Movatterモバイル変換

• Improved performance by introducing a more efficient synchronization primitive instead of previously used spinlocks for a shared invalidation message queue, which allows avoiding a bottleneck. This may be useful for1C.

• UpdatedCREATE DATABASE,createdb, andinitdb to display an information message when the value oflc_collate is used as the default foricu_locale or--icu-locale.

• Optimized the behavior when working with theCOALESCE function. Now this function is replaced with one of its arguments if all other arguments were evaluated tonull. This allows improving the selectivity of query clauses and producing better execution plans.

• Introduced the following changes to the implementation ofcrash_info:

  • Added thecrash_info_timer,crash_info_timer_interval, andcrash_info_query_threshold configuration parameters that allow profiling long queries incrash_info output files.

  • Extended the diagnostic information that thecrash_info_dump configuration parameter can specify to write in case of a backend crash. By default, now the general information on the signal is dumped and for Linux x64, also the information on the processor state and bytes near RIP (return instruction pointer). All this ensures better diagnostics, in particular, in case ofSIGILL (illegal instruction) signals. The way of dumping stack files in case of crashes has also been changed for Linux x64. Besides,memory_context has been removed from the default list of sources to be dumped in order to avoid failures that could occur when writing the diagnostic information.

• Fixed an issue inLISTEN/NOTIFY queue handling, which caused failure to obtain transaction status. Previously,VACUUM FREEZE could advancedatfrozenxid too far forward.

• Fixed an issue where the planner could choose non-optimal execution plans for queries that use theLIKE operator with themchar data type.

• Added support for Debian 13, Red Hat Enterprise Linux 10, Rocky Linux 10, Oracle Linux 10, and AlmaLinux 10.

• Ended support for Ubuntu 20.04.

• Added thewal2json extension for logical decoding that converts database changes from the write-ahead log (WAL) intoJSON format.

• Upgradedaqo to provide the following bugfixes:

  • Added thetmpoids field to theaqo_data view and thetmpoids parameter to theaqo_data_update function to support queries that involve temporary tables along with persistent tables.

  • Fixed the race condition when calling theaqo_cleanup function from different sessions.

  • Fixed an issue whereaqo learning data was not saved for queries that use materialization.

  • Fixed an issue inaqo, which could lead to a standby crash withSIGABRT if the standby had a differentaqo version than the primary.

• Upgradedauto_explain to move the planning time of a query in the output to a separate line under the planner tree, as in theEXPLAIN output. With the former location of the planning time in the output,pgbadger could not recognize the information on a query and therefore did not include it in the report.

• Updatedpg_pathman to fix an issue with partition access handling. The issue could cause commands that reuse the same query plan, for example, inPL/pgSQL, to fail with errors such as“relation with OIDNNNN does not exist”.

• Upgradedpg_proaudit to provide the following enhancements:

  • Added support forCEF format for event log files.

  • Adjustedpg_proaudit output.xid andvxid fields are no longer displayed if they are zero.

• Upgradedpg_probackup to version 2.8.10, which provides the following new features, optimizations, and bug fixes:

  • Added the--wal-tree option for theadd-instance command to store archived WAL files in a tree-like structure. This improves WAL operation performance with large volumes of logs.

  • Added thepostgres-edition field to theshow command output in theJSON format to display DBMS edition information.

  • Improved the--dry-run option operation during tablespace recovery. The operation estimation and plan output are now more precise.

  • Fixed an issue related to the retention window calculation. The time value is now taken from the host wherepg_probackup is running. The retention window is now calculated correctly even with the timezone differences between the servers.

  • Fixed a problem with thearchive-get command operation. Thepbk_prefetch directory is now created with correct access rights, inheriting permissions fromPGDATA.

  • Fixed false failure of thecheckdb command in the remote mode during concurrentTRUNCATE operations or table drops on standby nodes.

• Fixed the following issues in thepgpro_bindump module:

  • Fixed potentialpg_probackup3 failure caused by temporary tables during backup validation using thepgpro_bindump module in the PRO mode.

  • Fixed a memory leak inpg_probackup3 using thepgpro_bindump module for PTRACK operations.

  • Fixed an issue that could cause infinite looping inpg_probackup3 when using thepgpro_bindump module.

  • Fixed an issue with multi-timeline backup in the PTRACK mode inpg_probackup3 using thepgpro_bindump module in the PRO mode.

• Upgradedpgpro_pwr to version 4.10, which completely redesigns the report and supportsPostgreSQL 18.

• Upgradedpgpro_stats to provide the following improvements and bug fixes:

  • Implemented the ability to track utility commands by theirquery_id, as it is done inpg_stat_statements. Previously, utility commands were tracked based on strict comparison of their textual query strings, and therefore could not be properly normalized.

  • Fixed a segmentation fault that could occur when defining custom metrics. The error originated from a memory allocation issue, which was fixed.

• Fixed an issue with incorrect configuration parameter values specified when using the1c.tune preset of thepgpro_tune utility.

• Upgradedpg_query_state to version 1.2 to add thepg_progress_bar andpg_progress_bar_visual functions that allow monitoring the current progress of query execution.

• Upgradedrum to provide the following bugfixes:

  • Fixed a segmentation fault caused by invalid memory access when a buffer pointer was used after the underlying buffer was modified by another thread.

  • Fixed an index scan infinite loop when using multipleORDER BY clauses.

Important

If you used a previous version ofaqo and need to upgrade it, after the upgrade of the primary server, the standby servers may have a differentaqo version.

To avoid any related issues, disableaqo by settingaqo.enable andaqo.wal_rw tooff first on the primary server and then on the standby servers. After the upgrade completes successfully, re-enableaqo by settingaqo.wal_rw andaqo.enable toon first on the standby servers and then on the primary server.