Replication shinanigans

postgresql replication

In Debian, package upgrades trigger restart of the related service. After I've restarted all my postgresql nodes (1 primary + 2 hot standby) at the same time, I've realized one of the secondaries were not able to process WAL files any further. Checking the logs gave me an idea: 2019-07-09 13:20:32 UTC [25065-1] postgres@[local] FATAL: the database system is starting up gzip: stdin: unexpected end of file ERROR: Remote 'barman get-wal' command has failed!

Read more →