This issue was a result of STORAGE_REPO_PATH prepending an extra stanza when the stanza was specified on the command line.
The tests missed this because by some strange coincidence the WAL dirs were empty for each test that specified a stanza. Add new tests to prevent a regression.
Fixed by Stefan Fercot.
Free all cached objects in the storage helper, especially the stanza name.
This clears the storage environment for tests that switch stanza names or go from a stanza name to no stanza name or vice versa. This is only useful for testing right now, but may be used in the future for commands than act on multiple stanzas.
This was previously 256, which was too small to log protocol parameters. Not only did this truncate important debug information but varying path lengths caused spurious differences in the expect logs.
This command was previously forked off from the archive-get command which required a bit of artificial option and log manipulation.
A separate command is easier to test and will work on platforms that don't have fork(), e.g. Windows.
These are intended to be temporary until a fully automated report is developed.
Since we don't know when that will happen, at least make it easier to generate the current report.
Prior to this the Perl remote was used to satisfy C requests. This worked fine but since the remote needed to be migrated to C anyway there was no reason to wait.
Add the ProtocolServer object and tweak ProtocolClient to work with it. It was also necessary to add a mechanism to get option values from the remote so that encryption settings could be read and used in the storage object.
Update the remote storage objects to comply with the protocol changes and add the storage protocol handler.
Ideally this commit would have been broken up into smaller chunks but there are cross-dependencies in the protocol layer and it didn't seem worth the extra effort.
The file write object destructors called close() and finalized the file even if it was not completely written. This was an issue in both the C and Perl code.
Rewrite the destructors to simply free resources (like file handles) rather than calling the close() method. This leaves the temp file in place for filesystems that use temp files.
Add unit tests to prevent regression.
Reported by blogh.
execRead() should be returning a size_t, not a void. Thankfully, this isn't actually used and therefore shouldn't be an issue, but we should fix it anyway.
Contributed by Stephen Frost.
This was not being caught because the integration tests for S3 were running remotely and going through the Perl code rather than the new C code.
Implement the exists method for the S3 driver and add tests to prevent a regression.
Reported by mibiio.
The check to verify that pg-path and data_directory are equal was not working because pg-path was getting overwritten with data_directory before validation took place.
Reported by James Chanco Jr.
This already worked in reverse, but this case is needed when a command that only uses protocol-timeout (e.g. info) calls a remote process where protocol-timeout and db-timeout can be set. If protocol-timeout was set to less than the default db-timeout then an error resulted.
Bug Fixes:
* Fix issue with multiple async status files causing a hard error. (Reported by Vidhya Gurumoorthi, Joe Ayers, Douglas J Hunley.)
Improvements:
* The info command is implemented entirely in C.
* Simplify info command text message when no stanzas are present by replacing the repository path with "the repository".
* Add _DARWIN_C_SOURCE flag to Makefile for MacOS builds. (Contributed by Douglas J Hunley.)
* Update address lookup in C TLS client to use modern methods. (Suggested by Bruno Friedmann.)
* Include Posix-compliant header for strcasecmp() and fd_set. (Suggested by ucando.)
This prevented packages from being passed to the documentation unless they were in the /backrest directory on the host.
Also make the local path /pgbackrest instead of the deprecated /backrest.
Reported by Heath Lord.
Rather than create _P/_PP variants for every type that needs to pass/return pointers, create FUNCTION_*_P/PP() macros that will properly pass or return any single/double pointer types.
There remain a few unresolved edge cases such as CHARPY but this handles the majority of types well.
This parameter was always useless but commit 7333b630 removed all references to it so remove the parameter at all call sites as well.
The original intention was probably to allow logging of TEST return values but that never happened.
Rather than create a CONST_ variant for every type that needs to be returned const, create a FUNCTION_LOG_RETURN_CONST() macro that will return any type as const.
The string object was reallocating memory with every concatenation which is not very efficient. This is especially true for JSON rendering which does a lot of concatenations.
Instead allocate a pool of extra memory on the first concatenation (50% of size) to be used for future concatenations and reallocate when needed.
Also add a 1GB size limit to ensure that there are no overflows.
Multiple status files were being created by asynchronous archiving if a high-level error occurred after one or more WAL segments had already been transferred successfully. Error files were being written for every file in the queue regardless of whether it had already succeeded. To fix this, add an option to skip writing error files when an ok file already exists.
There are other situations where both files might exist (various fsync and filesystem error scenarios) so it seems best to retry in the case that multiple status files are found rather than throwing a hard error (which then means that archiving is completely stuck). In the case of multiple status files, a warning will be logged to alert the user that something unusual is happening and the command will be retried.
Reported by fpa-postgres, Joe Ayers, Douglas J Hunley.
gcc has apparently merged this function in string.h but Posix specifies that it should be in strings.h. FreeBSD at at least is sticking to the standard.
In the long run it might be better to implement our own strcasecmp() function but for now just add the header.
Suggested by ucando.
The implementation using gethostbyname() was only intended to be used during prototyping but was forgotten when the code was finalized.
Replace it with gettaddrinfo() which is more modern and supports IPv6.
Suggested by Bruno Friedmann.
Rename FUNCTION_DEBUG_* macros to FUNCTION_LOG_* to more accurately reflect what they do. Further rename FUNCTION_DEBUG_RESULT* macros to FUNCTION_LOG_RETURN* to make it clearer that they return from the function as well as logging. Leave FUNCTION_TEST_* macros as they are.
Consolidate the various ASSERT* macros into a single ASSERT macro that is always compiled out of production builds. It was difficult to figure out when an assert would be checked with all the different types in play. When ASSERTs are compiled in they will always be checked regardless of the log level -- tying these two concepts together was not a good idea.
The C info code has already been committed but this commit wires it into main.
Also remove the info Perl code and tests since they are no longer called.
This is a partial implementation of remote storage with just enough functionality to get the info command working. The client is written in C but the server is still in Perl, which limits progress until a C server is written.
This is a complete protocol client implementation in C.
Currently there is no C server implementation so the C client is talking to a Perl server. This won't work very long, though, as the protocol format, even though in JSON, has a lot of language-specific structure. While it would be possible to maintain compatibility between C and Perl it's probably not worth the effort in the long run.
Just as in Perl there are helper functions to make constructing protocol objects easier. Currently only repository remotes are supported.
Executes a child process and allows the calling process to communicate with it using read/write io.
This object is specially tailored to implement the protocol layer and may or may not be generally applicable to general purpose
execution.
Parameters for the local/remote commands are based on parameters that are passed to the current command.
Generate parameters for the new command based on the intersection of parameters between the current command and the command to be executed.
General i/o objects for reading and writing file descriptors, in particular those that can block. In other words, these are not generally to be used with file descriptors for actual files, but rather pipes, sockets, etc.
Replace the repository path with just "the repository". The path is not important in this context and it is clearer to state where the stanzas are missing from.
