EDB PgPool-II 4.2.5 release notes v4
Released: 10 Jun 2021
EDB Pgpool-II 4.2.5 includes the following upstream merges and bug fixes:
| Type | Description |
|---|---|
| Upstream merge | Merged with community Pgpool-II 4.2.5. See the community Release Notes for details. |
| Upstream merge | Merged with community Pgpool-II 4.2.4. See the community Release Notes for details. |
| Bug fix | Fixed maximum length of hostnames, including domain name. Now the correct value is 256 instead of 128. |
| Bug fix | Fixed query cache so that it does not cache SQLValueFunctions. |
| Bug fix | Implemented follow_primary command-locking over the watchdog channel. This is a supplementary fix for a race condition between detach_false_primary and follow_primary_command. |
| Bug fix | Fixed the issue wherein the client-side hangs when a NoData response follows a Describe statement. |
| Bug fix | Fixed an issue that sent an incorrect length error query (to abort transactions running on the non-main nodes), throwing an “invalid string in message” error at the backend in SI mode. |
| Bug fix | Fixed an issue wherein child processes created by follow_primary command were orphaned when pgpool was shut down. |
| Bug fix | Fixed an issue wherein, after executing the follow_primary command, the status of a standby node was set as down using the pcp_detach_node. |
| Bug fix | Fixed the watchdog communication race condition wherein the watchdog notified the new message to the main process, which took significant time to process the message. |
| Bug fix | Fixed an issue wherein the watchdog node status was not updating after rebooting one of the cluster nodes. |
| Bug fix | Fixed an issue wherein the follow_primary_command failed when watchdog was enabled. |
| Bug fix | Fixed a bug related to PGPOOL SHOW heartbeat and PGPOOL SHOW ALL commands where the last entry of command result was not shown. |
| Bug fix | Fixed an issue to allow the value of the parameter log_rotation_age to be set to 0 to disable it. |
| Bug fix | Fixed an incorrect display of the load balancing node status in raw clustering mode. |
| Bug fix | Fixed the pgpool logger process, which was utilizing 100% CPU. |
| Bug fix | Fixed the SI mode to acquire a snapshot with an internal transaction. |
| Bug fix | Fixed the backend_flag* parameter, which was shown twice while executing the PGPOOL SHOW ALL command. |
| Bug fix | Updated the other_pgpool parameter (removed in the 4.2 release) to watchdog in the SHOW ALL command. |