The Pgpool-II project has released maintenance updates for five branches, including version 4.7.2, to address a watchdog synchronization race that can lock standby nodes during cluster restarts, as well as routing errors and handling of aborted transactions. These updates are critical for system administrators managing PostgreSQL connection pools to prevent silent failover failures and memory leaks under heavy load. Key fixes include improvements in how the middleware processes aborted transactions, debug logging, query cache invalidation for the MERGE command, and corrections to routing errors that could lead to outdated data being sent to the wrong database nodes. Administrators are advised to apply these updates promptly and monitor their systems during the upgrade process to ensure proper functionality
Pgpool-II 4.7.2 and Older Branches Released to Fix Watchdog Races and Routing Bugs
The Pgpool-II project just pushed maintenance releases across five branches to patch a watchdog synchronization race that locks standby nodes during cluster restarts. Engineers also corrected how the middleware handles aborted transactions, debug logging, and query cache invalidation for PostgreSQL MERGE commands. Routing errors for read-only settings and temporary table cache leaks get fixed to stop stale data from hitting the wrong database nodes. System administrators managing PostgreSQL connection pools should apply these updates immediately to prevent silent failover failures and memory leaks under heavy load.
