A.36. Release 4.2.4

Fix maximum length of hostnames including domain name. (Tatsuo Ishii)

The maximum length of hostnames was 128, but the correct value is 254. Discussion: [pgpool-hackers: 3903]

Fix query cache to not cache SQLValueFunctions. (Tatsuo Ishii)

The list of SQLValueFunctions is follow.

• CURRENT_DATE

• CURRENT_TIME

• CURRENT_TIMESTAMP

• LOCALTIME

• LOCALTIMESTAMP

• CURRENT_ROLE

• CURRENT_USER

• SESSION_USER

• USER

• CURRENT_CATALOG

• CURRENT_SCHEMA

Implementing the follow_primary command-locking over the watchdog channel. (Muhammad Usama)

This is supplementary fix for "Fix race condition between detach_false_primary and follow_primary_command". The design is good enough for most cases, but has the potential to fail in some case.

Fix client side hang when describe message is followed by NoData response. (Tatsuo Ishii)

Problem reported and patch provided by Daniel van de Giessen. Discussion: [pgpool-hackers: 3950]

Fix typo in pgpool.conf samples. (Tatsuo Ishii)

Fix sending invalid message in SI mode. (Tatsuo Ishii)

When a query is aborted by specific reason like serialization error, Pgpool-II sends error query to abort transactions running on non main nodes. The message length of the query was incorrect and it caused "invalid string in message" error on backend.

Fix orphan process is left when pgpool is going down. (Tatsuo Ishii)

When pgpool is going down while follow primary command is ongoing, some process started by follow primary child process could be left.

Fix pcp_detach_node leaves down node. (Tatsuo Ishii)

As detaching primary node using pcp_detach_node, a standby node is kept down status after follow primary command was executed. Discussion: [pgpool-hackers: 3915]

Fix backend_application_name cannot be changed by reloading pgpool.pgpool.conf. (Tatsuo Ishii)

Fix watchdog communication race condition. (Tatsuo Ishii)

When watchdog notifies the new message to main process while main process is executing the process notified by watchdog process, there was a significant delay before the new message was processing. Discussion: [pgpool-hackers: 3900]

Fix watchdog node status not updating after rebooting. (bug 684) (Muhammad Usama)

Doc: Update configuration example "Pgpool-II on Kubernetes". (Bo Peng)

Doc: add more explanation about backend_application_name. (Tatsuo Ishii)

Doc: fix documentation typos. (Bo Peng)

Doc: fix typo in in memory query cache document. (Tatsuo Ishii)

Doc: fix parameter name typo in ldap option (Takuma Hoshiai)

Doc: fix wd_life_point description (Masaya Kawamoto)

Fix pgpool_setup to generate portable follow_primary.sh. (Tatsuo Ishii)

Fix rsync parameter in pgpool_setup. (Tatsuo Ishii)

Fix follow primary script creation in pgpool_setup. (Tatsuo Ishii)

Fix pgpool_setup in creating base backup script. (Tatsuo Ishii)

Enhance watchdog_setup script. (Tatsuo Ishii)

shutdownall script generated by watchdog_setup caused failover event which is not necessary in the whole shutdown sequence.

Fix 075.detach_primary_left_down_node. (Tatsuo Ishii)

Fix 031.connection_life_time. (Tatsuo Ishii)

Fix 018.detach_primary error in the log. (Tatsuo Ishii)

Fix occasional failure of 028.watchdog_enable_consensus_with_half_votes test. (Tatsuo Ishii)

Fix occasional regression test 018.detach_primary error. (Tatsuo Ishii)