Loading...
Loading...
A PostgreSQL outage occurred when a cluster hit transaction ID wraparound after teams disabled autovacuum following an earlier incident where autovacuum had been mistakenly blamed. With autovacuum left off, transaction IDs aged without routine cleanup, forcing a disruptive recovery and service interruption. The incident underscores human-error risks in operational decisions, the necessity of background maintenance like autovacuum to prevent wraparound, and the danger of quick fixes that ignore long-term maintenance. Engineers running PostgreSQL at scale should treat autovacuum as critical, review operational changes, and monitor maintenance features to safeguard availability and data integrity.
The PostgreSQL Global Development Group released minor updates for all supported branches — 18.4, 17.10, 16.14, 15.18, and 14.23 — addressing 11 security vulnerabilities and more than 60 bug fixes. Notable CVEs include CVE-2026-6472 (privilege check bypass in CREATE TYPE), CVE-2026-6473 (integer wraparound causing out-of-bounds writes, high severity), CVE-2026-6474 (information leak via timeofday()), and CVE-2026-6475 (symlink follow allowing file overwrite during pg_basebackup/pg_rewind). The advisory lists affected versions (14–18) and credits multiple external reporters. PostgreSQL 14 is also announced to reach end-of-life on November 12, 2026, prompting production users to plan upgrades. Operators should apply patches promptly to mitigate high-severity risks.
UNLOGGED Tables in PostgreSQL: When Speed Matters More Than Durability by Chandan Shukla SQLServerCentral PostgreSQL Learn about unlogged tables and how they work in PostgreSQL. ★ ★ ★ ★ ★ ★ ★ ★ ★ ★ ( 3 ) Log in or register to rate You rated this post out of 5. Change rating 2025-11-21 5,018 reads Discuss
UNLOGGED Tables in PostgreSQL: When Speed Matters More Than Durability by Chandan Shukla SQLServerCentral PostgreSQL Learn about unlogged tables and how they work in PostgreSQL. ★ ★ ★ ★ ★ ★ ★ ★ ★ ★ ( 3 ) Log in or register to rate You rated this post out of 5. Change rating 2025-11-21 5,017 reads Discuss
A production outage occurred when a PostgreSQL cluster hit transaction ID wraparound after teams disabled autovacuum following an earlier incident where it was mistakenly blamed. Autovacuum stayed off and long-term maintenance consequences were not considered, allowing transaction IDs to age out and force a disruptive recovery. The post highlights human error in database ops, the critical role of autovacuum for preventing wraparound, and the operational risk of shortcut fixes without understanding maintenance features. This matters to engineers running PostgreSQL at scale because it underscores how background housekeeping safeguards data integrity and availability and why operational changes must be reviewed and monitored.