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.
Master and standby can both be configured on the backup server and pgBackRest will automatically determine which is the master. This means no configuration changes for backup are required after failing over from a master to standby when a separate backup server is used.
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.
* Tablespace paths that had $PGDATA as a substring would be identified as a subdirectories of $PGDATA even when they were not.
* Also hardened relative path checking a bit.
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.
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.
* Make the code more modular and object-oriented.
* Multiple Docker containers can now be created for a single test to simulate more realistic environments.
The pg_xlogfile_name() function is no longer used to construct WAL filenames from LSNs. While this function is convenient it is not available on a standby. Instead, the archive is searched for the LSN in order to find the timeline. If due to some misadventure the LSN appears on multiple timelines then an error will be thrown, whereas before this condition would have passed unnoticed.
* Fixed an issue where keep-alives could be starved out by lots of small files during multi-threaded operation and were completely absent during single-threaded operation when resuming from a previous incomplete backup.
Reported by Janice Parkinson.
* Added the protocol-timeout option. Previously protocol-timeout was set as db-timeout + 30 seconds.
* Failure to shutdown remotes at the end of the backup no longer throws an exception. A warning is still generated that recommends a higher protocol-timeout.
* Fixed an issue where the expire command would refuse to run when explicitly called from the command line if the db-host option was set. This was not an issue when expire was run after a backup, which is the usual case.
* Option handling is now far more strict. Previously it was possible for a command to use an option that was not explicitly assigned to it. This was especially true for the backup-host and db-host options which are used to determine locality.
Reported by Chris Barber.
* Containers now use a squid proxy for apt/yum to speed builds.
* Obsolete containers are removed by the <br-option>--vm-force</br-option> option.
* Greatly reduced the quantity of Docker containers built by default. Containers are only built for PostgreSQL versions specified in db-minimal and those required to build documentation. Additional containers can be built with --db-version=all or by specifying a version, e.g. --db-version=9.4.
Added an execution cache so that documentation can be generated without setting up the full container environment. This is useful for packaging, keeps the documentation consistent for a release, and speeds up generation when no changes are made in the execution list.
Added database version constants and changed version identification code to use hash tables instead of if-else. Propagated the db version constants to the rest of the code and in passing fixed some path/filename constants.
Added new regression tests to check that specific files are never copied.
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.
This change allows for easier testing since all files are local on the host VM and can be easily accessed without using `docker exec`. In addition, this change is required to allow multiple Docker containers per test case which is coming soon.
* 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.
* 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.
* Specific VMs can now be built by using --vm along with --vm-build.
* Docker caching can be disabled with --vm-force.
* ControlMaster is now used for al VMs to improve test speed.
Fixed an issue where the master process was passing --repo-remote-path instead of --repo-path to the remote and causing the lock files to be created in the default repository directory (/var/lib/backup), generally ending in failure. This was only an issue when --repo-remote-path was defined on the command line rather than in pg_backrest.conf.
Perl Critic added and passes on gentle. A policy file has been created with some permanent exceptions and a list of policies to be fixed in approximately the order they should be fixed in.
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.
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.
Fixed an issue where document generation failed because some OSs are not tolerant of having multiple installed versions of PostgreSQL. A separate VM is now created for each version. Also added a sleep after database starts during document generation to ensure the database is running before the next command runs.
Reported by John Harvey.
1) Tests for all operating systems can now be run with a single command.
2) Tests can be run in parallel with --process-max.
3) Container generation now integrated into test.pl
4) Some basic test documentation.
Keepalives are now used to make sure the remote for the main process does not timeout while the thread remotes do all the work. The error messages for timeouts was also improved to make debugging easier.
* Better messaging for expiration.
* Fixed already stopped message.
* retention-archive and retention-archive-type now use retention-full and 'full' when not specified.
* Fixed issue where backup-user was required (should default to backrest).
* ExecuteTest now supports retries.
* Fixed issue where log test was not comparing test logs.
* Fixed issue where test logs would not match for ssh connection errors
Unable to reproduce this anymore. It seems to have been fixed with the last round of config changes. Add regression tests to make sure it doesn't happen again.
Most tests are working now. What's not working:
1) --target-resume option fails because pause_on_recovery setting was removed. Need to implement to the new 9.5 option and make that work with older versions in a consistent way.
2) No tests for the new .partial WAL segments that can be generated on timeline switch.
* Major refactoring of the protocol layer to support this work.
* Fixed protocol issue that was preventing ssh errors (especially connect) from being logged.
Replaced IPC::System::Simple and Net::OpenSSH with IPC::Open3 to eliminate CPAN dependency for multiple distros. Using open3 will also be used for local processes so it make sense to switch now.