pg_upgrade: the tricky and dangerous parts
Postgres FM
00:00
PIGI Upgrade Tough and Dangerous Parts
The problem is related to how we upgrade standby nodes and post quiz documentation describes the recipe involving R-Sync. If you really want to minimize downtime in case of without logical downtime, usually a few minutes, we want to achieve a few minutes even with clusters huge. A best mode for R- Sync would be check some. But it would take a lot. Maybe it's easier to recreate replica than just wait until R-sync check some is working. So I would encourage you to check this explore this and communicate with author maybe I will do it as well.
Transcript
Play full episode
