E.13. Postgres Pro Enterprise 14.8.1

Release Date: 2023-06-06

E.13.1. Overview

This release is based on PostgreSQL 14.8 and Postgres Pro Enterprise 14.7.1. All changes inherited from PostgreSQL 14.8 are listed in PostgreSQL 14.8 Release Notes. As compared with Postgres Pro Enterprise 14.7.1, this version also provides the following changes:

  • Implemented lock deduplication that allows to effectively store in memory and track all exclusive locks held by a standby server's startup process during WAL replay.

  • Fixed pruning logic to slightly improve performance.

  • Optimized CFS garbage collector: reduced CPU costs and reads from disk during file analysis.

  • Added the unicode_nul_character_replacement_in_jsonb configuration parameter to replace '\u0000' with the specified unicode character when calling a function processing JSONB.

  • Improved performance of DDL operations on temporary tables. When max_connections is high, the costs to acquire locks for the tables that are marked ON COMMIT DELETE ROWS are now reduced due to the change of the locking mechanism.

  • Added an operation log, which stores information about system events such as an upgrade, execution of pg_resetwal and so on, which is highly useful for vendor's technical support. Recording to the operation log is only done at the system level, and SQL functions are used to read it. (See Section 9.27.12.)

  • Fixed an issue in the connection pooler, which could cause backend termination if the client aborted the connection during authentication.

  • Fixed an issue with much larger CFS data files on the replica compared to the primary. Now the replica writes the files in the same way as the primary when consistent recovery state is reached.

  • Fixed an internal error that occurred during a self join removal. Previously, it was not taken into account that when an expression A.x=B.x was transformed to A.x IS NOT NULL, it was necessary to correct meta parameters of the transformed expression because of its different semantics. In particular, the initial expression can be used in cases such as selection of an attribute to sort by or being a JOIN parametrization expression, but this is inapplicable for the transformed expression.

  • Removed the deprecated configuration parameter cfs_gc_verify_file from postgresql.conf. The configuration parameter cfs_encryption is deprecated and will be removed in a future release.

  • Added support for ROSA Chrome 2021.2.

  • Ended support for Ubuntu 18.04.

  • Upgraded mamonsu to version 3.5.3, which provides new features and bugfixes. Notable changes are as follows:

    • Removed the limitation that did not allow a user different from mamonsu to read and edit the configuration file agent.conf. Specifically, the pg_probackup plugin now works correctly when the mamonsu user owns the pg_probackup backup directory.

    • Removed metrics that conflict native Zabbix agents from Zabbix configuration file being exported.

    • Fixed mamonsu choice of the default extension to collect statistics. Now when pgpro_stats and pg_stat_statements are both installed, pgpro_stats is correctly treated as the default one.

    • Added a check that the server where mamonsu WAL plugin runs is not in recovery.

  • Excluded CREATE TABLESPACE and DROP TABLESPACE from multimaster replication. These commands should be executed separately on each peer node.

  • Upgraded pg_hint_plan to version 1.4.1, which provides the hints_anywhere parameter that allows to place hints anywhere in the query.

  • Upgraded pg_probackup to version 2.6.4 Enterprise, which provides optimizations and bug fixes. Lower versions are incompatible with Postgres Pro Enterprise 14.8.1. Notable changes are as follows:

    • Included the catchup command, which was previously provided only in a test mode, in the production delivery.

    • Added optimizations for incremental backups that reduce disk read amount almost twice.

    • Improved error and warning messaging and unified the error message style.

    • Fixed issues and improved the reliability of CFS table merge and backup operations.

    • Fixed usability issues with the show command, in particular: fixed displaying the recovery time, which could sometimes not be displayed after a successful backup; made the backup time be displayed in a table view even if the backup ended with an error.

    • Fixed an error in specifying the relative path after restoring a cluster from a backup.

    • Corrected an improper assignment of backup IDs that took place in some cases.

    • Added an option to skip creation or initialization of an existing backup instance.

    • Fixed an issue with inclusion of the whole data in an incremental backup rather than only the changed data.

  • Upgraded pgpro_stats to version 1.6, which provides new features and bugfixes:

    • Implemented tracing of application sessions, based on filters, which trigger logging the execution of queries that match filtering conditions. You can use pgpro_stats functions to create, update or delete query filters. (See Section G.2.6.1 for details.)

    • Added functions for superuser to create views that emulate pg_stat_statements and pg_stat_kcache extensions. Specifically, pg_stat_statements, pg_stat_statements_info, pg_stat_kcache and pg_stat_kcache_detail views can be created. Once they are created, you can work with them as if the appropriate extension is installed. (See Section G.2.6.2 for details.)

    • Added pgpro_stats_archiver view, which extends the pg_stat_archiver view by adding two fields, showing overall time that the archiver process was active and overall execution time of the archive command.

    • Implemented backward compatibility of the vacuum statistics interface. Now a newer version of the pgpro_stats shared library can be safely used with old declarations of SQL functions. Previously such cases caused a server crash.

    • Fixed possible server crashes that could occur due to PGDATA overflow when in pgpro_stats, one or several long queries were executed. Now instead of storing query texts in a temporary file located in the PGDATA/pg_stat_tmp directory, it is possible to choose the directory with the external file to store query texts by setting the value of the pgpro_stats.stats_temp_directory configuration parameter. If you wish, you can set this value equal to the value of the system configuration parameter stats_temp_directory. But keep in mind that pgpro_stats.stats_temp_directory can only be set at the server start.

  • Upgraded pgpro_pwr to version 4.1.3, which supports pgpro_stats 1.6 and provides the following new features:

    • Added report tables Top tables by removed all-visible/all-frozed marks.

    • Added a few new fields to the Cluster Statistics report table.

  • Upgraded PTRACK to version 2.4.3.

  • Upgraded orafce to version 4.2.6.

E.13.2. Migration to Version 14.8.1

If you are upgrading from a Postgres Pro Enterprise release based on the same PostgreSQL major version, it is enough to install the new version into your current installation directory.

If you are migrating to the current major version of Postgres Pro Enterprise using pg_upgrade, and you want to keep the existing visibility map representation in pageinspect, you should run the VACUUM command before the upgrade.

When upgrading to Postgres Pro Enterprise versions starting with 14.8.1, make sure to upgrade pg_probackup Enterprise to version 2.6.4 or higher since lower versions are incompatible with these Postgres Pro Enterprise versions.

To migrate from PostgreSQL, as well as Postgres Pro Standard or Postgres Pro Enterprise based on a previous PostgreSQL major version, see the migration instructions for version 14.