A.16. Release 4.3.1

Suppress message length log for in_hot_standby. (Tatsuo Ishii)

PostgreSQL 14 introduced new config parameter in_hot_standby.

The value is either "on" for standby servers or "off" for primary servers. As a result pgpool log is flooded by the messages:

     LOG:  reading message length
     DETAIL:  message length (22) in slot 1 does not match with slot 0(23)
    

To avoid this, only complain if the parameter name is not in_hot_standby.

Also the message is enhanced to show the parameter name.

     LOG:  ParameterStatus "TimeZone": node 1 message length 30 is different from main node message length 24
    

Discussion: [pgpool-hackers: 4076]

Change the default value for log_line_prefix. (Tatsuo Ishii)

The default value for log_line_prefix is changed from "%t pid: %p" to "%m %a pid: %p". Discussion: [pgpool-hackers: 4097]

Add validations of wd_lifecheck_password and recovery_password format (Masaya Kawamoto)

wd_lifecheck_password and recovery_password are not allowed to be md5 hashed password format but pgpool just reported authentication failure and did not check them.

Enhance parameter status handling. (Tatsuo Ishii)

When a parameter status message arrives from backend Pgpool-II forwards it to frontend not just memorize.

Fix pgpool_remote_start.sample script. (Bo Peng)

Rename the parameter "wd_initial_node_showup_time" to wd_no_show_node_removal_timeout. (Bo Peng)

Differently from documentation wd_no_show_node_removal_timeout was written as wd_initial_node_showup_time in source code.

Fix for a small mistake in pgpool-recovery SQL script (Muhammad Usama)

Fix pcp_node_info hang when pgpool cannot connect to backend. (Tatsuo Ishii)

Problem reported and patch reviewed by Emond Papegaaij. Discussion: [pgpool-general: 7983]

Fix long standing bug with pcp_node_info. (Tatsuo Ishii)

Discussion: [pgpool-hackers: 4109]

Fix health check process issues pointed out by Coverity. (Tatsuo Ishii)

Fix memory leak pointed out by Coverity. (Tatsuo Ishii)

Fix failover() to deal with the case when no former primary node exists. (Tatsuo Ishii)

In case Req_info->primary_node_id is -1 like no primary node exists when Pgpool-II starts up, failover() skipped to call find_primary_node_repeatedly(). Also follow_master_command was not executed.

Discussion: [pgpool-hackers: 4113]

Fixed sample failover script. (Bo Peng)

This script did not consider the case when the old primary node id is "-1".

Fixed the streaming replication check process not to retry if it cannot connect to the backend.(bug 742) (Bo Peng)

This retry caused a long time failover.

Sync with English manual. (Tatsuo Ishii)

Fix typo in pcp_watchdog_info manual. (Tatsuo Ishii)

Fix documentation typos. (pengbo)

Patch is created by Umar Hayat.

Add "exclude" settings to /etc/yum.repos.d/pgdg-redhat-all.repo so that Pgpool-II is not installed from PostgreSQL YUM repository. (Bo Peng)

Add restriction regarding ParameterStatus and in_hot_standby parameter. (Tatsuo Ishii)

Add restriction about set_config. (Tatsuo Ishii)

Fix pgpool_setup in failover scrip creation. (Tatsuo Ishii)

When pgpool_setup creates failover.sh, it did not consider the case when no primary server existed.

Allow to run regression test against existing installation without recompiling. (Tatsuo Ishii)

Discussion: [pgpool-hackers: 4077]

Fix regression test 075. (Tatsuo Ishii)

The test reported success even if pgpool does not start up. Problem reported and patch provided by Qiang Lingjie. Discussion: [pgpool-hackers: 4085]

Fix pgpool_setup and watchdog_setup binary PATH in noinstall mode. (Bo Peng)