp-stop":[TIMESTAMP],"backup-type":"incr","db-id":1,"option-archive-check":false,"option-archive-copy":true,"option-backup-standby":false,"option-checksum-page":true,"option-compress":false,"option-hardlink":false,"option-online":false}
[BACKUP-DIFF-1]={"backrest-format":5,"backrest-version":"[VERSION-1]","backup-archive-start":null,"backup-archive-stop":null,"backup-info-repo-size":[SIZE],"backup-info-repo-size-delta":[DELTA],"backup-info-size":[SIZE],"backup-info-size-delta":[DELTA],"backup-prior":"[BACKUP-FULL-2]","backup-reference":["[BACKUP-FULL-2]"],"backup-timestamp-start":[TIMESTAMP],"backup-timestamp-stop":[TIMESTAMP],"backup-type":"diff","db-id":1,"option-archive-check":false,"option-archive-copy":true,"option-backup-standby":false,"option-checksum-page":true,"option-compress":false,"option-hardlink":false,"option-online":false}
[BACKUP-INCR-3]={"backrest-format":5,"backrest-version":"[VERSION-1]","backup-archive-start":null,"backup-archive-stop":null,"backup-info-repo-size":[SIZE],"backup-info-repo-size-delta":[DELTA],"backup-info-size":[SIZE],"backup-info-size-delta":[DELTA],"backup-prior":"[BACKUP-DIFF-1]","backup-reference":["[BACKUP-FULL-2]","[BACKUP-DIFF-1]"],"backup-timestamp-start":[TIMESTAMP],"backup-timestamp-stop":[TIMESTAMP],"backup-type":"incr","db-id":1,"option-archive-check":false,"option-archive-copy":true,"option-backup-standby":false,"option-checksum-page":true,"option-compress":false,"option-hardlink":false,"option-online":false}
[BACKUP-INCR-4]={"backrest-format":5,"backrest-version":"[VERSION-1]","backup-archive-start":null,"backup-archive-stop":null,"backup-info-repo-size":[SIZE],"backup-info-repo-size-delta":[DELTA],"backup-info-size":[SIZE],"backup-info-size-delta":[DELTA],"backup-prior":"[BACKUP-INCR-3]","backup-reference":["[BACKUP-FULL-2]","[BACKUP-DIFF-1]","[BACKUP-INCR-3]"],"backup-timestamp-start":[TIMESTAMP],"backup-timestamp-stop":[TIMESTAMP],"backup-type":"incr","db-id":1,"option-archive-check":false,"option-archive-copy":true,"option-backup-standby":false,"option-checksum-page":true,"option-compress":false,"option-hardlink":false,"option-online":false}
Set archive-check option in manifest correctly when offline. Archive check does not run when in offline backup mode but the option was set to true in the manifest. It's harmless since these options are informational only but it could cause confusion when debugging.
2019-11-28 15:27:21 +02:00
[BACKUP-DIFF-2]={"backrest-format":5,"backrest-version":"[VERSION-1]","backup-archive-start":null,"backup-archive-stop":null,"backup-info-repo-size":[SIZE],"backup-info-repo-size-delta":[DELTA],"backup-info-size":[SIZE],"backup-info-size-delta":[DELTA],"backup-prior":"[BACKUP-FULL-2]","backup-reference":["[BACKUP-FULL-2]"],"backup-timestamp-start":[TIMESTAMP],"backup-timestamp-stop":[TIMESTAMP],"backup-type":"diff","db-id":1,"option-archive-check":false,"option-archive-copy":true,"option-backup-standby":false,"option-checksum-page":true,"option-compress":false,"option-hardlink":false,"option-online":false}
[BACKUP-DIFF-3]={"backrest-format":5,"backrest-version":"[VERSION-1]","backup-archive-start":null,"backup-archive-stop":null,"backup-info-repo-size":[SIZE],"backup-info-repo-size-delta":[DELTA],"backup-info-size":[SIZE],"backup-info-size-delta":[DELTA],"backup-prior":"[BACKUP-FULL-2]","backup-reference":["[BACKUP-FULL-2]"],"backup-timestamp-start":[TIMESTAMP],"backup-timestamp-stop":[TIMESTAMP],"backup-type":"diff","db-id":1,"option-archive-check":false,"option-archive-copy":true,"option-backup-standby":false,"option-checksum-page":true,"option-compress":false,"option-hardlink":false,"option-online":false}
[BACKUP-FULL-3]={"backrest-format":5,"backrest-version":"[VERSION-1]","backup-archive-start":null,"backup-archive-stop":null,"backup-info-repo-size":[SIZE],"backup-info-repo-size-delta":[DELTA],"backup-info-size":[SIZE],"backup-info-size-delta":[DELTA],"backup-timestamp-start":[TIMESTAMP],"backup-timestamp-stop":[TIMESTAMP],"backup-type":"full","db-id":1,"option-archive-check":false,"option-archive-copy":true,"option-backup-standby":false,"option-checksum-page":false,"option-compress":true,"option-hardlink":true,"option-online":false}
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
[db]
Database version refactoring. * Refactor db version constants into a separate module. * Update synthetic backup tests to PostgreSQL 9.4.
2016-08-12 04:35:24 +02:00
db-catalog-version=201409291
db-control-version=942
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
db-id=1
Automate generation of WAL and pg_control test files. The existing static files would not work with 32-bit or big-endian systems so create functions to generate these files dynamically rather than creating a bunch of new static files.
2017-11-19 03:02:54 +02:00
db-system-id=1000000000000000094
Database version refactoring. * Refactor db version constants into a separate module. * Update synthetic backup tests to PostgreSQL 9.4.
2016-08-12 04:35:24 +02:00
db-version="9.4"
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
[db:history]
Automate generation of WAL and pg_control test files. The existing static files would not work with 32-bit or big-endian systems so create functions to generate these files dynamically rather than creating a bunch of new static files.
2017-11-19 03:02:54 +02:00
1={"db-catalog-version":201409291,"db-control-version":942,"db-system-id":1000000000000000094,"db-version":"9.4"}
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
Improve performance of info file load/save. Info files required three copies in memory to be loaded (the original string, an ini representation, and the final info object). Not only was this memory inefficient but the Ini object does sequential scans when searching for keys making large files very slow to load. This has not been an issue since archive.info and backup.info are very small, but it becomes a big deal when loading manifests with hundreds of thousands of files. Instead of holding copies of the data in memory, use a callback to deliver the ini data directly to the object when loading. Use a similar method for save to avoid having an intermediate copy. Save is a bit complex because sections/keys must be written in alpha order or older versions of pgBackRest will not calculate the correct checksum. Also move the load retry logic to helper functions rather than embedding it in the Info object. This allows for more flexibility in loading and ensures that stack traces will be available when developing unit tests. Reviewed by Cynthia Shang.
2019-09-06 19:48:28 +02:00
[backrest]
backrest-checksum="[CHECKSUM]"
Major refactor of the test suite. * Make the code more modular and object-oriented. * Multiple Docker containers can now be created for a single test to simulate more realistic environments.
2016-06-24 14:12:58 +02:00
expire full=1 (db-master host)
Reduce log level for all expect tests to detail. The C code is designed to be efficient rather than deterministic at the debug log level. As we move more testing from integration to unit tests it makes less sense to try and maintain the expect logs at this log level. Most of the expect logs have already been moved to detail level but mock/all still had tests at debug level. Change the logging defaults in the config file and remove as many references to log-level-console as possible.
2019-05-23 00:23:44 +02:00
> [CONTAINER-EXEC] db-master [BACKREST-BIN] --config=[TEST_PATH]/db-master/pgbackrest.conf --repo1-retention-full=1 --stanza=db expire
Closed #131: Offline backups make expire command fail.
2016-01-16 06:37:17 +02:00
------------------------------------------------------------------------------------------------------------------------------------
Integration test improvements for disk and memory efficiency. Set log-level-file=off when more that one test will run. In this case is it impossible to see the logs anyway since they will be automatically cleaned up after the test. This improves performance pretty dramatically since trace-level logging is expensive. If a singe integration test is run then log-level-file is trace by default but can be changed with the --log-level-test-file option. Reduce buffer-size to 64k to save memory during testing and allow more processes to run in parallel. Update log replacement rules so that these options can change without affecting expect logs.
2019-12-17 22:23:07 +02:00
P00 INFO: expire command begin [BACKREST-VERSION]: --buffer-size=[BUFFER-SIZE] --config=[TEST_PATH]/db-master/pgbackrest.conf --lock-path=[TEST_PATH]/db-master/lock --log-level-console=detail --log-level-file=[LOG-LEVEL-FILE] --log-level-stderr=off --log-path=[TEST_PATH]/db-master/log[] --no-log-timestamp --repo1-path=[TEST_PATH]/db-master/repo --repo1-retention-full=1 --stanza=db
Remove integration tests that are now covered in the unit tests. Most of these tests are just checking that errors are thrown when required. These are well covered in various unit tests. The "cannot resume" tests are also well covered in the backup unit tests. Finally, config warnings are well covered in the config unit tests. There is more to be done here, but this accounts for the low-hanging fruit.
2019-12-18 03:14:45 +02:00
P00 INFO: expire full backup set: [BACKUP-FULL-2], [BACKUP-INCR-2], [BACKUP-DIFF-1], [BACKUP-INCR-3], [BACKUP-INCR-4], [BACKUP-DIFF-2], [BACKUP-DIFF-3]
Show Process ID in log output instead of filtering it out with the timestamp.
2016-09-14 23:30:57 +02:00
P00 INFO: remove expired backup [BACKUP-DIFF-3]
Remove integration tests that are now covered in the unit tests. Most of these tests are just checking that errors are thrown when required. These are well covered in various unit tests. The "cannot resume" tests are also well covered in the backup unit tests. Finally, config warnings are well covered in the config unit tests. There is more to be done here, but this accounts for the low-hanging fruit.
2019-12-18 03:14:45 +02:00
P00 INFO: remove expired backup [BACKUP-DIFF-2]
Show Process ID in log output instead of filtering it out with the timestamp.
2016-09-14 23:30:57 +02:00
P00 INFO: remove expired backup [BACKUP-INCR-4]
P00 INFO: remove expired backup [BACKUP-INCR-3]
Remove integration tests that are now covered in the unit tests. Most of these tests are just checking that errors are thrown when required. These are well covered in various unit tests. The "cannot resume" tests are also well covered in the backup unit tests. Finally, config warnings are well covered in the config unit tests. There is more to be done here, but this accounts for the low-hanging fruit.
2019-12-18 03:14:45 +02:00
P00 INFO: remove expired backup [BACKUP-DIFF-1]
P00 INFO: remove expired backup [BACKUP-INCR-2]
Show Process ID in log output instead of filtering it out with the timestamp.
2016-09-14 23:30:57 +02:00
P00 INFO: remove expired backup [BACKUP-FULL-2]
Added a predictable log entry to signal that a command has completed successfully. For example a backup ends successfully with: INFO: backup command end: completed successfully.
2016-12-04 21:44:53 +02:00
P00 INFO: expire command end: completed successfully
Closed #131: Offline backups make expire command fail.
2016-01-16 06:37:17 +02:00
Major refactor of the test suite. * Make the code more modular and object-oriented. * Multiple Docker containers can now be created for a single test to simulate more realistic environments.
2016-06-24 14:12:58 +02:00
diff backup - add file (db-master host)
Reduce log level for all expect tests to detail. The C code is designed to be efficient rather than deterministic at the debug log level. As we move more testing from integration to unit tests it makes less sense to try and maintain the expect logs at this log level. Most of the expect logs have already been moved to detail level but mock/all still had tests at debug level. Change the logging defaults in the config file and remove as many references to log-level-console as possible.
2019-05-23 00:23:44 +02:00
> [CONTAINER-EXEC] db-master [BACKREST-BIN] --config=[TEST_PATH]/db-master/pgbackrest.conf --no-online --checksum-page --type=diff --stanza=db backup
Log testing can now be enabled for certain deterministic tests. This works by comparing the generated logs against a previous copy. Currently only enabled for the backup/synthetic tests.
2015-05-05 19:08:48 +02:00
------------------------------------------------------------------------------------------------------------------------------------
Add Zstandard compression support. Zstandard is a fast lossless compression algorithm targeting real-time compression scenarios at zlib-level and better compression ratios. It's backed by a very fast entropy stage, provided by Huff0 and FSE library. Zstandard version >= 1.0 is required, which is generally only available on newer distributions.
2020-05-04 21:25:27 +02:00
P00 INFO: backup command begin [BACKREST-VERSION]: --buffer-size=[BUFFER-SIZE] --checksum-page --compress-level=3 --compress-type=zst --config=[TEST_PATH]/db-master/pgbackrest.conf --db-timeout=45 --lock-path=[TEST_PATH]/db-master/lock --log-level-console=detail --log-level-file=[LOG-LEVEL-FILE] --log-level-stderr=off --log-path=[TEST_PATH]/db-master/log[] --no-log-timestamp --no-online --pg1-path=[TEST_PATH]/db-master/db/base-2 --protocol-timeout=60 --repo1-hardlink --repo1-path=[TEST_PATH]/db-master/repo --stanza=db --start-fast --type=diff
Add time-based retention for full backups. The --repo-retention-full-type option allows retention of full backups based on a time period, specified in days. The new option will default to 'count' and therefore will not affect current installations. Setting repo-retention-full-type to 'time' will allow the user to use a time period, in days, to indicate full backup retention. Using this method, a full backup can be expired only if the time the backup completed is older than the number of days set with repo-retention-full (calculated from the moment the 'expire' command is run) and at least one full backup meets the retention period. If archive retention has not been configured, then the default settings will expire archives that are prior to the oldest retained full backup. For example, if there are three full backups ending in times that are 25 days old (F1), 20 days old (F2) and 10 days old (F3), then if the full retention period is 15 days, then only F1 will be expired; F2 will be retained because F1 is not at least 15 days old.
2020-05-08 21:25:03 +02:00
P00 WARN: option 'repo1-retention-full' is not set for 'repo1-retention-full-type=count', the repository may run out of space
Rename retention-* options to repo-retention-*.
2018-02-19 22:07:24 +02:00
HINT: to retain full backups indefinitely (without warning), set option 'repo1-retention-full' to the maximum.
Show Process ID in log output instead of filtering it out with the timestamp.
2016-09-14 23:30:57 +02:00
P00 INFO: last backup label = [BACKUP-FULL-3], version = [VERSION-1]
Add checksum delta for incremental backups. Use checksums rather than timestamps to determine if files have changed. This is useful in cases where the timestamps may not be trustworthy, e.g. when performing an incremental after failing over to a standby. If checksum delta is enabled then checksums will be used for verification of resumed backups, even if they are full. Resumes have always used checksums to verify the files in the repository, enabling delta performs checksums on the database files as well. Note that the user must manually enable this feature in cases were it would be useful or just keep in enabled all the time. A future commit will address automatically enabling the feature in cases where it seems likely to be useful. Contributed by Cynthia Shang.
2018-09-19 17:12:45 +02:00
P01 INFO: backup file [TEST_PATH]/db-master/db/base-2/base/base2.txt (9B, 100%) checksum cafac3c59553f2cfde41ce2e62e7662295f108c0
P00 DETAIL: hardlink pg_data/PG_VERSION to [BACKUP-FULL-3]
P00 DETAIL: hardlink pg_data/badchecksum.txt to [BACKUP-FULL-3]
P00 DETAIL: hardlink pg_data/base/1/12000 to [BACKUP-FULL-3]
P00 DETAIL: hardlink pg_data/base/1/PG_VERSION to [BACKUP-FULL-3]
P00 DETAIL: hardlink pg_data/base/16384/17000 to [BACKUP-FULL-3]
P00 DETAIL: hardlink pg_data/base/16384/PG_VERSION to [BACKUP-FULL-3]
P00 DETAIL: hardlink pg_data/base/32768/33000 to [BACKUP-FULL-3]
P00 DETAIL: hardlink pg_data/base/32768/33000.32767 to [BACKUP-FULL-3]
Full/Synthetic test refactor. * Combine hardlink and non/compressed in synthetic tests to reduce test time and improve coverage. * Change log level of hardlink logging to detail. * Cast size in S3 manifest to integer.
2017-06-15 21:32:10 +02:00
P00 DETAIL: hardlink pg_data/base/32768/33001 to [BACKUP-FULL-3]
Exclude temporary and unlogged relation (table/index) files from backup. Implemented using the same logic as the patches adding this feature to PostgreSQL, 8694cc96 and 920a5e50. Temporary relation exclusion is enabled in PostgreSQL ≥ 9.0. Unlogged relation exclusion is enabled in PostgreSQL ≥ 9.1, where the feature was introduced. Contributed by Cynthia Shang.
2018-07-31 00:53:34 +02:00
P00 DETAIL: hardlink pg_data/base/32768/44000_init to [BACKUP-FULL-3]
Add checksum delta for incremental backups. Use checksums rather than timestamps to determine if files have changed. This is useful in cases where the timestamps may not be trustworthy, e.g. when performing an incremental after failing over to a standby. If checksum delta is enabled then checksums will be used for verification of resumed backups, even if they are full. Resumes have always used checksums to verify the files in the repository, enabling delta performs checksums on the database files as well. Note that the user must manually enable this feature in cases were it would be useful or just keep in enabled all the time. A future commit will address automatically enabling the feature in cases where it seems likely to be useful. Contributed by Cynthia Shang.
2018-09-19 17:12:45 +02:00
P00 DETAIL: hardlink pg_data/base/32768/PG_VERSION to [BACKUP-FULL-3]
P00 DETAIL: hardlink pg_data/changecontent.txt to [BACKUP-FULL-3]
P00 DETAIL: hardlink pg_data/changetime.txt to [BACKUP-FULL-3]
Full/Synthetic test refactor. * Combine hardlink and non/compressed in synthetic tests to reduce test time and improve coverage. * Change log level of hardlink logging to detail. * Cast size in S3 manifest to integer.
2017-06-15 21:32:10 +02:00
P00 DETAIL: hardlink pg_data/global/pg_control to [BACKUP-FULL-3]
P00 DETAIL: hardlink pg_data/pg_stat/global.stat to [BACKUP-FULL-3]
Add checksum delta for incremental backups. Use checksums rather than timestamps to determine if files have changed. This is useful in cases where the timestamps may not be trustworthy, e.g. when performing an incremental after failing over to a standby. If checksum delta is enabled then checksums will be used for verification of resumed backups, even if they are full. Resumes have always used checksums to verify the files in the repository, enabling delta performs checksums on the database files as well. Note that the user must manually enable this feature in cases were it would be useful or just keep in enabled all the time. A future commit will address automatically enabling the feature in cases where it seems likely to be useful. Contributed by Cynthia Shang.
2018-09-19 17:12:45 +02:00
P00 DETAIL: hardlink pg_data/postgresql.conf to [BACKUP-FULL-3]
Use minio for integration testing. ScalityS3 has not received any maintenance in years and is slow to start which is bad for testing. Replace it with minio which starts quickly and ships as a single executable or a tiny container. Minio has stricter limits on allowable characters but should still provide enough coverage to show that our encoding is working correctly. This commit also includes the upgrade to openssl 1.1.1 in the Ubuntu 18.04 container.
2019-07-03 04:20:35 +02:00
P00 DETAIL: hardlink pg_data/special-!_.*'()&!@;:+,? to [BACKUP-FULL-3]
Add checksum delta for incremental backups. Use checksums rather than timestamps to determine if files have changed. This is useful in cases where the timestamps may not be trustworthy, e.g. when performing an incremental after failing over to a standby. If checksum delta is enabled then checksums will be used for verification of resumed backups, even if they are full. Resumes have always used checksums to verify the files in the repository, enabling delta performs checksums on the database files as well. Note that the user must manually enable this feature in cases were it would be useful or just keep in enabled all the time. A future commit will address automatically enabling the feature in cases where it seems likely to be useful. Contributed by Cynthia Shang.
2018-09-19 17:12:45 +02:00
P00 DETAIL: hardlink pg_data/zero_from_start to [BACKUP-FULL-3]
P00 DETAIL: hardlink pg_data/zerosize.txt to [BACKUP-FULL-3]
P00 DETAIL: hardlink pg_tblspc/2/[TS_PATH-1]/32768/tablespace2.txt to [BACKUP-FULL-3]
P00 DETAIL: hardlink pg_tblspc/2/[TS_PATH-1]/32768/tablespace2c.txt to [BACKUP-FULL-3]
Show Process ID in log output instead of filtering it out with the timestamp.
2016-09-14 23:30:57 +02:00
P00 INFO: diff backup size = 9B
Remove integration tests that are now covered in the unit tests. Most of these tests are just checking that errors are thrown when required. These are well covered in various unit tests. The "cannot resume" tests are also well covered in the backup unit tests. Finally, config warnings are well covered in the config unit tests. There is more to be done here, but this accounts for the low-hanging fruit.
2019-12-18 03:14:45 +02:00
P00 INFO: new backup label = [BACKUP-DIFF-4]
Added a predictable log entry to signal that a command has completed successfully. For example a backup ends successfully with: INFO: backup command end: completed successfully.
2016-12-04 21:44:53 +02:00
P00 INFO: backup command end: completed successfully
Integration test improvements for disk and memory efficiency. Set log-level-file=off when more that one test will run. In this case is it impossible to see the logs anyway since they will be automatically cleaned up after the test. This improves performance pretty dramatically since trace-level logging is expensive. If a singe integration test is run then log-level-file is trace by default but can be changed with the --log-level-test-file option. Reduce buffer-size to 64k to save memory during testing and allow more processes to run in parallel. Update log replacement rules so that these options can change without affecting expect logs.
2019-12-17 22:23:07 +02:00
P00 INFO: expire command begin [BACKREST-VERSION]: --buffer-size=[BUFFER-SIZE] --config=[TEST_PATH]/db-master/pgbackrest.conf --lock-path=[TEST_PATH]/db-master/lock --log-level-console=detail --log-level-file=[LOG-LEVEL-FILE] --log-level-stderr=off --log-path=[TEST_PATH]/db-master/log[] --no-log-timestamp --repo1-path=[TEST_PATH]/db-master/repo --stanza=db
Rename retention-* options to repo-retention-*.
2018-02-19 22:07:24 +02:00
P00 INFO: option 'repo1-retention-archive' is not set - archive logs will not be expired
Added a predictable log entry to signal that a command has completed successfully. For example a backup ends successfully with: INFO: backup command end: completed successfully.
2016-12-04 21:44:53 +02:00
P00 INFO: expire command end: completed successfully
Added tablespace setting to allow compact database restores for development, staging, etc.
2015-05-09 00:34:27 +02:00
Rename test paths for clarity. This was worked out as part of the test suite refactor [c8f806a] but not committed with it because of the large number of expect logs changes involved. Keeping them separate made it easier to audit the changes in the refactor.
2016-06-24 16:54:31 +02:00
+ supplemental file: [TEST_PATH]/db-master/pgbackrest.conf
----------------------------------------------------------
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
[db]
Rename db-* options to pg-* and backup-* options to repo-* to improve consistency. * repo-* options are now indexed although only one is allowed. * List deprecated option names in documentation and command-line help.
2018-02-04 01:27:38 +02:00
pg1-path=[TEST_PATH]/db-master/db/base-2
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
New simpler configuration and consistent project/exe/path naming. * The repo-path option now always refers to the repository where backups and archive are stored, whether local or remote, so the repo-remote-path option has been removed. The new spool-path option can be used to define a location for queueing WAL segments when archiving asynchronously. Otherwise, a local repository is no longer required. * Implemented a new config format which should be far simpler to use. See the User Guide and Configuration Reference for details but for a simple configuration all options can now be placed in the stanza section. Options that are shared between stanzas can be placed in the [global] section. More complex configurations can still make use of command sections though this should be a rare use case. * The default configuration filename is now pgbackrest.conf instead of pg_backrest.conf. This was done for consistency with other naming changes but also to prevent old config files from being loaded accidentally. * The default repository name was changed from /var/lib/backup to /var/lib/pgbackrest. * Lock files are now stored in /tmp/pgbackrest by default. These days /run/pgbackrest would be the preferred location but that would require init scripts which are not part of this release. The lock-path option can be used to configure the lock directory. * Log files are now stored in /var/log/pgbackrest by default and no longer have the date appended so they can be managed with logrotate. The log-path option can be used to configure the lock directory. * Executable filename changed from pg_backrest to pgbackrest.
2016-04-14 15:30:54 +02:00
[db:restore]
Rename test paths for clarity. This was worked out as part of the test suite refactor [c8f806a] but not committed with it because of the large number of expect logs changes involved. Keeping them separate made it easier to audit the changes in the refactor.
2016-06-24 16:54:31 +02:00
tablespace-map=1=[TEST_PATH]/db-master/db/tablespace/ts1-2
tablespace-map=2=[TEST_PATH]/db-master/db/tablespace/ts2-2
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
New simpler configuration and consistent project/exe/path naming. * The repo-path option now always refers to the repository where backups and archive are stored, whether local or remote, so the repo-remote-path option has been removed. The new spool-path option can be used to define a location for queueing WAL segments when archiving asynchronously. Otherwise, a local repository is no longer required. * Implemented a new config format which should be far simpler to use. See the User Guide and Configuration Reference for details but for a simple configuration all options can now be placed in the stanza section. Options that are shared between stanzas can be placed in the [global] section. More complex configurations can still make use of command sections though this should be a rare use case. * The default configuration filename is now pgbackrest.conf instead of pg_backrest.conf. This was done for consistency with other naming changes but also to prevent old config files from being loaded accidentally. * The default repository name was changed from /var/lib/backup to /var/lib/pgbackrest. * Lock files are now stored in /tmp/pgbackrest by default. These days /run/pgbackrest would be the preferred location but that would require init scripts which are not part of this release. The lock-path option can be used to configure the lock directory. * Log files are now stored in /var/log/pgbackrest by default and no longer have the date appended so they can be managed with logrotate. The log-path option can be used to configure the lock directory. * Executable filename changed from pg_backrest to pgbackrest.
2016-04-14 15:30:54 +02:00
[global]
Integration test improvements for disk and memory efficiency. Set log-level-file=off when more that one test will run. In this case is it impossible to see the logs anyway since they will be automatically cleaned up after the test. This improves performance pretty dramatically since trace-level logging is expensive. If a singe integration test is run then log-level-file is trace by default but can be changed with the --log-level-test-file option. Reduce buffer-size to 64k to save memory during testing and allow more processes to run in parallel. Update log replacement rules so that these options can change without affecting expect logs.
2019-12-17 22:23:07 +02:00
buffer-size=[BUFFER-SIZE]
Fix custom settings for compress-level option being ignored. Reported by Jens Wilke.
2017-10-23 00:05:46 +02:00
compress-level=3
Add Zstandard compression support. Zstandard is a fast lossless compression algorithm targeting real-time compression scenarios at zlib-level and better compression ratios. It's backed by a very fast entropy stage, provided by Huff0 and FSE library. Zstandard version >= 1.0 is required, which is generally only available on newer distributions.
2020-05-04 21:25:27 +02:00
compress-type=zst
Storage and IO layer refactor: Refactor storage layer to allow for new repository filesystems using drivers. (Reviewed by Cynthia Shang.) Refactor IO layer to allow for new compression formats, checksum types, and other capabilities using filters. (Reviewed by Cynthia Shang.)
2017-06-09 23:51:41 +02:00
db-timeout=45
lock-path=[TEST_PATH]/db-master/lock
Reduce log level for all expect tests to detail. The C code is designed to be efficient rather than deterministic at the debug log level. As we move more testing from integration to unit tests it makes less sense to try and maintain the expect logs at this log level. Most of the expect logs have already been moved to detail level but mock/all still had tests at debug level. Change the logging defaults in the config file and remove as many references to log-level-console as possible.
2019-05-23 00:23:44 +02:00
log-level-console=detail
Integration test improvements for disk and memory efficiency. Set log-level-file=off when more that one test will run. In this case is it impossible to see the logs anyway since they will be automatically cleaned up after the test. This improves performance pretty dramatically since trace-level logging is expensive. If a singe integration test is run then log-level-file is trace by default but can be changed with the --log-level-test-file option. Reduce buffer-size to 64k to save memory during testing and allow more processes to run in parallel. Update log replacement rules so that these options can change without affecting expect logs.
2019-12-17 22:23:07 +02:00
log-level-file=[LOG-LEVEL-FILE]
Added the log-level-stderr option. Controls whether console log messages are sent to stderr or stdout. By default this is set to warn which represents a change in behavior from previous versions, even though it may be more intuitive. Setting log-level-stderr=off will preserve the old behavior. Suggested by Sascha Biberhofer.
2016-10-05 15:09:30 +02:00
log-level-stderr=off
Storage and IO layer refactor: Refactor storage layer to allow for new repository filesystems using drivers. (Reviewed by Cynthia Shang.) Refactor IO layer to allow for new compression formats, checksum types, and other capabilities using filters. (Reviewed by Cynthia Shang.)
2017-06-09 23:51:41 +02:00
log-path=[TEST_PATH]/db-master/log
Integration test improvements for disk and memory efficiency. Set log-level-file=off when more that one test will run. In this case is it impossible to see the logs anyway since they will be automatically cleaned up after the test. This improves performance pretty dramatically since trace-level logging is expensive. If a singe integration test is run then log-level-file is trace by default but can be changed with the --log-level-test-file option. Reduce buffer-size to 64k to save memory during testing and allow more processes to run in parallel. Update log replacement rules so that these options can change without affecting expect logs.
2019-12-17 22:23:07 +02:00
log-subprocess=[LOG-SUBPROCESS]
Set log-timestamp=n for integration tests. This means less filtering of logs needs to be done and new timestamps can be added without adding new filters.
2018-05-11 17:24:38 +02:00
log-timestamp=n
Storage and IO layer refactor: Refactor storage layer to allow for new repository filesystems using drivers. (Reviewed by Cynthia Shang.) Refactor IO layer to allow for new compression formats, checksum types, and other capabilities using filters. (Reviewed by Cynthia Shang.)
2017-06-09 23:51:41 +02:00
protocol-timeout=60
Rename db-* options to pg-* and backup-* options to repo-* to improve consistency. * repo-* options are now indexed although only one is allowed. * List deprecated option names in documentation and command-line help.
2018-02-04 01:27:38 +02:00
repo1-hardlink=y
repo1-path=[TEST_PATH]/db-master/repo
Consolidate archive-push tests into new archive test.
2017-06-21 14:38:53 +02:00
spool-path=[TEST_PATH]/db-master/spool
New simpler configuration and consistent project/exe/path naming. * The repo-path option now always refers to the repository where backups and archive are stored, whether local or remote, so the repo-remote-path option has been removed. The new spool-path option can be used to define a location for queueing WAL segments when archiving asynchronously. Otherwise, a local repository is no longer required. * Implemented a new config format which should be far simpler to use. See the User Guide and Configuration Reference for details but for a simple configuration all options can now be placed in the stanza section. Options that are shared between stanzas can be placed in the [global] section. More complex configurations can still make use of command sections though this should be a rare use case. * The default configuration filename is now pgbackrest.conf instead of pg_backrest.conf. This was done for consistency with other naming changes but also to prevent old config files from being loaded accidentally. * The default repository name was changed from /var/lib/backup to /var/lib/pgbackrest. * Lock files are now stored in /tmp/pgbackrest by default. These days /run/pgbackrest would be the preferred location but that would require init scripts which are not part of this release. The lock-path option can be used to configure the lock directory. * Log files are now stored in /var/log/pgbackrest by default and no longer have the date appended so they can be managed with logrotate. The log-path option can be used to configure the lock directory. * Executable filename changed from pg_backrest to pgbackrest.
2016-04-14 15:30:54 +02:00
[global:backup]
archive-copy=y
start-fast=y
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
Remove integration tests that are now covered in the unit tests. Most of these tests are just checking that errors are thrown when required. These are well covered in various unit tests. The "cannot resume" tests are also well covered in the backup unit tests. Finally, config warnings are well covered in the config unit tests. There is more to be done here, but this accounts for the low-hanging fruit.
2019-12-18 03:14:45 +02:00
+ supplemental file: [TEST_PATH]/db-master/repo/backup/db/[BACKUP-DIFF-4]/backup.manifest
Rename test paths for clarity. This was worked out as part of the test suite refactor [c8f806a] but not committed with it because of the large number of expect logs changes involved. Keeping them separate made it easier to audit the changes in the refactor.
2016-06-24 16:54:31 +02:00
-----------------------------------------------------------------------------------------
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
[backrest]
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
backrest-format=5
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
backrest-version="[VERSION-1]"
[backup]
Remove integration tests that are now covered in the unit tests. Most of these tests are just checking that errors are thrown when required. These are well covered in various unit tests. The "cannot resume" tests are also well covered in the backup unit tests. Finally, config warnings are well covered in the config unit tests. There is more to be done here, but this accounts for the low-hanging fruit.
2019-12-18 03:14:45 +02:00
backup-label="[BACKUP-DIFF-4]"
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
backup-prior="[BACKUP-FULL-3]"
backup-timestamp-copy-start=[TIMESTAMP]
backup-timestamp-start=[TIMESTAMP]
backup-timestamp-stop=[TIMESTAMP]
backup-type="diff"
[backup:db]
Database version refactoring. * Refactor db version constants into a separate module. * Update synthetic backup tests to PostgreSQL 9.4.
2016-08-12 04:35:24 +02:00
db-catalog-version=201409291
db-control-version=942
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
db-id=1
Automate generation of WAL and pg_control test files. The existing static files would not work with 32-bit or big-endian systems so create functions to generate these files dynamically rather than creating a bunch of new static files.
2017-11-19 03:02:54 +02:00
db-system-id=1000000000000000094
Database version refactoring. * Refactor db version constants into a separate module. * Update synthetic backup tests to PostgreSQL 9.4.
2016-08-12 04:35:24 +02:00
db-version="9.4"
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
[backup:option]
Set archive-check option in manifest correctly when offline. Archive check does not run when in offline backup mode but the option was set to true in the manifest. It's harmless since these options are informational only but it could cause confusion when debugging.
2019-11-28 15:27:21 +02:00
option-archive-check=false
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
option-archive-copy=true
Backup from a standby cluster. A connection to the primary cluster is still required to start/stop the backup and copy files that are not replicated, but the vast majority of files are copied from the standby in order to reduce load on the master.
2016-08-25 17:25:46 +02:00
option-backup-standby=false
Integration test improvements for disk and memory efficiency. Set log-level-file=off when more that one test will run. In this case is it impossible to see the logs anyway since they will be automatically cleaned up after the test. This improves performance pretty dramatically since trace-level logging is expensive. If a singe integration test is run then log-level-file is trace by default but can be changed with the --log-level-test-file option. Reduce buffer-size to 64k to save memory during testing and allow more processes to run in parallel. Update log replacement rules so that these options can change without affecting expect logs.
2019-12-17 22:23:07 +02:00
option-buffer-size=[BUFFER-SIZE]
Added the --checksum-page option. This option allows pgBackRest to validate page checksums in data files when checksums are enabled on PostgreSQL >= 9.3. Note that this functionality requires a C library which may not initially be available in OS packages. The option will automatically be enabled when the library is present and checksums are enabled on the cluster.
2016-12-13 01:54:07 +02:00
option-checksum-page=false
Full/Synthetic test refactor. * Combine hardlink and non/compressed in synthetic tests to reduce test time and improve coverage. * Change log level of hardlink logging to detail. * Cast size in S3 manifest to integer.
2017-06-15 21:32:10 +02:00
option-compress=true
Add additional options to backup.manifest for debugging purposes. Add the buffer-size, compress-level, compress-level-network, and process-max options to the backup:option section in backup.manifest to aid in debugging. It may also make sense to propagate these options up to backup.info so they can be displayed in the info command, but for now this is deemed sufficient. Contributed by blogh.
2019-03-10 11:03:52 +02:00
option-compress-level=3
option-compress-level-network=3
Add Zstandard compression support. Zstandard is a fast lossless compression algorithm targeting real-time compression scenarios at zlib-level and better compression ratios. It's backed by a very fast entropy stage, provided by Huff0 and FSE library. Zstandard version >= 1.0 is required, which is generally only available on newer distributions.
2020-05-04 21:25:27 +02:00
option-compress-type="zst"
Add checksum delta for incremental backups. Use checksums rather than timestamps to determine if files have changed. This is useful in cases where the timestamps may not be trustworthy, e.g. when performing an incremental after failing over to a standby. If checksum delta is enabled then checksums will be used for verification of resumed backups, even if they are full. Resumes have always used checksums to verify the files in the repository, enabling delta performs checksums on the database files as well. Note that the user must manually enable this feature in cases were it would be useful or just keep in enabled all the time. A future commit will address automatically enabling the feature in cases where it seems likely to be useful. Contributed by Cynthia Shang.
2018-09-19 17:12:45 +02:00
option-delta=false
Full/Synthetic test refactor. * Combine hardlink and non/compressed in synthetic tests to reduce test time and improve coverage. * Change log level of hardlink logging to detail. * Cast size in S3 manifest to integer.
2017-06-15 21:32:10 +02:00
option-hardlink=true
Closed #167: Rename `--no-start-stop` option to `--no-online`.
2016-02-20 02:32:02 +02:00
option-online=false
Add additional options to backup.manifest for debugging purposes. Add the buffer-size, compress-level, compress-level-network, and process-max options to the backup:option section in backup.manifest to aid in debugging. It may also make sense to propagate these options up to backup.info so they can be displayed in the info command, but for now this is deemed sufficient. Contributed by blogh.
2019-03-10 11:03:52 +02:00
option-process-max=1
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
[backup:target]
Rename test paths for clarity. This was worked out as part of the test suite refactor [c8f806a] but not committed with it because of the large number of expect logs changes involved. Keeping them separate made it easier to audit the changes in the refactor.
2016-06-24 16:54:31 +02:00
pg_data={"path":"[TEST_PATH]/db-master/db/base-2","type":"path"}
pg_tblspc/2={"path":"[TEST_PATH]/db-master/db/tablespace/ts2-2","tablespace-id":"2","tablespace-name":"ts2","type":"link"}
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
[target:file]
Full/Synthetic test refactor. * Combine hardlink and non/compressed in synthetic tests to reduce test time and improve coverage. * Change log level of hardlink logging to detail. * Cast size in S3 manifest to integer.
2017-06-15 21:32:10 +02:00
pg_data/PG_VERSION={"checksum":"184473f470864e067ee3a22e64b47b0a1c356f29","master":true,"reference":"[BACKUP-FULL-3]","repo-size":[SIZE],"size":[SIZE],"timestamp":[TIMESTAMP-1]}
pg_data/badchecksum.txt={"checksum":"f927212cd08d11a42a666b2f04235398e9ceeb51","master":true,"reference":"[BACKUP-FULL-3]","repo-size":[SIZE],"size":[SIZE],"timestamp":[TIMESTAMP-1]}
pg_data/base/1/12000={"checksum":"22c98d248ff548311eda88559e4a8405ed77c003","reference":"[BACKUP-FULL-3]","repo-size":[SIZE],"size":[SIZE],"timestamp":[TIMESTAMP-1]}
pg_data/base/1/PG_VERSION={"checksum":"184473f470864e067ee3a22e64b47b0a1c356f29","mode":"0660","reference":"[BACKUP-FULL-3]","repo-size":[SIZE],"size":[SIZE],"timestamp":[TIMESTAMP-1]}
pg_data/base/16384/17000={"checksum":"7579ada0808d7f98087a0a586d0df9de009cdc33","reference":"[BACKUP-FULL-3]","repo-size":[SIZE],"size":[SIZE],"timestamp":[TIMESTAMP-1]}
pg_data/base/16384/PG_VERSION={"checksum":"184473f470864e067ee3a22e64b47b0a1c356f29","reference":"[BACKUP-FULL-3]","repo-size":[SIZE],"size":[SIZE],"timestamp":[TIMESTAMP-1]}
Fix incorrect offline upper bound for ignoring page checksum errors. For offline backups the upper bound was being set to 0x0000FFFF0000FFFF rather than UINT64_MAX. This meant that page checksum errors might be ignored for databases with a lot of past WAL in offline mode. Online mode is not affected since the upper bound is retrieved from pg_start_backup().
2019-07-11 15:13:56 +02:00
pg_data/base/32768/33000={"checksum":"7a16d165e4775f7c92e8cdf60c0af57313f0bf90","reference":"[BACKUP-FULL-3]","repo-size":[SIZE],"size":[SIZE],"timestamp":[TIMESTAMP-1]}
pg_data/base/32768/33000.32767={"checksum":"6e99b589e550e68e934fd235ccba59fe5b592a9e","reference":"[BACKUP-FULL-3]","repo-size":[SIZE],"size":[SIZE],"timestamp":[TIMESTAMP-1]}
Full/Synthetic test refactor. * Combine hardlink and non/compressed in synthetic tests to reduce test time and improve coverage. * Change log level of hardlink logging to detail. * Cast size in S3 manifest to integer.
2017-06-15 21:32:10 +02:00
pg_data/base/32768/33001={"checksum":"6bf316f11d28c28914ea9be92c00de9bea6d9a6b","reference":"[BACKUP-FULL-3]","repo-size":[SIZE],"size":[SIZE],"timestamp":[TIMESTAMP-1]}
Fix incorrect offline upper bound for ignoring page checksum errors. For offline backups the upper bound was being set to 0x0000FFFF0000FFFF rather than UINT64_MAX. This meant that page checksum errors might be ignored for databases with a lot of past WAL in offline mode. Online mode is not affected since the upper bound is retrieved from pg_start_backup().
2019-07-11 15:13:56 +02:00
pg_data/base/32768/44000_init={"checksum":"7a16d165e4775f7c92e8cdf60c0af57313f0bf90","reference":"[BACKUP-FULL-3]","repo-size":[SIZE],"size":[SIZE],"timestamp":[TIMESTAMP-1]}
Full/Synthetic test refactor. * Combine hardlink and non/compressed in synthetic tests to reduce test time and improve coverage. * Change log level of hardlink logging to detail. * Cast size in S3 manifest to integer.
2017-06-15 21:32:10 +02:00
pg_data/base/32768/PG_VERSION={"checksum":"184473f470864e067ee3a22e64b47b0a1c356f29","reference":"[BACKUP-FULL-3]","repo-size":[SIZE],"size":[SIZE],"timestamp":[TIMESTAMP-1]}
pg_data/base/base2.txt={"checksum":"cafac3c59553f2cfde41ce2e62e7662295f108c0","repo-size":[SIZE],"size":[SIZE],"timestamp":[TIMESTAMP-1]}
Add checksum delta for incremental backups. Use checksums rather than timestamps to determine if files have changed. This is useful in cases where the timestamps may not be trustworthy, e.g. when performing an incremental after failing over to a standby. If checksum delta is enabled then checksums will be used for verification of resumed backups, even if they are full. Resumes have always used checksums to verify the files in the repository, enabling delta performs checksums on the database files as well. Note that the user must manually enable this feature in cases were it would be useful or just keep in enabled all the time. A future commit will address automatically enabling the feature in cases where it seems likely to be useful. Contributed by Cynthia Shang.
2018-09-19 17:12:45 +02:00
pg_data/changecontent.txt={"checksum":"a094d94583e209556d03c3c5da33131a065f1689","master":true,"reference":"[BACKUP-FULL-3]","repo-size":[SIZE],"size":[SIZE],"timestamp":[TIMESTAMP-1]}
pg_data/changetime.txt={"checksum":"88087292ed82e26f3eb824d0bffc05ccf7a30f8d","master":true,"reference":"[BACKUP-FULL-3]","repo-size":[SIZE],"size":[SIZE],"timestamp":[TIMESTAMP-2]}
Support configurable WAL segment size. PostgreSQL 11 introduces configurable WAL segment sizes, from 1MB to 1GB. There are two areas that needed to be updated to support this: building the archive-get queue and checking that WAL has been archived after a backup. Both operations require the WAL segment size to properly build a list. Checking the archive after a backup is still implemented in Perl and has an active database connection, so just get the WAL segment size from the database. The archive-get command does not have a connection to the database, so get the WAL segment size from pg_control instead. This requires a deeper inspection of pg_control than has been done in the past, so it seemed best to copy the relevant data structures from each version of PostgreSQL and build a generic interface layer to address them. While this approach is a bit verbose, it has the advantage of being relatively simple, and can easily be updated for new versions of PostgreSQL. Since the integration tests generate pg_control files for testing, teach Perl how to generate files with the correct offsets for both 32-bit and 64-bit architectures.
2018-09-25 11:24:42 +02:00
pg_data/global/pg_control={"checksum":"4c77c900f7af0d9ab13fa9982051a42e0b637f6c","master":true,"reference":"[BACKUP-FULL-3]","repo-size":[SIZE],"size":[SIZE],"timestamp":[TIMESTAMP-2]}
Full/Synthetic test refactor. * Combine hardlink and non/compressed in synthetic tests to reduce test time and improve coverage. * Change log level of hardlink logging to detail. * Cast size in S3 manifest to integer.
2017-06-15 21:32:10 +02:00
pg_data/pg_stat/global.stat={"checksum":"e350d5ce0153f3e22d5db21cf2a4eff00f3ee877","master":true,"reference":"[BACKUP-FULL-3]","repo-size":[SIZE],"size":[SIZE],"timestamp":[TIMESTAMP-2]}
pg_data/postgresql.conf={"checksum":"6721d92c9fcdf4248acff1f9a1377127d9064807","master":true,"reference":"[BACKUP-FULL-3]","repo-size":[SIZE],"size":[SIZE],"timestamp":[TIMESTAMP-2]}
Use minio for integration testing. ScalityS3 has not received any maintenance in years and is slow to start which is bad for testing. Replace it with minio which starts quickly and ships as a single executable or a tiny container. Minio has stricter limits on allowable characters but should still provide enough coverage to show that our encoding is working correctly. This commit also includes the upgrade to openssl 1.1.1 in the Ubuntu 18.04 container.
2019-07-03 04:20:35 +02:00
pg_data/special-!_.*'()&!@;:+,?={"master":true,"reference":"[BACKUP-FULL-3]","repo-size":[SIZE],"size":[SIZE],"timestamp":[TIMESTAMP-1]}
Allow zero-size files in backup manifest to reference a prior manifest regardless of timestamp delta. Contributed by Cynthia Shang.
2018-08-24 22:50:33 +02:00
pg_data/zero_from_start={"master":true,"reference":"[BACKUP-FULL-3]","repo-size":[SIZE],"size":[SIZE],"timestamp":[TIMESTAMP-1]}
Add zero-length file to mock/all test.
2018-07-05 21:40:50 +02:00
pg_data/zerosize.txt={"master":true,"reference":"[BACKUP-FULL-3]","repo-size":[SIZE],"size":[SIZE],"timestamp":[TIMESTAMP-2]}
Full/Synthetic test refactor. * Combine hardlink and non/compressed in synthetic tests to reduce test time and improve coverage. * Change log level of hardlink logging to detail. * Cast size in S3 manifest to integer.
2017-06-15 21:32:10 +02:00
pg_tblspc/2/[TS_PATH-1]/32768/tablespace2.txt={"checksum":"dc7f76e43c46101b47acc55ae4d593a9e6983578","reference":"[BACKUP-FULL-3]","repo-size":[SIZE],"size":[SIZE],"timestamp":[TIMESTAMP-1]}
pg_tblspc/2/[TS_PATH-1]/32768/tablespace2c.txt={"checksum":"dfcb8679956b734706cf87259d50c88f83e80e66","reference":"[BACKUP-FULL-3]","repo-size":[SIZE],"size":[SIZE],"timestamp":[TIMESTAMP-1]}
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
[target:file:default]
Continuous integration using travis-ci.
2017-02-21 15:59:23 +02:00
group="[GROUP-1]"
Fixed a issue where tablespaces were copied from the master during standby backup.
2016-09-04 15:19:44 +02:00
master=false
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
mode="0600"
Regression tests can now be run as any properly-configured user, not just vagrant.
2016-10-01 19:39:44 +02:00
user="[USER-1]"
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
[target:link]
Rename test paths for clarity. This was worked out as part of the test suite refactor [c8f806a] but not committed with it because of the large number of expect logs changes involved. Keeping them separate made it easier to audit the changes in the refactor.
2016-06-24 16:54:31 +02:00
pg_data/pg_tblspc/2={"destination":"[TEST_PATH]/db-master/db/tablespace/ts2-2"}
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
[target:link:default]
Continuous integration using travis-ci.
2017-02-21 15:59:23 +02:00
group="[GROUP-1]"
Regression tests can now be run as any properly-configured user, not just vagrant.
2016-10-01 19:39:44 +02:00
user="[USER-1]"
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
[target:path]
The restore command is implemented entirely in C. For the most part this is a direct migration of the Perl code into C. There is one important behavioral change with regard to how file permissions are handled. The Perl code tried to set ownership as it was in the manifest even when running as an unprivileged user. This usually just led to errors and frustration. The C code works like this: If a restore is run as a non-root user (the typical scenario) then all files restored will belong to the user/group executing pgBackRest. If existing files are not owned by the executing user/group then an error will result if the ownership cannot be updated to the executing user/group. In that case the file ownership will need to be updated by a privileged user before the restore can be retried. If a restore is run as the root user then pgBackRest will attempt to recreate the ownership recorded in the manifest when the backup was made. Only user/group names are stored in the manifest so the same names must exist on the restore host for this to work. If the user/group name cannot be found locally then the user/group of the PostgreSQL data directory will be used and finally root if the data directory user/group cannot be mapped to a name. Reviewed by Cynthia Shang.
2019-09-26 13:52:02 +02:00
pg_data={"mode":"0777"}
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
pg_data/base={}
Allow selective restore of databases from a cluster backup. This feature can result in major space and time savings when only specific databases are restored. Unrestored databases will not be accessible but must be manually dropped before they will be removed from the shared catalogue.
2016-05-11 15:21:39 +02:00
pg_data/base/1={}
pg_data/base/16384={}
pg_data/base/32768={}
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
pg_data/global={}
pg_data/pg_clog={}
Exclude directories during backup that are cleaned, recreated, or zeroed by PostgreSQL at startup. These include (depending on the version where they were introduced): pgsql_tmp, pg_dynshmem, pg_notify, pg_replslot, pg_serial, pg_snapshots, pg_stat_tmp, pg_subtrans. The postgresql.auto.conf.tmp file is now excluded in addition to files that were already excluded: backup_label.old, postmaster.opts, postmaster.pid, recovery.conf, recovery.done.
2016-08-16 15:35:16 +02:00
pg_data/pg_dynshmem={}
Allow arbitrary directories and/or files to be excluded from a backup. Misuse of this feature can lead to inconsistent backups so read the --exclude documentation carefully before using.
2018-08-27 21:51:05 +02:00
pg_data/pg_log={}
Exclude directories during backup that are cleaned, recreated, or zeroed by PostgreSQL at startup. These include (depending on the version where they were introduced): pgsql_tmp, pg_dynshmem, pg_notify, pg_replslot, pg_serial, pg_snapshots, pg_stat_tmp, pg_subtrans. The postgresql.auto.conf.tmp file is now excluded in addition to files that were already excluded: backup_label.old, postmaster.opts, postmaster.pid, recovery.conf, recovery.done.
2016-08-16 15:35:16 +02:00
pg_data/pg_notify={}
pg_data/pg_replslot={}
pg_data/pg_serial={}
pg_data/pg_snapshots={}
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
pg_data/pg_stat={}
Exclude directories during backup that are cleaned, recreated, or zeroed by PostgreSQL at startup. These include (depending on the version where they were introduced): pgsql_tmp, pg_dynshmem, pg_notify, pg_replslot, pg_serial, pg_snapshots, pg_stat_tmp, pg_subtrans. The postgresql.auto.conf.tmp file is now excluded in addition to files that were already excluded: backup_label.old, postmaster.opts, postmaster.pid, recovery.conf, recovery.done.
2016-08-16 15:35:16 +02:00
pg_data/pg_stat_tmp={}
pg_data/pg_subtrans={}
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
pg_data/pg_tblspc={}
The restore command is implemented entirely in C. For the most part this is a direct migration of the Perl code into C. There is one important behavioral change with regard to how file permissions are handled. The Perl code tried to set ownership as it was in the manifest even when running as an unprivileged user. This usually just led to errors and frustration. The C code works like this: If a restore is run as a non-root user (the typical scenario) then all files restored will belong to the user/group executing pgBackRest. If existing files are not owned by the executing user/group then an error will result if the ownership cannot be updated to the executing user/group. In that case the file ownership will need to be updated by a privileged user before the restore can be retried. If a restore is run as the root user then pgBackRest will attempt to recreate the ownership recorded in the manifest when the backup was made. Only user/group names are stored in the manifest so the same names must exist on the restore host for this to work. If the user/group name cannot be found locally then the user/group of the PostgreSQL data directory will be used and finally root if the data directory user/group cannot be mapped to a name. Reviewed by Cynthia Shang.
2019-09-26 13:52:02 +02:00
pg_tblspc={"mode":"0777"}
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
pg_tblspc/2={}
pg_tblspc/2/[TS_PATH-1]={}
Allow selective restore of databases from a cluster backup. This feature can result in major space and time savings when only specific databases are restored. Unrestored databases will not be accessible but must be manually dropped before they will be removed from the shared catalogue.
2016-05-11 15:21:39 +02:00
pg_tblspc/2/[TS_PATH-1]/32768={}
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
[target:path:default]
Continuous integration using travis-ci.
2017-02-21 15:59:23 +02:00
group="[GROUP-1]"
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
mode="0700"
Regression tests can now be run as any properly-configured user, not just vagrant.
2016-10-01 19:39:44 +02:00
user="[USER-1]"
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
Move info file checksum to the end of the file. Putting the checksum at the beginning of the file made it impossible to stream the file out when saving. The entire file had to be held in memory while it was checksummed so the checksum could be written at the beginning. Instead place the checksum at the end. This does not break the existing Perl or C code since the read is not order dependent. There are no plans to improve the Perl code to take advantage of this change, but it will make the C implementation more efficient. Reviewed by Cynthia Shang.
2019-08-22 01:45:48 +02:00
[backrest]
backrest-checksum="[CHECKSUM]"
Rename test paths for clarity. This was worked out as part of the test suite refactor [c8f806a] but not committed with it because of the large number of expect logs changes involved. Keeping them separate made it easier to audit the changes in the refactor.
2016-06-24 16:54:31 +02:00
+ supplemental file: [TEST_PATH]/db-master/repo/backup/db/backup.info
---------------------------------------------------------------------
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
[backrest]
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
backrest-format=5
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
backrest-version="[VERSION-1]"
[backup:current]
Set archive-check option in manifest correctly when offline. Archive check does not run when in offline backup mode but the option was set to true in the manifest. It's harmless since these options are informational only but it could cause confusion when debugging.
2019-11-28 15:27:21 +02:00
[BACKUP-FULL-3]={"backrest-format":5,"backrest-version":"[VERSION-1]","backup-archive-start":null,"backup-archive-stop":null,"backup-info-repo-size":[SIZE],"backup-info-repo-size-delta":[DELTA],"backup-info-size":[SIZE],"backup-info-size-delta":[DELTA],"backup-timestamp-start":[TIMESTAMP],"backup-timestamp-stop":[TIMESTAMP],"backup-type":"full","db-id":1,"option-archive-check":false,"option-archive-copy":true,"option-backup-standby":false,"option-checksum-page":false,"option-compress":true,"option-hardlink":true,"option-online":false}
Remove integration tests that are now covered in the unit tests. Most of these tests are just checking that errors are thrown when required. These are well covered in various unit tests. The "cannot resume" tests are also well covered in the backup unit tests. Finally, config warnings are well covered in the config unit tests. There is more to be done here, but this accounts for the low-hanging fruit.
2019-12-18 03:14:45 +02:00
[BACKUP-DIFF-4]={"backrest-format":5,"backrest-version":"[VERSION-1]","backup-archive-start":null,"backup-archive-stop":null,"backup-info-repo-size":[SIZE],"backup-info-repo-size-delta":[DELTA],"backup-info-size":[SIZE],"backup-info-size-delta":[DELTA],"backup-prior":"[BACKUP-FULL-3]","backup-reference":["[BACKUP-FULL-3]"],"backup-timestamp-start":[TIMESTAMP],"backup-timestamp-stop":[TIMESTAMP],"backup-type":"diff","db-id":1,"option-archive-check":false,"option-archive-copy":true,"option-backup-standby":false,"option-checksum-page":false,"option-compress":true,"option-hardlink":true,"option-online":false}
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
[db]
Database version refactoring. * Refactor db version constants into a separate module. * Update synthetic backup tests to PostgreSQL 9.4.
2016-08-12 04:35:24 +02:00
db-catalog-version=201409291
db-control-version=942
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
db-id=1
Automate generation of WAL and pg_control test files. The existing static files would not work with 32-bit or big-endian systems so create functions to generate these files dynamically rather than creating a bunch of new static files.
2017-11-19 03:02:54 +02:00
db-system-id=1000000000000000094
Database version refactoring. * Refactor db version constants into a separate module. * Update synthetic backup tests to PostgreSQL 9.4.
2016-08-12 04:35:24 +02:00
db-version="9.4"
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
[db:history]
Automate generation of WAL and pg_control test files. The existing static files would not work with 32-bit or big-endian systems so create functions to generate these files dynamically rather than creating a bunch of new static files.
2017-11-19 03:02:54 +02:00
1={"db-catalog-version":201409291,"db-control-version":942,"db-system-id":1000000000000000094,"db-version":"9.4"}
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
Improve performance of info file load/save. Info files required three copies in memory to be loaded (the original string, an ini representation, and the final info object). Not only was this memory inefficient but the Ini object does sequential scans when searching for keys making large files very slow to load. This has not been an issue since archive.info and backup.info are very small, but it becomes a big deal when loading manifests with hundreds of thousands of files. Instead of holding copies of the data in memory, use a callback to deliver the ini data directly to the object when loading. Use a similar method for save to avoid having an intermediate copy. Save is a bit complex because sections/keys must be written in alpha order or older versions of pgBackRest will not calculate the correct checksum. Also move the load retry logic to helper functions rather than embedding it in the Info object. This allows for more flexibility in loading and ensures that stack traces will be available when developing unit tests. Reviewed by Cynthia Shang.
2019-09-06 19:48:28 +02:00
[backrest]
backrest-checksum="[CHECKSUM]"
Major refactor of the test suite. * Make the code more modular and object-oriented. * Multiple Docker containers can now be created for a single test to simulate more realistic environments.
2016-06-24 14:12:58 +02:00
restore delta, remap - selective restore 16384 (db-master host)
Reduce log level for all expect tests to detail. The C code is designed to be efficient rather than deterministic at the debug log level. As we move more testing from integration to unit tests it makes less sense to try and maintain the expect logs at this log level. Most of the expect logs have already been moved to detail level but mock/all still had tests at debug level. Change the logging defaults in the config file and remove as many references to log-level-console as possible.
2019-05-23 00:23:44 +02:00
> [CONTAINER-EXEC] db-master [BACKREST-BIN] --config=[TEST_PATH]/db-master/pgbackrest.conf --delta --db-include=16384 --stanza=db restore
Allow selective restore of databases from a cluster backup. This feature can result in major space and time savings when only specific databases are restored. Unrestored databases will not be accessible but must be manually dropped before they will be removed from the shared catalogue.
2016-05-11 15:21:39 +02:00
------------------------------------------------------------------------------------------------------------------------------------
Remove compress/compress-level options from commands where unused. These commands (e.g. restore, archive-get) never used the compress options but allowed them to be passed on the command line. Now they will error when these options are passed on the command line. If these errors occur then remove the unused options.
2020-02-27 19:25:32 +02:00
P00 INFO: restore command begin [BACKREST-VERSION]: --buffer-size=[BUFFER-SIZE] --config=[TEST_PATH]/db-master/pgbackrest.conf --db-include=16384 --delta --lock-path=[TEST_PATH]/db-master/lock --log-level-console=detail --log-level-file=[LOG-LEVEL-FILE] --log-level-stderr=off --log-path=[TEST_PATH]/db-master/log[] --no-log-timestamp --pg1-path=[TEST_PATH]/db-master/db/base-2 --protocol-timeout=60 --repo1-path=[TEST_PATH]/db-master/repo --stanza=db --tablespace-map=2=[TEST_PATH]/db-master/db/tablespace/ts2-2
Remove integration tests that are now covered in the unit tests. Most of these tests are just checking that errors are thrown when required. These are well covered in various unit tests. The "cannot resume" tests are also well covered in the backup unit tests. Finally, config warnings are well covered in the config unit tests. There is more to be done here, but this accounts for the low-hanging fruit.
2019-12-18 03:14:45 +02:00
P00 INFO: restore backup set [BACKUP-DIFF-4]
The restore command is implemented entirely in C. For the most part this is a direct migration of the Perl code into C. There is one important behavioral change with regard to how file permissions are handled. The Perl code tried to set ownership as it was in the manifest even when running as an unprivileged user. This usually just led to errors and frustration. The C code works like this: If a restore is run as a non-root user (the typical scenario) then all files restored will belong to the user/group executing pgBackRest. If existing files are not owned by the executing user/group then an error will result if the ownership cannot be updated to the executing user/group. In that case the file ownership will need to be updated by a privileged user before the restore can be retried. If a restore is run as the root user then pgBackRest will attempt to recreate the ownership recorded in the manifest when the backup was made. Only user/group names are stored in the manifest so the same names must exist on the restore host for this to work. If the user/group name cannot be found locally then the user/group of the PostgreSQL data directory will be used and finally root if the data directory user/group cannot be mapped to a name. Reviewed by Cynthia Shang.
2019-09-26 13:52:02 +02:00
P00 INFO: map tablespace 'pg_tblspc/2' to '[TEST_PATH]/db-master/db/tablespace/ts2-2'
P00 DETAIL: databases found for selective restore (1, 16384, 32768)
P00 DETAIL: check '[TEST_PATH]/db-master/db/base-2' exists
P00 DETAIL: check '[TEST_PATH]/db-master/db/tablespace/ts2-2/[TS_PATH-1]' exists
P00 DETAIL: remove 'global/pg_control' so cluster will not start if restore does not complete
P00 INFO: remove invalid files/links/paths from '[TEST_PATH]/db-master/db/base-2'
P00 DETAIL: remove invalid file '[TEST_PATH]/db-master/db/base-2/recovery.conf'
P00 INFO: remove invalid files/links/paths from '[TEST_PATH]/db-master/db/tablespace/ts2-2/[TS_PATH-1]'
Exclude temporary and unlogged relation (table/index) files from backup. Implemented using the same logic as the patches adding this feature to PostgreSQL, 8694cc96 and 920a5e50. Temporary relation exclusion is enabled in PostgreSQL ≥ 9.0. Unlogged relation exclusion is enabled in PostgreSQL ≥ 9.1, where the feature was introduced. Contributed by Cynthia Shang.
2018-07-31 00:53:34 +02:00
P01 DETAIL: restore zeroed file [TEST_PATH]/db-master/db/base-2/base/32768/33001 (64KB, 36%)
P01 DETAIL: restore zeroed file [TEST_PATH]/db-master/db/base-2/base/32768/44000_init (32KB, 54%)
P01 DETAIL: restore zeroed file [TEST_PATH]/db-master/db/base-2/base/32768/33000.32767 (32KB, 72%)
P01 DETAIL: restore zeroed file [TEST_PATH]/db-master/db/base-2/base/32768/33000 (32KB, 90%)
Support configurable WAL segment size. PostgreSQL 11 introduces configurable WAL segment sizes, from 1MB to 1GB. There are two areas that needed to be updated to support this: building the archive-get queue and checking that WAL has been archived after a backup. Both operations require the WAL segment size to properly build a list. Checking the archive after a backup is still implemented in Perl and has an active database connection, so just get the WAL segment size from the database. The archive-get command does not have a connection to the database, so get the WAL segment size from pg_control instead. This requires a deeper inspection of pg_control than has been done in the past, so it seemed best to copy the relevant data structures from each version of PostgreSQL and build a generic interface layer to address them. While this approach is a bit verbose, it has the advantage of being relatively simple, and can easily be updated for new versions of PostgreSQL. Since the integration tests generate pg_control files for testing, teach Perl how to generate files with the correct offsets for both 32-bit and 64-bit architectures.
2018-09-25 11:24:42 +02:00
P01 INFO: restore file [TEST_PATH]/db-master/db/base-2/global/pg_control.pgbackrest.tmp (8KB, 95%) checksum 4c77c900f7af0d9ab13fa9982051a42e0b637f6c
Add more realistic data files to synthetic backup and restore tests.
2016-12-10 16:13:55 +02:00
P01 DETAIL: restore file [TEST_PATH]/db-master/db/base-2/base/1/12000 - exists and matches backup (8KB, 99%) checksum 22c98d248ff548311eda88559e4a8405ed77c003
For simplicity, the pg_control file is now copied with the rest of the files instead of by itself of at the end of the process. The backup command does not require this behavior and the restore copies to a temporary file which is renamed at the end of the restore.
2016-11-30 21:23:05 +02:00
P01 DETAIL: restore file [TEST_PATH]/db-master/db/base-2/postgresql.conf - exists and matches backup (21B, 99%) checksum 6721d92c9fcdf4248acff1f9a1377127d9064807
P01 DETAIL: restore file [TEST_PATH]/db-master/db/base-2/badchecksum.txt - exists and matches backup (11B, 99%) checksum f927212cd08d11a42a666b2f04235398e9ceeb51
P01 DETAIL: restore file [TEST_PATH]/db-master/db/base-2/base/base2.txt - exists and matches backup (9B, 99%) checksum cafac3c59553f2cfde41ce2e62e7662295f108c0
P01 DETAIL: restore file [TEST_PATH]/db-master/db/base-2/base/16384/17000 - exists and matches backup (9B, 99%) checksum 7579ada0808d7f98087a0a586d0df9de009cdc33
Add checksum delta for incremental backups. Use checksums rather than timestamps to determine if files have changed. This is useful in cases where the timestamps may not be trustworthy, e.g. when performing an incremental after failing over to a standby. If checksum delta is enabled then checksums will be used for verification of resumed backups, even if they are full. Resumes have always used checksums to verify the files in the repository, enabling delta performs checksums on the database files as well. Note that the user must manually enable this feature in cases were it would be useful or just keep in enabled all the time. A future commit will address automatically enabling the feature in cases where it seems likely to be useful. Contributed by Cynthia Shang.
2018-09-19 17:12:45 +02:00
P01 DETAIL: restore file [TEST_PATH]/db-master/db/base-2/changecontent.txt - exists and matches backup (7B, 99%) checksum a094d94583e209556d03c3c5da33131a065f1689
For simplicity, the pg_control file is now copied with the rest of the files instead of by itself of at the end of the process. The backup command does not require this behavior and the restore copies to a temporary file which is renamed at the end of the restore.
2016-11-30 21:23:05 +02:00
P01 DETAIL: restore file [TEST_PATH]/db-master/db/base-2/pg_stat/global.stat - exists and matches backup (5B, 99%) checksum e350d5ce0153f3e22d5db21cf2a4eff00f3ee877
Add checksum delta for incremental backups. Use checksums rather than timestamps to determine if files have changed. This is useful in cases where the timestamps may not be trustworthy, e.g. when performing an incremental after failing over to a standby. If checksum delta is enabled then checksums will be used for verification of resumed backups, even if they are full. Resumes have always used checksums to verify the files in the repository, enabling delta performs checksums on the database files as well. Note that the user must manually enable this feature in cases were it would be useful or just keep in enabled all the time. A future commit will address automatically enabling the feature in cases where it seems likely to be useful. Contributed by Cynthia Shang.
2018-09-19 17:12:45 +02:00
P01 DETAIL: restore file [TEST_PATH]/db-master/db/base-2/changetime.txt - exists and matches backup (4B, 99%) checksum 88087292ed82e26f3eb824d0bffc05ccf7a30f8d
For simplicity, the pg_control file is now copied with the rest of the files instead of by itself of at the end of the process. The backup command does not require this behavior and the restore copies to a temporary file which is renamed at the end of the restore.
2016-11-30 21:23:05 +02:00
P01 DETAIL: restore file [TEST_PATH]/db-master/db/base-2/base/32768/PG_VERSION - exists and matches backup (3B, 99%) checksum 184473f470864e067ee3a22e64b47b0a1c356f29
P01 DETAIL: restore file [TEST_PATH]/db-master/db/base-2/base/16384/PG_VERSION - exists and matches backup (3B, 99%) checksum 184473f470864e067ee3a22e64b47b0a1c356f29
P01 DETAIL: restore file [TEST_PATH]/db-master/db/base-2/base/1/PG_VERSION - exists and matches backup (3B, 99%) checksum 184473f470864e067ee3a22e64b47b0a1c356f29
P01 DETAIL: restore file [TEST_PATH]/db-master/db/base-2/PG_VERSION - exists and matches backup (3B, 99%) checksum 184473f470864e067ee3a22e64b47b0a1c356f29
Add zero-length file to mock/all test.
2018-07-05 21:40:50 +02:00
P01 DETAIL: restore file [TEST_PATH]/db-master/db/base-2/zerosize.txt - exists and is zero size (0B, 99%)
Allow zero-size files in backup manifest to reference a prior manifest regardless of timestamp delta. Contributed by Cynthia Shang.
2018-08-24 22:50:33 +02:00
P01 DETAIL: restore file [TEST_PATH]/db-master/db/base-2/zero_from_start - exists and is zero size (0B, 99%)
Use minio for integration testing. ScalityS3 has not received any maintenance in years and is slow to start which is bad for testing. Replace it with minio which starts quickly and ships as a single executable or a tiny container. Minio has stricter limits on allowable characters but should still provide enough coverage to show that our encoding is working correctly. This commit also includes the upgrade to openssl 1.1.1 in the Ubuntu 18.04 container.
2019-07-03 04:20:35 +02:00
P01 DETAIL: restore file [TEST_PATH]/db-master/db/base-2/special-!_.*'()&!@;:+,? - exists and is zero size (0B, 99%)
For simplicity, the pg_control file is now copied with the rest of the files instead of by itself of at the end of the process. The backup command does not require this behavior and the restore copies to a temporary file which is renamed at the end of the restore.
2016-11-30 21:23:05 +02:00
P01 DETAIL: restore zeroed file [TEST_PATH]/db-master/db/base-2/pg_tblspc/2/[TS_PATH-1]/32768/tablespace2c.txt (12B, 99%)
Additional backup exclusions. * Exclude contents of pg_snapshots, pg_serial, pg_notify, and pg_dynshmem from backup since they are rebuilt on startup. * Exclude pg_internal.init files from backup since they are rebuilt on startup.
2017-09-04 14:26:57 +02:00
P01 DETAIL: restore zeroed file [TEST_PATH]/db-master/db/base-2/pg_tblspc/2/[TS_PATH-1]/32768/tablespace2.txt (7B, 100%)
For simplicity, the pg_control file is now copied with the rest of the files instead of by itself of at the end of the process. The backup command does not require this behavior and the restore copies to a temporary file which is renamed at the end of the restore.
2016-11-30 21:23:05 +02:00
P00 INFO: write [TEST_PATH]/db-master/db/base-2/recovery.conf
The restore command is implemented entirely in C. For the most part this is a direct migration of the Perl code into C. There is one important behavioral change with regard to how file permissions are handled. The Perl code tried to set ownership as it was in the manifest even when running as an unprivileged user. This usually just led to errors and frustration. The C code works like this: If a restore is run as a non-root user (the typical scenario) then all files restored will belong to the user/group executing pgBackRest. If existing files are not owned by the executing user/group then an error will result if the ownership cannot be updated to the executing user/group. In that case the file ownership will need to be updated by a privileged user before the restore can be retried. If a restore is run as the root user then pgBackRest will attempt to recreate the ownership recorded in the manifest when the backup was made. Only user/group names are stored in the manifest so the same names must exist on the restore host for this to work. If the user/group name cannot be found locally then the user/group of the PostgreSQL data directory will be used and finally root if the data directory user/group cannot be mapped to a name. Reviewed by Cynthia Shang.
2019-09-26 13:52:02 +02:00
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base-2'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base-2/base'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base-2/base/1'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base-2/base/16384'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base-2/base/32768'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base-2/pg_clog'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base-2/pg_dynshmem'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base-2/pg_log'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base-2/pg_notify'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base-2/pg_replslot'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base-2/pg_serial'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base-2/pg_snapshots'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base-2/pg_stat'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base-2/pg_stat_tmp'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base-2/pg_subtrans'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base-2/pg_tblspc'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base-2/pg_tblspc/2'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base-2/pg_tblspc/2/[TS_PATH-1]'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base-2/pg_tblspc/2/[TS_PATH-1]/32768'
Storage and IO layer refactor: Refactor storage layer to allow for new repository filesystems using drivers. (Reviewed by Cynthia Shang.) Refactor IO layer to allow for new compression formats, checksum types, and other capabilities using filters. (Reviewed by Cynthia Shang.)
2017-06-09 23:51:41 +02:00
P00 INFO: restore global/pg_control (performed last to ensure aborted restores cannot be started)
The restore command is implemented entirely in C. For the most part this is a direct migration of the Perl code into C. There is one important behavioral change with regard to how file permissions are handled. The Perl code tried to set ownership as it was in the manifest even when running as an unprivileged user. This usually just led to errors and frustration. The C code works like this: If a restore is run as a non-root user (the typical scenario) then all files restored will belong to the user/group executing pgBackRest. If existing files are not owned by the executing user/group then an error will result if the ownership cannot be updated to the executing user/group. In that case the file ownership will need to be updated by a privileged user before the restore can be retried. If a restore is run as the root user then pgBackRest will attempt to recreate the ownership recorded in the manifest when the backup was made. Only user/group names are stored in the manifest so the same names must exist on the restore host for this to work. If the user/group name cannot be found locally then the user/group of the PostgreSQL data directory will be used and finally root if the data directory user/group cannot be mapped to a name. Reviewed by Cynthia Shang.
2019-09-26 13:52:02 +02:00
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base-2/global'
Added a predictable log entry to signal that a command has completed successfully. For example a backup ends successfully with: INFO: backup command end: completed successfully.
2016-12-04 21:44:53 +02:00
P00 INFO: restore command end: completed successfully
Allow selective restore of databases from a cluster backup. This feature can result in major space and time savings when only specific databases are restored. Unrestored databases will not be accessible but must be manually dropped before they will be removed from the shared catalogue.
2016-05-11 15:21:39 +02:00
Rename test paths for clarity. This was worked out as part of the test suite refactor [c8f806a] but not committed with it because of the large number of expect logs changes involved. Keeping them separate made it easier to audit the changes in the refactor.
2016-06-24 16:54:31 +02:00
+ supplemental file: [TEST_PATH]/db-master/db/base-2/recovery.conf
------------------------------------------------------------------
PostgreSQL 12 support. Recovery settings are now written into postgresql.auto.conf instead of recovery.conf. Existing recovery_target* settings will be commented out to help avoid conflicts. A comment is added before recovery settings to identify them as written by pgBackRest since it is unclear how, in general, old settings will be removed. recovery.signal and standby.signal are automatically created based on the recovery settings.
2019-10-01 19:20:43 +02:00
# Recovery settings generated by pgBackRest restore on [TIMESTAMP]
Reduce log level for all expect tests to detail. The C code is designed to be efficient rather than deterministic at the debug log level. As we move more testing from integration to unit tests it makes less sense to try and maintain the expect logs at this log level. Most of the expect logs have already been moved to detail level but mock/all still had tests at debug level. Change the logging defaults in the config file and remove as many references to log-level-console as possible.
2019-05-23 00:23:44 +02:00
restore_command = '[BACKREST-BIN] --config=[TEST_PATH]/db-master/pgbackrest.conf --stanza=db archive-get %f "%p"'
Allow selective restore of databases from a cluster backup. This feature can result in major space and time savings when only specific databases are restored. Unrestored databases will not be accessible but must be manually dropped before they will be removed from the shared catalogue.
2016-05-11 15:21:39 +02:00
Major refactor of the test suite. * Make the code more modular and object-oriented. * Multiple Docker containers can now be created for a single test to simulate more realistic environments.
2016-06-24 14:12:58 +02:00
restore delta, remap - selective restore 32768 (db-master host)
Reduce log level for all expect tests to detail. The C code is designed to be efficient rather than deterministic at the debug log level. As we move more testing from integration to unit tests it makes less sense to try and maintain the expect logs at this log level. Most of the expect logs have already been moved to detail level but mock/all still had tests at debug level. Change the logging defaults in the config file and remove as many references to log-level-console as possible.
2019-05-23 00:23:44 +02:00
> [CONTAINER-EXEC] db-master [BACKREST-BIN] --config=[TEST_PATH]/db-master/pgbackrest.conf --delta --db-include=32768 --stanza=db restore
Allow selective restore of databases from a cluster backup. This feature can result in major space and time savings when only specific databases are restored. Unrestored databases will not be accessible but must be manually dropped before they will be removed from the shared catalogue.
2016-05-11 15:21:39 +02:00
------------------------------------------------------------------------------------------------------------------------------------
Remove compress/compress-level options from commands where unused. These commands (e.g. restore, archive-get) never used the compress options but allowed them to be passed on the command line. Now they will error when these options are passed on the command line. If these errors occur then remove the unused options.
2020-02-27 19:25:32 +02:00
P00 INFO: restore command begin [BACKREST-VERSION]: --buffer-size=[BUFFER-SIZE] --config=[TEST_PATH]/db-master/pgbackrest.conf --db-include=32768 --delta --lock-path=[TEST_PATH]/db-master/lock --log-level-console=detail --log-level-file=[LOG-LEVEL-FILE] --log-level-stderr=off --log-path=[TEST_PATH]/db-master/log[] --no-log-timestamp --pg1-path=[TEST_PATH]/db-master/db/base-2 --protocol-timeout=60 --repo1-path=[TEST_PATH]/db-master/repo --stanza=db --tablespace-map=2=[TEST_PATH]/db-master/db/tablespace/ts2-2
Remove integration tests that are now covered in the unit tests. Most of these tests are just checking that errors are thrown when required. These are well covered in various unit tests. The "cannot resume" tests are also well covered in the backup unit tests. Finally, config warnings are well covered in the config unit tests. There is more to be done here, but this accounts for the low-hanging fruit.
2019-12-18 03:14:45 +02:00
P00 INFO: restore backup set [BACKUP-DIFF-4]
The restore command is implemented entirely in C. For the most part this is a direct migration of the Perl code into C. There is one important behavioral change with regard to how file permissions are handled. The Perl code tried to set ownership as it was in the manifest even when running as an unprivileged user. This usually just led to errors and frustration. The C code works like this: If a restore is run as a non-root user (the typical scenario) then all files restored will belong to the user/group executing pgBackRest. If existing files are not owned by the executing user/group then an error will result if the ownership cannot be updated to the executing user/group. In that case the file ownership will need to be updated by a privileged user before the restore can be retried. If a restore is run as the root user then pgBackRest will attempt to recreate the ownership recorded in the manifest when the backup was made. Only user/group names are stored in the manifest so the same names must exist on the restore host for this to work. If the user/group name cannot be found locally then the user/group of the PostgreSQL data directory will be used and finally root if the data directory user/group cannot be mapped to a name. Reviewed by Cynthia Shang.
2019-09-26 13:52:02 +02:00
P00 INFO: map tablespace 'pg_tblspc/2' to '[TEST_PATH]/db-master/db/tablespace/ts2-2'
P00 DETAIL: databases found for selective restore (1, 16384, 32768)
P00 DETAIL: check '[TEST_PATH]/db-master/db/base-2' exists
P00 DETAIL: check '[TEST_PATH]/db-master/db/tablespace/ts2-2/[TS_PATH-1]' exists
P00 DETAIL: remove 'global/pg_control' so cluster will not start if restore does not complete
P00 INFO: remove invalid files/links/paths from '[TEST_PATH]/db-master/db/base-2'
P00 DETAIL: remove invalid file '[TEST_PATH]/db-master/db/base-2/recovery.conf'
P00 INFO: remove invalid files/links/paths from '[TEST_PATH]/db-master/db/tablespace/ts2-2/[TS_PATH-1]'
Exclude temporary and unlogged relation (table/index) files from backup. Implemented using the same logic as the patches adding this feature to PostgreSQL, 8694cc96 and 920a5e50. Temporary relation exclusion is enabled in PostgreSQL ≥ 9.0. Unlogged relation exclusion is enabled in PostgreSQL ≥ 9.1, where the feature was introduced. Contributed by Cynthia Shang.
2018-07-31 00:53:34 +02:00
P01 INFO: restore file [TEST_PATH]/db-master/db/base-2/base/32768/33001 (64KB, 36%) checksum 6bf316f11d28c28914ea9be92c00de9bea6d9a6b
Fix incorrect offline upper bound for ignoring page checksum errors. For offline backups the upper bound was being set to 0x0000FFFF0000FFFF rather than UINT64_MAX. This meant that page checksum errors might be ignored for databases with a lot of past WAL in offline mode. Online mode is not affected since the upper bound is retrieved from pg_start_backup().
2019-07-11 15:13:56 +02:00
P01 INFO: restore file [TEST_PATH]/db-master/db/base-2/base/32768/44000_init (32KB, 54%) checksum 7a16d165e4775f7c92e8cdf60c0af57313f0bf90
P01 INFO: restore file [TEST_PATH]/db-master/db/base-2/base/32768/33000.32767 (32KB, 72%) checksum 6e99b589e550e68e934fd235ccba59fe5b592a9e
P01 INFO: restore file [TEST_PATH]/db-master/db/base-2/base/32768/33000 (32KB, 90%) checksum 7a16d165e4775f7c92e8cdf60c0af57313f0bf90
Support configurable WAL segment size. PostgreSQL 11 introduces configurable WAL segment sizes, from 1MB to 1GB. There are two areas that needed to be updated to support this: building the archive-get queue and checking that WAL has been archived after a backup. Both operations require the WAL segment size to properly build a list. Checking the archive after a backup is still implemented in Perl and has an active database connection, so just get the WAL segment size from the database. The archive-get command does not have a connection to the database, so get the WAL segment size from pg_control instead. This requires a deeper inspection of pg_control than has been done in the past, so it seemed best to copy the relevant data structures from each version of PostgreSQL and build a generic interface layer to address them. While this approach is a bit verbose, it has the advantage of being relatively simple, and can easily be updated for new versions of PostgreSQL. Since the integration tests generate pg_control files for testing, teach Perl how to generate files with the correct offsets for both 32-bit and 64-bit architectures.
2018-09-25 11:24:42 +02:00
P01 INFO: restore file [TEST_PATH]/db-master/db/base-2/global/pg_control.pgbackrest.tmp (8KB, 95%) checksum 4c77c900f7af0d9ab13fa9982051a42e0b637f6c
Add more realistic data files to synthetic backup and restore tests.
2016-12-10 16:13:55 +02:00
P01 DETAIL: restore file [TEST_PATH]/db-master/db/base-2/base/1/12000 - exists and matches backup (8KB, 99%) checksum 22c98d248ff548311eda88559e4a8405ed77c003
For simplicity, the pg_control file is now copied with the rest of the files instead of by itself of at the end of the process. The backup command does not require this behavior and the restore copies to a temporary file which is renamed at the end of the restore.
2016-11-30 21:23:05 +02:00
P01 DETAIL: restore file [TEST_PATH]/db-master/db/base-2/postgresql.conf - exists and matches backup (21B, 99%) checksum 6721d92c9fcdf4248acff1f9a1377127d9064807
P01 DETAIL: restore file [TEST_PATH]/db-master/db/base-2/badchecksum.txt - exists and matches backup (11B, 99%) checksum f927212cd08d11a42a666b2f04235398e9ceeb51
P01 DETAIL: restore file [TEST_PATH]/db-master/db/base-2/base/base2.txt - exists and matches backup (9B, 99%) checksum cafac3c59553f2cfde41ce2e62e7662295f108c0
P01 DETAIL: restore zeroed file [TEST_PATH]/db-master/db/base-2/base/16384/17000 (9B, 99%)
Add checksum delta for incremental backups. Use checksums rather than timestamps to determine if files have changed. This is useful in cases where the timestamps may not be trustworthy, e.g. when performing an incremental after failing over to a standby. If checksum delta is enabled then checksums will be used for verification of resumed backups, even if they are full. Resumes have always used checksums to verify the files in the repository, enabling delta performs checksums on the database files as well. Note that the user must manually enable this feature in cases were it would be useful or just keep in enabled all the time. A future commit will address automatically enabling the feature in cases where it seems likely to be useful. Contributed by Cynthia Shang.
2018-09-19 17:12:45 +02:00
P01 DETAIL: restore file [TEST_PATH]/db-master/db/base-2/changecontent.txt - exists and matches backup (7B, 99%) checksum a094d94583e209556d03c3c5da33131a065f1689
For simplicity, the pg_control file is now copied with the rest of the files instead of by itself of at the end of the process. The backup command does not require this behavior and the restore copies to a temporary file which is renamed at the end of the restore.
2016-11-30 21:23:05 +02:00
P01 DETAIL: restore file [TEST_PATH]/db-master/db/base-2/pg_stat/global.stat - exists and matches backup (5B, 99%) checksum e350d5ce0153f3e22d5db21cf2a4eff00f3ee877
Add checksum delta for incremental backups. Use checksums rather than timestamps to determine if files have changed. This is useful in cases where the timestamps may not be trustworthy, e.g. when performing an incremental after failing over to a standby. If checksum delta is enabled then checksums will be used for verification of resumed backups, even if they are full. Resumes have always used checksums to verify the files in the repository, enabling delta performs checksums on the database files as well. Note that the user must manually enable this feature in cases were it would be useful or just keep in enabled all the time. A future commit will address automatically enabling the feature in cases where it seems likely to be useful. Contributed by Cynthia Shang.
2018-09-19 17:12:45 +02:00
P01 DETAIL: restore file [TEST_PATH]/db-master/db/base-2/changetime.txt - exists and matches backup (4B, 99%) checksum 88087292ed82e26f3eb824d0bffc05ccf7a30f8d
For simplicity, the pg_control file is now copied with the rest of the files instead of by itself of at the end of the process. The backup command does not require this behavior and the restore copies to a temporary file which is renamed at the end of the restore.
2016-11-30 21:23:05 +02:00
P01 DETAIL: restore file [TEST_PATH]/db-master/db/base-2/base/32768/PG_VERSION - exists and matches backup (3B, 99%) checksum 184473f470864e067ee3a22e64b47b0a1c356f29
P01 DETAIL: restore file [TEST_PATH]/db-master/db/base-2/base/16384/PG_VERSION - exists and matches backup (3B, 99%) checksum 184473f470864e067ee3a22e64b47b0a1c356f29
P01 DETAIL: restore file [TEST_PATH]/db-master/db/base-2/base/1/PG_VERSION - exists and matches backup (3B, 99%) checksum 184473f470864e067ee3a22e64b47b0a1c356f29
P01 DETAIL: restore file [TEST_PATH]/db-master/db/base-2/PG_VERSION - exists and matches backup (3B, 99%) checksum 184473f470864e067ee3a22e64b47b0a1c356f29
Add zero-length file to mock/all test.
2018-07-05 21:40:50 +02:00
P01 DETAIL: restore file [TEST_PATH]/db-master/db/base-2/zerosize.txt - exists and is zero size (0B, 99%)
Allow zero-size files in backup manifest to reference a prior manifest regardless of timestamp delta. Contributed by Cynthia Shang.
2018-08-24 22:50:33 +02:00
P01 DETAIL: restore file [TEST_PATH]/db-master/db/base-2/zero_from_start - exists and is zero size (0B, 99%)
Use minio for integration testing. ScalityS3 has not received any maintenance in years and is slow to start which is bad for testing. Replace it with minio which starts quickly and ships as a single executable or a tiny container. Minio has stricter limits on allowable characters but should still provide enough coverage to show that our encoding is working correctly. This commit also includes the upgrade to openssl 1.1.1 in the Ubuntu 18.04 container.
2019-07-03 04:20:35 +02:00
P01 DETAIL: restore file [TEST_PATH]/db-master/db/base-2/special-!_.*'()&!@;:+,? - exists and is zero size (0B, 99%)
For simplicity, the pg_control file is now copied with the rest of the files instead of by itself of at the end of the process. The backup command does not require this behavior and the restore copies to a temporary file which is renamed at the end of the restore.
2016-11-30 21:23:05 +02:00
P01 INFO: restore file [TEST_PATH]/db-master/db/base-2/pg_tblspc/2/[TS_PATH-1]/32768/tablespace2c.txt (12B, 99%) checksum dfcb8679956b734706cf87259d50c88f83e80e66
Additional backup exclusions. * Exclude contents of pg_snapshots, pg_serial, pg_notify, and pg_dynshmem from backup since they are rebuilt on startup. * Exclude pg_internal.init files from backup since they are rebuilt on startup.
2017-09-04 14:26:57 +02:00
P01 INFO: restore file [TEST_PATH]/db-master/db/base-2/pg_tblspc/2/[TS_PATH-1]/32768/tablespace2.txt (7B, 100%) checksum dc7f76e43c46101b47acc55ae4d593a9e6983578
For simplicity, the pg_control file is now copied with the rest of the files instead of by itself of at the end of the process. The backup command does not require this behavior and the restore copies to a temporary file which is renamed at the end of the restore.
2016-11-30 21:23:05 +02:00
P00 INFO: write [TEST_PATH]/db-master/db/base-2/recovery.conf
The restore command is implemented entirely in C. For the most part this is a direct migration of the Perl code into C. There is one important behavioral change with regard to how file permissions are handled. The Perl code tried to set ownership as it was in the manifest even when running as an unprivileged user. This usually just led to errors and frustration. The C code works like this: If a restore is run as a non-root user (the typical scenario) then all files restored will belong to the user/group executing pgBackRest. If existing files are not owned by the executing user/group then an error will result if the ownership cannot be updated to the executing user/group. In that case the file ownership will need to be updated by a privileged user before the restore can be retried. If a restore is run as the root user then pgBackRest will attempt to recreate the ownership recorded in the manifest when the backup was made. Only user/group names are stored in the manifest so the same names must exist on the restore host for this to work. If the user/group name cannot be found locally then the user/group of the PostgreSQL data directory will be used and finally root if the data directory user/group cannot be mapped to a name. Reviewed by Cynthia Shang.
2019-09-26 13:52:02 +02:00
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base-2'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base-2/base'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base-2/base/1'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base-2/base/16384'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base-2/base/32768'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base-2/pg_clog'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base-2/pg_dynshmem'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base-2/pg_log'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base-2/pg_notify'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base-2/pg_replslot'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base-2/pg_serial'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base-2/pg_snapshots'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base-2/pg_stat'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base-2/pg_stat_tmp'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base-2/pg_subtrans'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base-2/pg_tblspc'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base-2/pg_tblspc/2'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base-2/pg_tblspc/2/[TS_PATH-1]'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base-2/pg_tblspc/2/[TS_PATH-1]/32768'
Storage and IO layer refactor: Refactor storage layer to allow for new repository filesystems using drivers. (Reviewed by Cynthia Shang.) Refactor IO layer to allow for new compression formats, checksum types, and other capabilities using filters. (Reviewed by Cynthia Shang.)
2017-06-09 23:51:41 +02:00
P00 INFO: restore global/pg_control (performed last to ensure aborted restores cannot be started)
The restore command is implemented entirely in C. For the most part this is a direct migration of the Perl code into C. There is one important behavioral change with regard to how file permissions are handled. The Perl code tried to set ownership as it was in the manifest even when running as an unprivileged user. This usually just led to errors and frustration. The C code works like this: If a restore is run as a non-root user (the typical scenario) then all files restored will belong to the user/group executing pgBackRest. If existing files are not owned by the executing user/group then an error will result if the ownership cannot be updated to the executing user/group. In that case the file ownership will need to be updated by a privileged user before the restore can be retried. If a restore is run as the root user then pgBackRest will attempt to recreate the ownership recorded in the manifest when the backup was made. Only user/group names are stored in the manifest so the same names must exist on the restore host for this to work. If the user/group name cannot be found locally then the user/group of the PostgreSQL data directory will be used and finally root if the data directory user/group cannot be mapped to a name. Reviewed by Cynthia Shang.
2019-09-26 13:52:02 +02:00
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base-2/global'
Added a predictable log entry to signal that a command has completed successfully. For example a backup ends successfully with: INFO: backup command end: completed successfully.
2016-12-04 21:44:53 +02:00
P00 INFO: restore command end: completed successfully
Allow selective restore of databases from a cluster backup. This feature can result in major space and time savings when only specific databases are restored. Unrestored databases will not be accessible but must be manually dropped before they will be removed from the shared catalogue.
2016-05-11 15:21:39 +02:00
Rename test paths for clarity. This was worked out as part of the test suite refactor [c8f806a] but not committed with it because of the large number of expect logs changes involved. Keeping them separate made it easier to audit the changes in the refactor.
2016-06-24 16:54:31 +02:00
+ supplemental file: [TEST_PATH]/db-master/db/base-2/recovery.conf
------------------------------------------------------------------
PostgreSQL 12 support. Recovery settings are now written into postgresql.auto.conf instead of recovery.conf. Existing recovery_target* settings will be commented out to help avoid conflicts. A comment is added before recovery settings to identify them as written by pgBackRest since it is unclear how, in general, old settings will be removed. recovery.signal and standby.signal are automatically created based on the recovery settings.
2019-10-01 19:20:43 +02:00
# Recovery settings generated by pgBackRest restore on [TIMESTAMP]
Reduce log level for all expect tests to detail. The C code is designed to be efficient rather than deterministic at the debug log level. As we move more testing from integration to unit tests it makes less sense to try and maintain the expect logs at this log level. Most of the expect logs have already been moved to detail level but mock/all still had tests at debug level. Change the logging defaults in the config file and remove as many references to log-level-console as possible.
2019-05-23 00:23:44 +02:00
restore_command = '[BACKREST-BIN] --config=[TEST_PATH]/db-master/pgbackrest.conf --stanza=db archive-get %f "%p"'
Allow selective restore of databases from a cluster backup. This feature can result in major space and time savings when only specific databases are restored. Unrestored databases will not be accessible but must be manually dropped before they will be removed from the shared catalogue.
2016-05-11 15:21:39 +02:00
Renumber return codes between 25 and 125 to avoid PostgreSQL interpreting some as fatal signal exceptions. Suggested by Yogesh Sharma.
2017-04-12 18:41:34 +02:00
restore delta, remap, expect exit 80 - error on invalid id (db-master host)
Rename test paths for clarity. This was worked out as part of the test suite refactor [c8f806a] but not committed with it because of the large number of expect logs changes involved. Keeping them separate made it easier to audit the changes in the refactor.
2016-06-24 16:54:31 +02:00
> [CONTAINER-EXEC] db-master [BACKREST-BIN] --config=[TEST_PATH]/db-master/pgbackrest.conf --delta --log-level-console=warn --db-include=7777 --stanza=db restore
Allow selective restore of databases from a cluster backup. This feature can result in major space and time savings when only specific databases are restored. Unrestored databases will not be accessible but must be manually dropped before they will be removed from the shared catalogue.
2016-05-11 15:21:39 +02:00
------------------------------------------------------------------------------------------------------------------------------------
Renumber return codes between 25 and 125 to avoid PostgreSQL interpreting some as fatal signal exceptions. Suggested by Yogesh Sharma.
2017-04-12 18:41:34 +02:00
P00 ERROR: [080]: database to include '7777' does not exist
Allow selective restore of databases from a cluster backup. This feature can result in major space and time savings when only specific databases are restored. Unrestored databases will not be accessible but must be manually dropped before they will be removed from the shared catalogue.
2016-05-11 15:21:39 +02:00
Renumber return codes between 25 and 125 to avoid PostgreSQL interpreting some as fatal signal exceptions. Suggested by Yogesh Sharma.
2017-04-12 18:41:34 +02:00
restore delta, remap, expect exit 81 - error on system id (db-master host)
Rename test paths for clarity. This was worked out as part of the test suite refactor [c8f806a] but not committed with it because of the large number of expect logs changes involved. Keeping them separate made it easier to audit the changes in the refactor.
2016-06-24 16:54:31 +02:00
> [CONTAINER-EXEC] db-master [BACKREST-BIN] --config=[TEST_PATH]/db-master/pgbackrest.conf --delta --log-level-console=warn --db-include=1 --stanza=db restore
Allow selective restore of databases from a cluster backup. This feature can result in major space and time savings when only specific databases are restored. Unrestored databases will not be accessible but must be manually dropped before they will be removed from the shared catalogue.
2016-05-11 15:21:39 +02:00
------------------------------------------------------------------------------------------------------------------------------------
Renumber return codes between 25 and 125 to avoid PostgreSQL interpreting some as fatal signal exceptions. Suggested by Yogesh Sharma.
2017-04-12 18:41:34 +02:00
P00 ERROR: [081]: system databases (template0, postgres, etc.) are included by default
Allow selective restore of databases from a cluster backup. This feature can result in major space and time savings when only specific databases are restored. Unrestored databases will not be accessible but must be manually dropped before they will be removed from the shared catalogue.
2016-05-11 15:21:39 +02:00
The restore command is implemented entirely in C. For the most part this is a direct migration of the Perl code into C. There is one important behavioral change with regard to how file permissions are handled. The Perl code tried to set ownership as it was in the manifest even when running as an unprivileged user. This usually just led to errors and frustration. The C code works like this: If a restore is run as a non-root user (the typical scenario) then all files restored will belong to the user/group executing pgBackRest. If existing files are not owned by the executing user/group then an error will result if the ownership cannot be updated to the executing user/group. In that case the file ownership will need to be updated by a privileged user before the restore can be retried. If a restore is run as the root user then pgBackRest will attempt to recreate the ownership recorded in the manifest when the backup was made. Only user/group names are stored in the manifest so the same names must exist on the restore host for this to work. If the user/group name cannot be found locally then the user/group of the PostgreSQL data directory will be used and finally root if the data directory user/group cannot be mapped to a name. Reviewed by Cynthia Shang.
2019-09-26 13:52:02 +02:00
restore, remap - no tablespace remap (db-master host)
Reduce log level for all expect tests to detail. The C code is designed to be efficient rather than deterministic at the debug log level. As we move more testing from integration to unit tests it makes less sense to try and maintain the expect logs at this log level. Most of the expect logs have already been moved to detail level but mock/all still had tests at debug level. Change the logging defaults in the config file and remove as many references to log-level-console as possible.
2019-05-23 00:23:44 +02:00
> [CONTAINER-EXEC] db-master [BACKREST-BIN] --config=[TEST_PATH]/db-master/pgbackrest.conf --tablespace-map-all=../../tablespace --stanza=db restore
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
------------------------------------------------------------------------------------------------------------------------------------
Remove compress/compress-level options from commands where unused. These commands (e.g. restore, archive-get) never used the compress options but allowed them to be passed on the command line. Now they will error when these options are passed on the command line. If these errors occur then remove the unused options.
2020-02-27 19:25:32 +02:00
P00 INFO: restore command begin [BACKREST-VERSION]: --buffer-size=[BUFFER-SIZE] --config=[TEST_PATH]/db-master/pgbackrest.conf --lock-path=[TEST_PATH]/db-master/lock --log-level-console=detail --log-level-file=[LOG-LEVEL-FILE] --log-level-stderr=off --log-path=[TEST_PATH]/db-master/log[] --no-log-timestamp --pg1-path=[TEST_PATH]/db-master/db/base-2/base --protocol-timeout=60 --repo1-path=[TEST_PATH]/db-master/repo --stanza=db --tablespace-map-all=../../tablespace
Remove integration tests that are now covered in the unit tests. Most of these tests are just checking that errors are thrown when required. These are well covered in various unit tests. The "cannot resume" tests are also well covered in the backup unit tests. Finally, config warnings are well covered in the config unit tests. There is more to be done here, but this accounts for the low-hanging fruit.
2019-12-18 03:14:45 +02:00
P00 INFO: restore backup set [BACKUP-DIFF-4]
The restore command is implemented entirely in C. For the most part this is a direct migration of the Perl code into C. There is one important behavioral change with regard to how file permissions are handled. The Perl code tried to set ownership as it was in the manifest even when running as an unprivileged user. This usually just led to errors and frustration. The C code works like this: If a restore is run as a non-root user (the typical scenario) then all files restored will belong to the user/group executing pgBackRest. If existing files are not owned by the executing user/group then an error will result if the ownership cannot be updated to the executing user/group. In that case the file ownership will need to be updated by a privileged user before the restore can be retried. If a restore is run as the root user then pgBackRest will attempt to recreate the ownership recorded in the manifest when the backup was made. Only user/group names are stored in the manifest so the same names must exist on the restore host for this to work. If the user/group name cannot be found locally then the user/group of the PostgreSQL data directory will be used and finally root if the data directory user/group cannot be mapped to a name. Reviewed by Cynthia Shang.
2019-09-26 13:52:02 +02:00
P00 INFO: remap data directory to '[TEST_PATH]/db-master/db/base-2/base'
P00 INFO: map tablespace 'pg_tblspc/2' to '../../tablespace/ts2'
P00 DETAIL: check '[TEST_PATH]/db-master/db/base-2/base' exists
P00 DETAIL: check '[TEST_PATH]/db-master/db/base-2/tablespace/ts2/[TS_PATH-1]' exists
P00 DETAIL: update mode for '[TEST_PATH]/db-master/db/base-2/base' to 0777
P00 DETAIL: create path '[TEST_PATH]/db-master/db/base-2/base/base'
P00 DETAIL: create path '[TEST_PATH]/db-master/db/base-2/base/base/1'
P00 DETAIL: create path '[TEST_PATH]/db-master/db/base-2/base/base/16384'
P00 DETAIL: create path '[TEST_PATH]/db-master/db/base-2/base/base/32768'
P00 DETAIL: create path '[TEST_PATH]/db-master/db/base-2/base/global'
P00 DETAIL: create path '[TEST_PATH]/db-master/db/base-2/base/pg_clog'
P00 DETAIL: create path '[TEST_PATH]/db-master/db/base-2/base/pg_dynshmem'
P00 DETAIL: create path '[TEST_PATH]/db-master/db/base-2/base/pg_log'
P00 DETAIL: create path '[TEST_PATH]/db-master/db/base-2/base/pg_notify'
P00 DETAIL: create path '[TEST_PATH]/db-master/db/base-2/base/pg_replslot'
P00 DETAIL: create path '[TEST_PATH]/db-master/db/base-2/base/pg_serial'
P00 DETAIL: create path '[TEST_PATH]/db-master/db/base-2/base/pg_snapshots'
P00 DETAIL: create path '[TEST_PATH]/db-master/db/base-2/base/pg_stat'
P00 DETAIL: create path '[TEST_PATH]/db-master/db/base-2/base/pg_stat_tmp'
P00 DETAIL: create path '[TEST_PATH]/db-master/db/base-2/base/pg_subtrans'
P00 DETAIL: create path '[TEST_PATH]/db-master/db/base-2/base/pg_tblspc'
P00 DETAIL: create symlink '[TEST_PATH]/db-master/db/base-2/base/pg_tblspc/2' to '../../tablespace/ts2'
P00 DETAIL: create path '[TEST_PATH]/db-master/db/base-2/base/pg_tblspc/2/[TS_PATH-1]/32768'
Exclude temporary and unlogged relation (table/index) files from backup. Implemented using the same logic as the patches adding this feature to PostgreSQL, 8694cc96 and 920a5e50. Temporary relation exclusion is enabled in PostgreSQL ≥ 9.0. Unlogged relation exclusion is enabled in PostgreSQL ≥ 9.1, where the feature was introduced. Contributed by Cynthia Shang.
2018-07-31 00:53:34 +02:00
P01 INFO: restore file [TEST_PATH]/db-master/db/base-2/base/base/32768/33001 (64KB, 36%) checksum 6bf316f11d28c28914ea9be92c00de9bea6d9a6b
Fix incorrect offline upper bound for ignoring page checksum errors. For offline backups the upper bound was being set to 0x0000FFFF0000FFFF rather than UINT64_MAX. This meant that page checksum errors might be ignored for databases with a lot of past WAL in offline mode. Online mode is not affected since the upper bound is retrieved from pg_start_backup().
2019-07-11 15:13:56 +02:00
P01 INFO: restore file [TEST_PATH]/db-master/db/base-2/base/base/32768/44000_init (32KB, 54%) checksum 7a16d165e4775f7c92e8cdf60c0af57313f0bf90
P01 INFO: restore file [TEST_PATH]/db-master/db/base-2/base/base/32768/33000.32767 (32KB, 72%) checksum 6e99b589e550e68e934fd235ccba59fe5b592a9e
P01 INFO: restore file [TEST_PATH]/db-master/db/base-2/base/base/32768/33000 (32KB, 90%) checksum 7a16d165e4775f7c92e8cdf60c0af57313f0bf90
Support configurable WAL segment size. PostgreSQL 11 introduces configurable WAL segment sizes, from 1MB to 1GB. There are two areas that needed to be updated to support this: building the archive-get queue and checking that WAL has been archived after a backup. Both operations require the WAL segment size to properly build a list. Checking the archive after a backup is still implemented in Perl and has an active database connection, so just get the WAL segment size from the database. The archive-get command does not have a connection to the database, so get the WAL segment size from pg_control instead. This requires a deeper inspection of pg_control than has been done in the past, so it seemed best to copy the relevant data structures from each version of PostgreSQL and build a generic interface layer to address them. While this approach is a bit verbose, it has the advantage of being relatively simple, and can easily be updated for new versions of PostgreSQL. Since the integration tests generate pg_control files for testing, teach Perl how to generate files with the correct offsets for both 32-bit and 64-bit architectures.
2018-09-25 11:24:42 +02:00
P01 INFO: restore file [TEST_PATH]/db-master/db/base-2/base/global/pg_control.pgbackrest.tmp (8KB, 95%) checksum 4c77c900f7af0d9ab13fa9982051a42e0b637f6c
Add more realistic data files to synthetic backup and restore tests.
2016-12-10 16:13:55 +02:00
P01 INFO: restore file [TEST_PATH]/db-master/db/base-2/base/base/1/12000 (8KB, 99%) checksum 22c98d248ff548311eda88559e4a8405ed77c003
For simplicity, the pg_control file is now copied with the rest of the files instead of by itself of at the end of the process. The backup command does not require this behavior and the restore copies to a temporary file which is renamed at the end of the restore.
2016-11-30 21:23:05 +02:00
P01 INFO: restore file [TEST_PATH]/db-master/db/base-2/base/postgresql.conf (21B, 99%) checksum 6721d92c9fcdf4248acff1f9a1377127d9064807
P01 INFO: restore file [TEST_PATH]/db-master/db/base-2/base/badchecksum.txt (11B, 99%) checksum f927212cd08d11a42a666b2f04235398e9ceeb51
P01 INFO: restore file [TEST_PATH]/db-master/db/base-2/base/base/base2.txt (9B, 99%) checksum cafac3c59553f2cfde41ce2e62e7662295f108c0
P01 INFO: restore file [TEST_PATH]/db-master/db/base-2/base/base/16384/17000 (9B, 99%) checksum 7579ada0808d7f98087a0a586d0df9de009cdc33
Add checksum delta for incremental backups. Use checksums rather than timestamps to determine if files have changed. This is useful in cases where the timestamps may not be trustworthy, e.g. when performing an incremental after failing over to a standby. If checksum delta is enabled then checksums will be used for verification of resumed backups, even if they are full. Resumes have always used checksums to verify the files in the repository, enabling delta performs checksums on the database files as well. Note that the user must manually enable this feature in cases were it would be useful or just keep in enabled all the time. A future commit will address automatically enabling the feature in cases where it seems likely to be useful. Contributed by Cynthia Shang.
2018-09-19 17:12:45 +02:00
P01 INFO: restore file [TEST_PATH]/db-master/db/base-2/base/changecontent.txt (7B, 99%) checksum a094d94583e209556d03c3c5da33131a065f1689
For simplicity, the pg_control file is now copied with the rest of the files instead of by itself of at the end of the process. The backup command does not require this behavior and the restore copies to a temporary file which is renamed at the end of the restore.
2016-11-30 21:23:05 +02:00
P01 INFO: restore file [TEST_PATH]/db-master/db/base-2/base/pg_stat/global.stat (5B, 99%) checksum e350d5ce0153f3e22d5db21cf2a4eff00f3ee877
Add checksum delta for incremental backups. Use checksums rather than timestamps to determine if files have changed. This is useful in cases where the timestamps may not be trustworthy, e.g. when performing an incremental after failing over to a standby. If checksum delta is enabled then checksums will be used for verification of resumed backups, even if they are full. Resumes have always used checksums to verify the files in the repository, enabling delta performs checksums on the database files as well. Note that the user must manually enable this feature in cases were it would be useful or just keep in enabled all the time. A future commit will address automatically enabling the feature in cases where it seems likely to be useful. Contributed by Cynthia Shang.
2018-09-19 17:12:45 +02:00
P01 INFO: restore file [TEST_PATH]/db-master/db/base-2/base/changetime.txt (4B, 99%) checksum 88087292ed82e26f3eb824d0bffc05ccf7a30f8d
For simplicity, the pg_control file is now copied with the rest of the files instead of by itself of at the end of the process. The backup command does not require this behavior and the restore copies to a temporary file which is renamed at the end of the restore.
2016-11-30 21:23:05 +02:00
P01 INFO: restore file [TEST_PATH]/db-master/db/base-2/base/base/32768/PG_VERSION (3B, 99%) checksum 184473f470864e067ee3a22e64b47b0a1c356f29
P01 INFO: restore file [TEST_PATH]/db-master/db/base-2/base/base/16384/PG_VERSION (3B, 99%) checksum 184473f470864e067ee3a22e64b47b0a1c356f29
P01 INFO: restore file [TEST_PATH]/db-master/db/base-2/base/base/1/PG_VERSION (3B, 99%) checksum 184473f470864e067ee3a22e64b47b0a1c356f29
P01 INFO: restore file [TEST_PATH]/db-master/db/base-2/base/PG_VERSION (3B, 99%) checksum 184473f470864e067ee3a22e64b47b0a1c356f29
Add zero-length file to mock/all test.
2018-07-05 21:40:50 +02:00
P01 INFO: restore file [TEST_PATH]/db-master/db/base-2/base/zerosize.txt (0B, 99%)
Allow zero-size files in backup manifest to reference a prior manifest regardless of timestamp delta. Contributed by Cynthia Shang.
2018-08-24 22:50:33 +02:00
P01 INFO: restore file [TEST_PATH]/db-master/db/base-2/base/zero_from_start (0B, 99%)
Use minio for integration testing. ScalityS3 has not received any maintenance in years and is slow to start which is bad for testing. Replace it with minio which starts quickly and ships as a single executable or a tiny container. Minio has stricter limits on allowable characters but should still provide enough coverage to show that our encoding is working correctly. This commit also includes the upgrade to openssl 1.1.1 in the Ubuntu 18.04 container.
2019-07-03 04:20:35 +02:00
P01 INFO: restore file [TEST_PATH]/db-master/db/base-2/base/special-!_.*'()&!@;:+,? (0B, 99%)
For simplicity, the pg_control file is now copied with the rest of the files instead of by itself of at the end of the process. The backup command does not require this behavior and the restore copies to a temporary file which is renamed at the end of the restore.
2016-11-30 21:23:05 +02:00
P01 INFO: restore file [TEST_PATH]/db-master/db/base-2/base/pg_tblspc/2/[TS_PATH-1]/32768/tablespace2c.txt (12B, 99%) checksum dfcb8679956b734706cf87259d50c88f83e80e66
Additional backup exclusions. * Exclude contents of pg_snapshots, pg_serial, pg_notify, and pg_dynshmem from backup since they are rebuilt on startup. * Exclude pg_internal.init files from backup since they are rebuilt on startup.
2017-09-04 14:26:57 +02:00
P01 INFO: restore file [TEST_PATH]/db-master/db/base-2/base/pg_tblspc/2/[TS_PATH-1]/32768/tablespace2.txt (7B, 100%) checksum dc7f76e43c46101b47acc55ae4d593a9e6983578
For simplicity, the pg_control file is now copied with the rest of the files instead of by itself of at the end of the process. The backup command does not require this behavior and the restore copies to a temporary file which is renamed at the end of the restore.
2016-11-30 21:23:05 +02:00
P00 INFO: write [TEST_PATH]/db-master/db/base-2/base/recovery.conf
The restore command is implemented entirely in C. For the most part this is a direct migration of the Perl code into C. There is one important behavioral change with regard to how file permissions are handled. The Perl code tried to set ownership as it was in the manifest even when running as an unprivileged user. This usually just led to errors and frustration. The C code works like this: If a restore is run as a non-root user (the typical scenario) then all files restored will belong to the user/group executing pgBackRest. If existing files are not owned by the executing user/group then an error will result if the ownership cannot be updated to the executing user/group. In that case the file ownership will need to be updated by a privileged user before the restore can be retried. If a restore is run as the root user then pgBackRest will attempt to recreate the ownership recorded in the manifest when the backup was made. Only user/group names are stored in the manifest so the same names must exist on the restore host for this to work. If the user/group name cannot be found locally then the user/group of the PostgreSQL data directory will be used and finally root if the data directory user/group cannot be mapped to a name. Reviewed by Cynthia Shang.
2019-09-26 13:52:02 +02:00
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base-2/base'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base-2/base/base'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base-2/base/base/1'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base-2/base/base/16384'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base-2/base/base/32768'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base-2/base/pg_clog'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base-2/base/pg_dynshmem'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base-2/base/pg_log'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base-2/base/pg_notify'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base-2/base/pg_replslot'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base-2/base/pg_serial'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base-2/base/pg_snapshots'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base-2/base/pg_stat'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base-2/base/pg_stat_tmp'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base-2/base/pg_subtrans'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base-2/base/pg_tblspc'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base-2/base/pg_tblspc/2'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base-2/base/pg_tblspc/2/[TS_PATH-1]'
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base-2/base/pg_tblspc/2/[TS_PATH-1]/32768'
Storage and IO layer refactor: Refactor storage layer to allow for new repository filesystems using drivers. (Reviewed by Cynthia Shang.) Refactor IO layer to allow for new compression formats, checksum types, and other capabilities using filters. (Reviewed by Cynthia Shang.)
2017-06-09 23:51:41 +02:00
P00 INFO: restore global/pg_control (performed last to ensure aborted restores cannot be started)
The restore command is implemented entirely in C. For the most part this is a direct migration of the Perl code into C. There is one important behavioral change with regard to how file permissions are handled. The Perl code tried to set ownership as it was in the manifest even when running as an unprivileged user. This usually just led to errors and frustration. The C code works like this: If a restore is run as a non-root user (the typical scenario) then all files restored will belong to the user/group executing pgBackRest. If existing files are not owned by the executing user/group then an error will result if the ownership cannot be updated to the executing user/group. In that case the file ownership will need to be updated by a privileged user before the restore can be retried. If a restore is run as the root user then pgBackRest will attempt to recreate the ownership recorded in the manifest when the backup was made. Only user/group names are stored in the manifest so the same names must exist on the restore host for this to work. If the user/group name cannot be found locally then the user/group of the PostgreSQL data directory will be used and finally root if the data directory user/group cannot be mapped to a name. Reviewed by Cynthia Shang.
2019-09-26 13:52:02 +02:00
P00 DETAIL: sync path '[TEST_PATH]/db-master/db/base-2/base/global'
Added a predictable log entry to signal that a command has completed successfully. For example a backup ends successfully with: INFO: backup command end: completed successfully.
2016-12-04 21:44:53 +02:00
P00 INFO: restore command end: completed successfully
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
Rename test paths for clarity. This was worked out as part of the test suite refactor [c8f806a] but not committed with it because of the large number of expect logs changes involved. Keeping them separate made it easier to audit the changes in the refactor.
2016-06-24 16:54:31 +02:00
+ supplemental file: [TEST_PATH]/db-master/db/base-2/base/recovery.conf
-----------------------------------------------------------------------
PostgreSQL 12 support. Recovery settings are now written into postgresql.auto.conf instead of recovery.conf. Existing recovery_target* settings will be commented out to help avoid conflicts. A comment is added before recovery settings to identify them as written by pgBackRest since it is unclear how, in general, old settings will be removed. recovery.signal and standby.signal are automatically created based on the recovery settings.
2019-10-01 19:20:43 +02:00
# Recovery settings generated by pgBackRest restore on [TIMESTAMP]
Reduce log level for all expect tests to detail. The C code is designed to be efficient rather than deterministic at the debug log level. As we move more testing from integration to unit tests it makes less sense to try and maintain the expect logs at this log level. Most of the expect logs have already been moved to detail level but mock/all still had tests at debug level. Change the logging defaults in the config file and remove as many references to log-level-console as possible.
2019-05-23 00:23:44 +02:00
restore_command = '[BACKREST-BIN] --config=[TEST_PATH]/db-master/pgbackrest.conf --stanza=db archive-get %f "%p"'
Implemented issue #26: Info command. * Includes updating the manifest to format 4. It turns out the manifest and .info files were not very good for providing information. A format update was required anyway so worked through the backlog of changes that would require a format change. * Multiple database versions are now supported in the archive. Does't actually work yet but the structure should be good. * Tests use more constants now that test logs can catch name regressions.
2015-06-14 00:25:49 +02:00
Make ls ordering deterministic in mock/all test.
2018-07-02 02:02:37 +02:00
> ls -1Rtr [TEST_PATH]/db-master/repo/backup/db/backup.history
Fix issue #137: Info command throws an exception if no stanzas exist.
2015-09-08 22:44:47 +02:00
------------------------------------------------------------------------------------------------------------------------------------
Rename test paths for clarity. This was worked out as part of the test suite refactor [c8f806a] but not committed with it because of the large number of expect logs changes involved. Keeping them separate made it easier to audit the changes in the refactor.
2016-06-24 16:54:31 +02:00
[TEST_PATH]/db-master/repo/backup/db/backup.history:
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
[YEAR-1]
Rename test paths for clarity. This was worked out as part of the test suite refactor [c8f806a] but not committed with it because of the large number of expect logs changes involved. Keeping them separate made it easier to audit the changes in the refactor.
2016-06-24 16:54:31 +02:00
[TEST_PATH]/db-master/repo/backup/db/backup.history/[YEAR-1]:
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
[BACKUP-FULL-1].manifest.gz
Make ls ordering deterministic in mock/all test.
2018-07-02 02:02:37 +02:00
[BACKUP-FULL-2].manifest.gz
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
[BACKUP-INCR-1].manifest.gz
[BACKUP-INCR-2].manifest.gz
[BACKUP-DIFF-1].manifest.gz
[BACKUP-INCR-3].manifest.gz
[BACKUP-INCR-4].manifest.gz
Remove integration tests that are now covered in the unit tests. Most of these tests are just checking that errors are thrown when required. These are well covered in various unit tests. The "cannot resume" tests are also well covered in the backup unit tests. Finally, config warnings are well covered in the config unit tests. There is more to be done here, but this accounts for the low-hanging fruit.
2019-12-18 03:14:45 +02:00
[BACKUP-DIFF-2].manifest.gz
New manifest format. * All files and directories linked from PGDATA are now included in the backup. By default links will be restored directly into PGDATA as files or directories. The --link-all option can be used to restore all links to their original locations. The --link-map option can be used to remap a link to a new location. * Removed --tablespace option and replaced with --tablespace-map-all option which should more clearly indicate its function. * Added detail log level which will output more information than info without being as verbose as debug.
2016-04-15 04:50:02 +02:00
[BACKUP-DIFF-3].manifest.gz
Major refactor of the test suite. * Make the code more modular and object-oriented. * Multiple Docker containers can now be created for a single test to simulate more realistic environments.
2016-06-24 14:12:58 +02:00
[BACKUP-FULL-3].manifest.gz
Remove integration tests that are now covered in the unit tests. Most of these tests are just checking that errors are thrown when required. These are well covered in various unit tests. The "cannot resume" tests are also well covered in the backup unit tests. Finally, config warnings are well covered in the config unit tests. There is more to be done here, but this accounts for the low-hanging fruit.
2019-12-18 03:14:45 +02:00
[BACKUP-DIFF-4].manifest.gz
Improve performance of info file load/save. Info files required three copies in memory to be loaded (the original string, an ini representation, and the final info object). Not only was this memory inefficient but the Ini object does sequential scans when searching for keys making large files very slow to load. This has not been an issue since archive.info and backup.info are very small, but it becomes a big deal when loading manifests with hundreds of thousands of files. Instead of holding copies of the data in memory, use a callback to deliver the ini data directly to the object when loading. Use a similar method for save to avoid having an intermediate copy. Save is a bit complex because sections/keys must be written in alpha order or older versions of pgBackRest will not calculate the correct checksum. Also move the load retry logic to helper functions rather than embedding it in the Info object. This allows for more flexibility in loading and ensures that stack traces will be available when developing unit tests. Reviewed by Cynthia Shang.
2019-09-06 19:48:28 +02:00
Fixed the backup command so the backup-standby option is reset (and the backup proceeds on the master) if the standby is not configured and/or reachable. Fixed by Cynthia Shang.
2017-05-12 22:05:44 +02:00
diff backup - option backup-standby reset - backup performed from master (db-master host)
Full/Synthetic test refactor. * Combine hardlink and non/compressed in synthetic tests to reduce test time and improve coverage. * Change log level of hardlink logging to detail. * Cast size in S3 manifest to integer.
2017-06-15 21:32:10 +02:00
> [CONTAINER-EXEC] db-master [BACKREST-BIN] --config=[TEST_PATH]/db-master/pgbackrest.conf --no-online --log-level-console=info --backup-standby --type=diff --stanza=db backup
Fixed the backup command so the backup-standby option is reset (and the backup proceeds on the master) if the standby is not configured and/or reachable. Fixed by Cynthia Shang.
2017-05-12 22:05:44 +02:00
------------------------------------------------------------------------------------------------------------------------------------
Add Zstandard compression support. Zstandard is a fast lossless compression algorithm targeting real-time compression scenarios at zlib-level and better compression ratios. It's backed by a very fast entropy stage, provided by Huff0 and FSE library. Zstandard version >= 1.0 is required, which is generally only available on newer distributions.
2020-05-04 21:25:27 +02:00
P00 INFO: backup command begin [BACKREST-VERSION]: --backup-standby --buffer-size=[BUFFER-SIZE] --compress-level=3 --compress-type=zst --config=[TEST_PATH]/db-master/pgbackrest.conf --db-timeout=45 --lock-path=[TEST_PATH]/db-master/lock --log-level-console=info --log-level-file=[LOG-LEVEL-FILE] --log-level-stderr=off --log-path=[TEST_PATH]/db-master/log[] --no-log-timestamp --no-online --pg1-path=[TEST_PATH]/db-master/db/base-2/base --protocol-timeout=60 --repo1-hardlink --repo1-path=[TEST_PATH]/db-master/repo --stanza=db --start-fast --type=diff
Add time-based retention for full backups. The --repo-retention-full-type option allows retention of full backups based on a time period, specified in days. The new option will default to 'count' and therefore will not affect current installations. Setting repo-retention-full-type to 'time' will allow the user to use a time period, in days, to indicate full backup retention. Using this method, a full backup can be expired only if the time the backup completed is older than the number of days set with repo-retention-full (calculated from the moment the 'expire' command is run) and at least one full backup meets the retention period. If archive retention has not been configured, then the default settings will expire archives that are prior to the oldest retained full backup. For example, if there are three full backups ending in times that are 25 days old (F1), 20 days old (F2) and 10 days old (F3), then if the full retention period is 15 days, then only F1 will be expired; F2 will be retained because F1 is not at least 15 days old.
2020-05-08 21:25:03 +02:00
P00 WARN: option 'repo1-retention-full' is not set for 'repo1-retention-full-type=count', the repository may run out of space
Rename retention-* options to repo-retention-*.
2018-02-19 22:07:24 +02:00
HINT: to retain full backups indefinitely (without warning), set option 'repo1-retention-full' to the maximum.
The backup command is implemented entirely in C. For the most part this is a direct migration of the Perl code into C except as noted below. A backup can now be initiated from a linked directory. The link will not be stored in the manifest or recreated on restore. If a link or directory does not already exist in the restore location then a directory will be created. The logic for creating backup labels has been improved and it should no longer be possible to get a backup label earlier than the latest backup even with timezone changes or clock skew. This has never been an issue in the field that we know of, but we found it in testing. For online backups all times are fetched from the PostgreSQL primary host (before only copy start was). This doesn't affect backup integrity but it does prevent clock skew between hosts affecting backup duration reporting. Archive copy now works as expected when the archive and backup have different compression settings, i.e. when one is compressed and the other is not. This was a long-standing bug in the Perl code. Resume will now work even if hardlink settings have been changed. Reviewed by Cynthia Shang.
2019-12-14 00:14:26 +02:00
P00 WARN: option backup-standby is enabled but backup is offline - backups will be performed from the primary
Fixed the backup command so the backup-standby option is reset (and the backup proceeds on the master) if the standby is not configured and/or reachable. Fixed by Cynthia Shang.
2017-05-12 22:05:44 +02:00
P00 INFO: last backup label = [BACKUP-FULL-3], version = [VERSION-1]
P01 INFO: backup file [TEST_PATH]/db-master/db/base-2/base/base/base2.txt (9B, 100%) checksum cafac3c59553f2cfde41ce2e62e7662295f108c0
P00 INFO: diff backup size = 9B
Remove integration tests that are now covered in the unit tests. Most of these tests are just checking that errors are thrown when required. These are well covered in various unit tests. The "cannot resume" tests are also well covered in the backup unit tests. Finally, config warnings are well covered in the config unit tests. There is more to be done here, but this accounts for the low-hanging fruit.
2019-12-18 03:14:45 +02:00
P00 INFO: new backup label = [BACKUP-DIFF-5]
Fixed the backup command so the backup-standby option is reset (and the backup proceeds on the master) if the standby is not configured and/or reachable. Fixed by Cynthia Shang.
2017-05-12 22:05:44 +02:00
P00 INFO: backup command end: completed successfully
Integration test improvements for disk and memory efficiency. Set log-level-file=off when more that one test will run. In this case is it impossible to see the logs anyway since they will be automatically cleaned up after the test. This improves performance pretty dramatically since trace-level logging is expensive. If a singe integration test is run then log-level-file is trace by default but can be changed with the --log-level-test-file option. Reduce buffer-size to 64k to save memory during testing and allow more processes to run in parallel. Update log replacement rules so that these options can change without affecting expect logs.
2019-12-17 22:23:07 +02:00
P00 INFO: expire command begin [BACKREST-VERSION]: --buffer-size=[BUFFER-SIZE] --config=[TEST_PATH]/db-master/pgbackrest.conf --lock-path=[TEST_PATH]/db-master/lock --log-level-console=info --log-level-file=[LOG-LEVEL-FILE] --log-level-stderr=off --log-path=[TEST_PATH]/db-master/log[] --no-log-timestamp --repo1-path=[TEST_PATH]/db-master/repo --stanza=db
Rename retention-* options to repo-retention-*.
2018-02-19 22:07:24 +02:00
P00 INFO: option 'repo1-retention-archive' is not set - archive logs will not be expired
Fixed the backup command so the backup-standby option is reset (and the backup proceeds on the master) if the standby is not configured and/or reachable. Fixed by Cynthia Shang.
2017-05-12 22:05:44 +02:00
P00 INFO: expire command end: completed successfully
+ supplemental file: [TEST_PATH]/db-master/pgbackrest.conf
----------------------------------------------------------
[db]
Rename db-* options to pg-* and backup-* options to repo-* to improve consistency. * repo-* options are now indexed although only one is allowed. * List deprecated option names in documentation and command-line help.
2018-02-04 01:27:38 +02:00
pg1-path=[TEST_PATH]/db-master/db/base-2/base
Fixed the backup command so the backup-standby option is reset (and the backup proceeds on the master) if the standby is not configured and/or reachable. Fixed by Cynthia Shang.
2017-05-12 22:05:44 +02:00
[db:restore]
[global]
Integration test improvements for disk and memory efficiency. Set log-level-file=off when more that one test will run. In this case is it impossible to see the logs anyway since they will be automatically cleaned up after the test. This improves performance pretty dramatically since trace-level logging is expensive. If a singe integration test is run then log-level-file is trace by default but can be changed with the --log-level-test-file option. Reduce buffer-size to 64k to save memory during testing and allow more processes to run in parallel. Update log replacement rules so that these options can change without affecting expect logs.
2019-12-17 22:23:07 +02:00
buffer-size=[BUFFER-SIZE]
Fix custom settings for compress-level option being ignored. Reported by Jens Wilke.
2017-10-23 00:05:46 +02:00
compress-level=3
Add Zstandard compression support. Zstandard is a fast lossless compression algorithm targeting real-time compression scenarios at zlib-level and better compression ratios. It's backed by a very fast entropy stage, provided by Huff0 and FSE library. Zstandard version >= 1.0 is required, which is generally only available on newer distributions.
2020-05-04 21:25:27 +02:00
compress-type=zst
Storage and IO layer refactor: Refactor storage layer to allow for new repository filesystems using drivers. (Reviewed by Cynthia Shang.) Refactor IO layer to allow for new compression formats, checksum types, and other capabilities using filters. (Reviewed by Cynthia Shang.)
2017-06-09 23:51:41 +02:00
db-timeout=45
lock-path=[TEST_PATH]/db-master/lock
Reduce log level for all expect tests to detail. The C code is designed to be efficient rather than deterministic at the debug log level. As we move more testing from integration to unit tests it makes less sense to try and maintain the expect logs at this log level. Most of the expect logs have already been moved to detail level but mock/all still had tests at debug level. Change the logging defaults in the config file and remove as many references to log-level-console as possible.
2019-05-23 00:23:44 +02:00
log-level-console=detail
Integration test improvements for disk and memory efficiency. Set log-level-file=off when more that one test will run. In this case is it impossible to see the logs anyway since they will be automatically cleaned up after the test. This improves performance pretty dramatically since trace-level logging is expensive. If a singe integration test is run then log-level-file is trace by default but can be changed with the --log-level-test-file option. Reduce buffer-size to 64k to save memory during testing and allow more processes to run in parallel. Update log replacement rules so that these options can change without affecting expect logs.
2019-12-17 22:23:07 +02:00
log-level-file=[LOG-LEVEL-FILE]
Fixed the backup command so the backup-standby option is reset (and the backup proceeds on the master) if the standby is not configured and/or reachable. Fixed by Cynthia Shang.
2017-05-12 22:05:44 +02:00
log-level-stderr=off
Storage and IO layer refactor: Refactor storage layer to allow for new repository filesystems using drivers. (Reviewed by Cynthia Shang.) Refactor IO layer to allow for new compression formats, checksum types, and other capabilities using filters. (Reviewed by Cynthia Shang.)
2017-06-09 23:51:41 +02:00
log-path=[TEST_PATH]/db-master/log
Integration test improvements for disk and memory efficiency. Set log-level-file=off when more that one test will run. In this case is it impossible to see the logs anyway since they will be automatically cleaned up after the test. This improves performance pretty dramatically since trace-level logging is expensive. If a singe integration test is run then log-level-file is trace by default but can be changed with the --log-level-test-file option. Reduce buffer-size to 64k to save memory during testing and allow more processes to run in parallel. Update log replacement rules so that these options can change without affecting expect logs.
2019-12-17 22:23:07 +02:00
log-subprocess=[LOG-SUBPROCESS]
Set log-timestamp=n for integration tests. This means less filtering of logs needs to be done and new timestamps can be added without adding new filters.
2018-05-11 17:24:38 +02:00
log-timestamp=n
Storage and IO layer refactor: Refactor storage layer to allow for new repository filesystems using drivers. (Reviewed by Cynthia Shang.) Refactor IO layer to allow for new compression formats, checksum types, and other capabilities using filters. (Reviewed by Cynthia Shang.)
2017-06-09 23:51:41 +02:00
protocol-timeout=60
Rename db-* options to pg-* and backup-* options to repo-* to improve consistency. * repo-* options are now indexed although only one is allowed. * List deprecated option names in documentation and command-line help.
2018-02-04 01:27:38 +02:00
repo1-hardlink=y
repo1-path=[TEST_PATH]/db-master/repo
Consolidate archive-push tests into new archive test.
2017-06-21 14:38:53 +02:00
spool-path=[TEST_PATH]/db-master/spool
Fixed the backup command so the backup-standby option is reset (and the backup proceeds on the master) if the standby is not configured and/or reachable. Fixed by Cynthia Shang.
2017-05-12 22:05:44 +02:00
[global:backup]
archive-copy=y
start-fast=y
Remove integration tests that are now covered in the unit tests. Most of these tests are just checking that errors are thrown when required. These are well covered in various unit tests. The "cannot resume" tests are also well covered in the backup unit tests. Finally, config warnings are well covered in the config unit tests. There is more to be done here, but this accounts for the low-hanging fruit.
2019-12-18 03:14:45 +02:00
+ supplemental file: [TEST_PATH]/db-master/repo/backup/db/[BACKUP-DIFF-5]/backup.manifest
Fixed the backup command so the backup-standby option is reset (and the backup proceeds on the master) if the standby is not configured and/or reachable. Fixed by Cynthia Shang.
2017-05-12 22:05:44 +02:00
-----------------------------------------------------------------------------------------
[backrest]
backrest-format=5
backrest-version="[VERSION-1]"
[backup]
Remove integration tests that are now covered in the unit tests. Most of these tests are just checking that errors are thrown when required. These are well covered in various unit tests. The "cannot resume" tests are also well covered in the backup unit tests. Finally, config warnings are well covered in the config unit tests. There is more to be done here, but this accounts for the low-hanging fruit.
2019-12-18 03:14:45 +02:00
backup-label="[BACKUP-DIFF-5]"
Fixed the backup command so the backup-standby option is reset (and the backup proceeds on the master) if the standby is not configured and/or reachable. Fixed by Cynthia Shang.
2017-05-12 22:05:44 +02:00
backup-prior="[BACKUP-FULL-3]"
backup-timestamp-copy-start=[TIMESTAMP]
backup-timestamp-start=[TIMESTAMP]
backup-timestamp-stop=[TIMESTAMP]
backup-type="diff"
[backup:db]
db-catalog-version=201409291
db-control-version=942
db-id=1
Automate generation of WAL and pg_control test files. The existing static files would not work with 32-bit or big-endian systems so create functions to generate these files dynamically rather than creating a bunch of new static files.
2017-11-19 03:02:54 +02:00
db-system-id=1000000000000000094
Fixed the backup command so the backup-standby option is reset (and the backup proceeds on the master) if the standby is not configured and/or reachable. Fixed by Cynthia Shang.
2017-05-12 22:05:44 +02:00
db-version="9.4"
[backup:option]
Set archive-check option in manifest correctly when offline. Archive check does not run when in offline backup mode but the option was set to true in the manifest. It's harmless since these options are informational only but it could cause confusion when debugging.
2019-11-28 15:27:21 +02:00
option-archive-check=false
Fixed the backup command so the backup-standby option is reset (and the backup proceeds on the master) if the standby is not configured and/or reachable. Fixed by Cynthia Shang.
2017-05-12 22:05:44 +02:00
option-archive-copy=true
option-backup-standby=false
Integration test improvements for disk and memory efficiency. Set log-level-file=off when more that one test will run. In this case is it impossible to see the logs anyway since they will be automatically cleaned up after the test. This improves performance pretty dramatically since trace-level logging is expensive. If a singe integration test is run then log-level-file is trace by default but can be changed with the --log-level-test-file option. Reduce buffer-size to 64k to save memory during testing and allow more processes to run in parallel. Update log replacement rules so that these options can change without affecting expect logs.
2019-12-17 22:23:07 +02:00
option-buffer-size=[BUFFER-SIZE]
Fixed the backup command so the backup-standby option is reset (and the backup proceeds on the master) if the standby is not configured and/or reachable. Fixed by Cynthia Shang.
2017-05-12 22:05:44 +02:00
option-checksum-page=false
Full/Synthetic test refactor. * Combine hardlink and non/compressed in synthetic tests to reduce test time and improve coverage. * Change log level of hardlink logging to detail. * Cast size in S3 manifest to integer.
2017-06-15 21:32:10 +02:00
option-compress=true
Add additional options to backup.manifest for debugging purposes. Add the buffer-size, compress-level, compress-level-network, and process-max options to the backup:option section in backup.manifest to aid in debugging. It may also make sense to propagate these options up to backup.info so they can be displayed in the info command, but for now this is deemed sufficient. Contributed by blogh.
2019-03-10 11:03:52 +02:00
option-compress-level=3
option-compress-level-network=3
Add Zstandard compression support. Zstandard is a fast lossless compression algorithm targeting real-time compression scenarios at zlib-level and better compression ratios. It's backed by a very fast entropy stage, provided by Huff0 and FSE library. Zstandard version >= 1.0 is required, which is generally only available on newer distributions.
2020-05-04 21:25:27 +02:00
option-compress-type="zst"
Add checksum delta for incremental backups. Use checksums rather than timestamps to determine if files have changed. This is useful in cases where the timestamps may not be trustworthy, e.g. when performing an incremental after failing over to a standby. If checksum delta is enabled then checksums will be used for verification of resumed backups, even if they are full. Resumes have always used checksums to verify the files in the repository, enabling delta performs checksums on the database files as well. Note that the user must manually enable this feature in cases were it would be useful or just keep in enabled all the time. A future commit will address automatically enabling the feature in cases where it seems likely to be useful. Contributed by Cynthia Shang.
2018-09-19 17:12:45 +02:00
option-delta=false
Full/Synthetic test refactor. * Combine hardlink and non/compressed in synthetic tests to reduce test time and improve coverage. * Change log level of hardlink logging to detail. * Cast size in S3 manifest to integer.
2017-06-15 21:32:10 +02:00
option-hardlink=true
Fixed the backup command so the backup-standby option is reset (and the backup proceeds on the master) if the standby is not configured and/or reachable. Fixed by Cynthia Shang.
2017-05-12 22:05:44 +02:00
option-online=false
Add additional options to backup.manifest for debugging purposes. Add the buffer-size, compress-level, compress-level-network, and process-max options to the backup:option section in backup.manifest to aid in debugging. It may also make sense to propagate these options up to backup.info so they can be displayed in the info command, but for now this is deemed sufficient. Contributed by blogh.
2019-03-10 11:03:52 +02:00
option-process-max=1
Fixed the backup command so the backup-standby option is reset (and the backup proceeds on the master) if the standby is not configured and/or reachable. Fixed by Cynthia Shang.
2017-05-12 22:05:44 +02:00
[backup:target]
pg_data={"path":"[TEST_PATH]/db-master/db/base-2/base","type":"path"}
pg_tblspc/2={"path":"../../tablespace/ts2","tablespace-id":"2","tablespace-name":"ts2","type":"link"}
[target:file]
Full/Synthetic test refactor. * Combine hardlink and non/compressed in synthetic tests to reduce test time and improve coverage. * Change log level of hardlink logging to detail. * Cast size in S3 manifest to integer.
2017-06-15 21:32:10 +02:00
pg_data/PG_VERSION={"checksum":"184473f470864e067ee3a22e64b47b0a1c356f29","master":true,"reference":"[BACKUP-FULL-3]","repo-size":[SIZE],"size":[SIZE],"timestamp":[TIMESTAMP-1]}
pg_data/badchecksum.txt={"checksum":"f927212cd08d11a42a666b2f04235398e9ceeb51","master":true,"reference":"[BACKUP-FULL-3]","repo-size":[SIZE],"size":[SIZE],"timestamp":[TIMESTAMP-1]}
pg_data/base/1/12000={"checksum":"22c98d248ff548311eda88559e4a8405ed77c003","reference":"[BACKUP-FULL-3]","repo-size":[SIZE],"size":[SIZE],"timestamp":[TIMESTAMP-1]}
pg_data/base/1/PG_VERSION={"checksum":"184473f470864e067ee3a22e64b47b0a1c356f29","mode":"0660","reference":"[BACKUP-FULL-3]","repo-size":[SIZE],"size":[SIZE],"timestamp":[TIMESTAMP-1]}
pg_data/base/16384/17000={"checksum":"7579ada0808d7f98087a0a586d0df9de009cdc33","reference":"[BACKUP-FULL-3]","repo-size":[SIZE],"size":[SIZE],"timestamp":[TIMESTAMP-1]}
pg_data/base/16384/PG_VERSION={"checksum":"184473f470864e067ee3a22e64b47b0a1c356f29","reference":"[BACKUP-FULL-3]","repo-size":[SIZE],"size":[SIZE],"timestamp":[TIMESTAMP-1]}
Fix incorrect offline upper bound for ignoring page checksum errors. For offline backups the upper bound was being set to 0x0000FFFF0000FFFF rather than UINT64_MAX. This meant that page checksum errors might be ignored for databases with a lot of past WAL in offline mode. Online mode is not affected since the upper bound is retrieved from pg_start_backup().
2019-07-11 15:13:56 +02:00
pg_data/base/32768/33000={"checksum":"7a16d165e4775f7c92e8cdf60c0af57313f0bf90","reference":"[BACKUP-FULL-3]","repo-size":[SIZE],"size":[SIZE],"timestamp":[TIMESTAMP-1]}
pg_data/base/32768/33000.32767={"checksum":"6e99b589e550e68e934fd235ccba59fe5b592a9e","reference":"[BACKUP-FULL-3]","repo-size":[SIZE],"size":[SIZE],"timestamp":[TIMESTAMP-1]}
Full/Synthetic test refactor. * Combine hardlink and non/compressed in synthetic tests to reduce test time and improve coverage. * Change log level of hardlink logging to detail. * Cast size in S3 manifest to integer.
2017-06-15 21:32:10 +02:00
pg_data/base/32768/33001={"checksum":"6bf316f11d28c28914ea9be92c00de9bea6d9a6b","reference":"[BACKUP-FULL-3]","repo-size":[SIZE],"size":[SIZE],"timestamp":[TIMESTAMP-1]}
Fix incorrect offline upper bound for ignoring page checksum errors. For offline backups the upper bound was being set to 0x0000FFFF0000FFFF rather than UINT64_MAX. This meant that page checksum errors might be ignored for databases with a lot of past WAL in offline mode. Online mode is not affected since the upper bound is retrieved from pg_start_backup().
2019-07-11 15:13:56 +02:00
pg_data/base/32768/44000_init={"checksum":"7a16d165e4775f7c92e8cdf60c0af57313f0bf90","reference":"[BACKUP-FULL-3]","repo-size":[SIZE],"size":[SIZE],"timestamp":[TIMESTAMP-1]}
Full/Synthetic test refactor. * Combine hardlink and non/compressed in synthetic tests to reduce test time and improve coverage. * Change log level of hardlink logging to detail. * Cast size in S3 manifest to integer.
2017-06-15 21:32:10 +02:00
pg_data/base/32768/PG_VERSION={"checksum":"184473f470864e067ee3a22e64b47b0a1c356f29","reference":"[BACKUP-FULL-3]","repo-size":[SIZE],"size":[SIZE],"timestamp":[TIMESTAMP-1]}
pg_data/base/base2.txt={"checksum":"cafac3c59553f2cfde41ce2e62e7662295f108c0","repo-size":[SIZE],"size":[SIZE],"timestamp":[TIMESTAMP-1]}
Add checksum delta for incremental backups. Use checksums rather than timestamps to determine if files have changed. This is useful in cases where the timestamps may not be trustworthy, e.g. when performing an incremental after failing over to a standby. If checksum delta is enabled then checksums will be used for verification of resumed backups, even if they are full. Resumes have always used checksums to verify the files in the repository, enabling delta performs checksums on the database files as well. Note that the user must manually enable this feature in cases were it would be useful or just keep in enabled all the time. A future commit will address automatically enabling the feature in cases where it seems likely to be useful. Contributed by Cynthia Shang.
2018-09-19 17:12:45 +02:00
pg_data/changecontent.txt={"checksum":"a094d94583e209556d03c3c5da33131a065f1689","master":true,"reference":"[BACKUP-FULL-3]","repo-size":[SIZE],"size":[SIZE],"timestamp":[TIMESTAMP-1]}
pg_data/changetime.txt={"checksum":"88087292ed82e26f3eb824d0bffc05ccf7a30f8d","master":true,"reference":"[BACKUP-FULL-3]","repo-size":[SIZE],"size":[SIZE],"timestamp":[TIMESTAMP-2]}
Support configurable WAL segment size. PostgreSQL 11 introduces configurable WAL segment sizes, from 1MB to 1GB. There are two areas that needed to be updated to support this: building the archive-get queue and checking that WAL has been archived after a backup. Both operations require the WAL segment size to properly build a list. Checking the archive after a backup is still implemented in Perl and has an active database connection, so just get the WAL segment size from the database. The archive-get command does not have a connection to the database, so get the WAL segment size from pg_control instead. This requires a deeper inspection of pg_control than has been done in the past, so it seemed best to copy the relevant data structures from each version of PostgreSQL and build a generic interface layer to address them. While this approach is a bit verbose, it has the advantage of being relatively simple, and can easily be updated for new versions of PostgreSQL. Since the integration tests generate pg_control files for testing, teach Perl how to generate files with the correct offsets for both 32-bit and 64-bit architectures.
2018-09-25 11:24:42 +02:00
pg_data/global/pg_control={"checksum":"4c77c900f7af0d9ab13fa9982051a42e0b637f6c","master":true,"reference":"[BACKUP-FULL-3]","repo-size":[SIZE],"size":[SIZE],"timestamp":[TIMESTAMP-2]}
Full/Synthetic test refactor. * Combine hardlink and non/compressed in synthetic tests to reduce test time and improve coverage. * Change log level of hardlink logging to detail. * Cast size in S3 manifest to integer.
2017-06-15 21:32:10 +02:00
pg_data/pg_stat/global.stat={"checksum":"e350d5ce0153f3e22d5db21cf2a4eff00f3ee877","master":true,"reference":"[BACKUP-FULL-3]","repo-size":[SIZE],"size":[SIZE],"timestamp":[TIMESTAMP-2]}
pg_data/postgresql.conf={"checksum":"6721d92c9fcdf4248acff1f9a1377127d9064807","master":true,"reference":"[BACKUP-FULL-3]","repo-size":[SIZE],"size":[SIZE],"timestamp":[TIMESTAMP-2]}
Use minio for integration testing. ScalityS3 has not received any maintenance in years and is slow to start which is bad for testing. Replace it with minio which starts quickly and ships as a single executable or a tiny container. Minio has stricter limits on allowable characters but should still provide enough coverage to show that our encoding is working correctly. This commit also includes the upgrade to openssl 1.1.1 in the Ubuntu 18.04 container.
2019-07-03 04:20:35 +02:00
pg_data/special-!_.*'()&!@;:+,?={"master":true,"reference":"[BACKUP-FULL-3]","repo-size":[SIZE],"size":[SIZE],"timestamp":[TIMESTAMP-1]}
Allow zero-size files in backup manifest to reference a prior manifest regardless of timestamp delta. Contributed by Cynthia Shang.
2018-08-24 22:50:33 +02:00
pg_data/zero_from_start={"master":true,"reference":"[BACKUP-FULL-3]","repo-size":[SIZE],"size":[SIZE],"timestamp":[TIMESTAMP-1]}
Add zero-length file to mock/all test.
2018-07-05 21:40:50 +02:00
pg_data/zerosize.txt={"master":true,"reference":"[BACKUP-FULL-3]","repo-size":[SIZE],"size":[SIZE],"timestamp":[TIMESTAMP-2]}
Full/Synthetic test refactor. * Combine hardlink and non/compressed in synthetic tests to reduce test time and improve coverage. * Change log level of hardlink logging to detail. * Cast size in S3 manifest to integer.
2017-06-15 21:32:10 +02:00
pg_tblspc/2/[TS_PATH-1]/32768/tablespace2.txt={"checksum":"dc7f76e43c46101b47acc55ae4d593a9e6983578","reference":"[BACKUP-FULL-3]","repo-size":[SIZE],"size":[SIZE],"timestamp":[TIMESTAMP-1]}
pg_tblspc/2/[TS_PATH-1]/32768/tablespace2c.txt={"checksum":"dfcb8679956b734706cf87259d50c88f83e80e66","reference":"[BACKUP-FULL-3]","repo-size":[SIZE],"size":[SIZE],"timestamp":[TIMESTAMP-1]}
Fixed the backup command so the backup-standby option is reset (and the backup proceeds on the master) if the standby is not configured and/or reachable. Fixed by Cynthia Shang.
2017-05-12 22:05:44 +02:00
[target:file:default]
group="[GROUP-1]"
master=false
mode="0600"
user="[USER-1]"
[target:link]
pg_data/pg_tblspc/2={"destination":"../../tablespace/ts2"}
[target:link:default]
group="[GROUP-1]"
user="[USER-1]"
[target:path]
The restore command is implemented entirely in C. For the most part this is a direct migration of the Perl code into C. There is one important behavioral change with regard to how file permissions are handled. The Perl code tried to set ownership as it was in the manifest even when running as an unprivileged user. This usually just led to errors and frustration. The C code works like this: If a restore is run as a non-root user (the typical scenario) then all files restored will belong to the user/group executing pgBackRest. If existing files are not owned by the executing user/group then an error will result if the ownership cannot be updated to the executing user/group. In that case the file ownership will need to be updated by a privileged user before the restore can be retried. If a restore is run as the root user then pgBackRest will attempt to recreate the ownership recorded in the manifest when the backup was made. Only user/group names are stored in the manifest so the same names must exist on the restore host for this to work. If the user/group name cannot be found locally then the user/group of the PostgreSQL data directory will be used and finally root if the data directory user/group cannot be mapped to a name. Reviewed by Cynthia Shang.
2019-09-26 13:52:02 +02:00
pg_data={"mode":"0777"}
Fixed the backup command so the backup-standby option is reset (and the backup proceeds on the master) if the standby is not configured and/or reachable. Fixed by Cynthia Shang.
2017-05-12 22:05:44 +02:00
pg_data/base={}
pg_data/base/1={}
pg_data/base/16384={}
pg_data/base/32768={}
pg_data/global={}
pg_data/pg_clog={}
pg_data/pg_dynshmem={}
Allow arbitrary directories and/or files to be excluded from a backup. Misuse of this feature can lead to inconsistent backups so read the --exclude documentation carefully before using.
2018-08-27 21:51:05 +02:00
pg_data/pg_log={}
Fixed the backup command so the backup-standby option is reset (and the backup proceeds on the master) if the standby is not configured and/or reachable. Fixed by Cynthia Shang.
2017-05-12 22:05:44 +02:00
pg_data/pg_notify={}
pg_data/pg_replslot={}
pg_data/pg_serial={}
pg_data/pg_snapshots={}
pg_data/pg_stat={}
pg_data/pg_stat_tmp={}
pg_data/pg_subtrans={}
pg_data/pg_tblspc={}
The restore command is implemented entirely in C. For the most part this is a direct migration of the Perl code into C. There is one important behavioral change with regard to how file permissions are handled. The Perl code tried to set ownership as it was in the manifest even when running as an unprivileged user. This usually just led to errors and frustration. The C code works like this: If a restore is run as a non-root user (the typical scenario) then all files restored will belong to the user/group executing pgBackRest. If existing files are not owned by the executing user/group then an error will result if the ownership cannot be updated to the executing user/group. In that case the file ownership will need to be updated by a privileged user before the restore can be retried. If a restore is run as the root user then pgBackRest will attempt to recreate the ownership recorded in the manifest when the backup was made. Only user/group names are stored in the manifest so the same names must exist on the restore host for this to work. If the user/group name cannot be found locally then the user/group of the PostgreSQL data directory will be used and finally root if the data directory user/group cannot be mapped to a name. Reviewed by Cynthia Shang.
2019-09-26 13:52:02 +02:00
pg_tblspc={"mode":"0777"}
Fixed the backup command so the backup-standby option is reset (and the backup proceeds on the master) if the standby is not configured and/or reachable. Fixed by Cynthia Shang.
2017-05-12 22:05:44 +02:00
pg_tblspc/2={}
pg_tblspc/2/[TS_PATH-1]={}
pg_tblspc/2/[TS_PATH-1]/32768={}
[target:path:default]
group="[GROUP-1]"
mode="0700"
user="[USER-1]"
Move info file checksum to the end of the file. Putting the checksum at the beginning of the file made it impossible to stream the file out when saving. The entire file had to be held in memory while it was checksummed so the checksum could be written at the beginning. Instead place the checksum at the end. This does not break the existing Perl or C code since the read is not order dependent. There are no plans to improve the Perl code to take advantage of this change, but it will make the C implementation more efficient. Reviewed by Cynthia Shang.
2019-08-22 01:45:48 +02:00
[backrest]
backrest-checksum="[CHECKSUM]"
Fixed the backup command so the backup-standby option is reset (and the backup proceeds on the master) if the standby is not configured and/or reachable. Fixed by Cynthia Shang.
2017-05-12 22:05:44 +02:00
+ supplemental file: [TEST_PATH]/db-master/repo/backup/db/backup.info
---------------------------------------------------------------------
[backrest]
backrest-format=5
backrest-version="[VERSION-1]"
[backup:current]
Set archive-check option in manifest correctly when offline. Archive check does not run when in offline backup mode but the option was set to true in the manifest. It's harmless since these options are informational only but it could cause confusion when debugging.
2019-11-28 15:27:21 +02:00
[BACKUP-FULL-3]={"backrest-format":5,"backrest-version":"[VERSION-1]","backup-archive-start":null,"backup-archive-stop":null,"backup-info-repo-size":[SIZE],"backup-info-repo-size-delta":[DELTA],"backup-info-size":[SIZE],"backup-info-size-delta":[DELTA],"backup-timestamp-start":[TIMESTAMP],"backup-timestamp-stop":[TIMESTAMP],"backup-type":"full","db-id":1,"option-archive-check":false,"option-archive-copy":true,"option-backup-standby":false,"option-checksum-page":false,"option-compress":true,"option-hardlink":true,"option-online":false}
Remove integration tests that are now covered in the unit tests. Most of these tests are just checking that errors are thrown when required. These are well covered in various unit tests. The "cannot resume" tests are also well covered in the backup unit tests. Finally, config warnings are well covered in the config unit tests. There is more to be done here, but this accounts for the low-hanging fruit.
2019-12-18 03:14:45 +02:00
[BACKUP-DIFF-4]={"backrest-format":5,"backrest-version":"[VERSION-1]","backup-archive-start":null,"backup-archive-stop":null,"backup-info-repo-size":[SIZE],"backup-info-repo-size-delta":[DELTA],"backup-info-size":[SIZE],"backup-info-size-delta":[DELTA],"backup-prior":"[BACKUP-FULL-3]","backup-reference":["[BACKUP-FULL-3]"],"backup-timestamp-start":[TIMESTAMP],"backup-timestamp-stop":[TIMESTAMP],"backup-type":"diff","db-id":1,"option-archive-check":false,"option-archive-copy":true,"option-backup-standby":false,"option-checksum-page":false,"option-compress":true,"option-hardlink":true,"option-online":false}
Set archive-check option in manifest correctly when offline. Archive check does not run when in offline backup mode but the option was set to true in the manifest. It's harmless since these options are informational only but it could cause confusion when debugging.
2019-11-28 15:27:21 +02:00
[BACKUP-DIFF-5]={"backrest-format":5,"backrest-version":"[VERSION-1]","backup-archive-start":null,"backup-archive-stop":null,"backup-info-repo-size":[SIZE],"backup-info-repo-size-delta":[DELTA],"backup-info-size":[SIZE],"backup-info-size-delta":[DELTA],"backup-prior":"[BACKUP-FULL-3]","backup-reference":["[BACKUP-FULL-3]"],"backup-timestamp-start":[TIMESTAMP],"backup-timestamp-stop":[TIMESTAMP],"backup-type":"diff","db-id":1,"option-archive-check":false,"option-archive-copy":true,"option-backup-standby":false,"option-checksum-page":false,"option-compress":true,"option-hardlink":true,"option-online":false}
Fixed the backup command so the backup-standby option is reset (and the backup proceeds on the master) if the standby is not configured and/or reachable. Fixed by Cynthia Shang.
2017-05-12 22:05:44 +02:00
[db]
db-catalog-version=201409291
db-control-version=942
db-id=1
Automate generation of WAL and pg_control test files. The existing static files would not work with 32-bit or big-endian systems so create functions to generate these files dynamically rather than creating a bunch of new static files.
2017-11-19 03:02:54 +02:00
db-system-id=1000000000000000094
Fixed the backup command so the backup-standby option is reset (and the backup proceeds on the master) if the standby is not configured and/or reachable. Fixed by Cynthia Shang.
2017-05-12 22:05:44 +02:00
db-version="9.4"
[db:history]
Automate generation of WAL and pg_control test files. The existing static files would not work with 32-bit or big-endian systems so create functions to generate these files dynamically rather than creating a bunch of new static files.
2017-11-19 03:02:54 +02:00
1={"db-catalog-version":201409291,"db-control-version":942,"db-system-id":1000000000000000094,"db-version":"9.4"}
Improve performance of info file load/save. Info files required three copies in memory to be loaded (the original string, an ini representation, and the final info object). Not only was this memory inefficient but the Ini object does sequential scans when searching for keys making large files very slow to load. This has not been an issue since archive.info and backup.info are very small, but it becomes a big deal when loading manifests with hundreds of thousands of files. Instead of holding copies of the data in memory, use a callback to deliver the ini data directly to the object when loading. Use a similar method for save to avoid having an intermediate copy. Save is a bit complex because sections/keys must be written in alpha order or older versions of pgBackRest will not calculate the correct checksum. Also move the load retry logic to helper functions rather than embedding it in the Info object. This allows for more flexibility in loading and ensures that stack traces will be available when developing unit tests. Reviewed by Cynthia Shang.
2019-09-06 19:48:28 +02:00
[backrest]
backrest-checksum="[CHECKSUM]"
Reference in New Issue Copy Permalink