master back in line and then my promoted server as slave what I have to do. I tried pg_rewind it's not working. Is there any procedure?
It is probably on the wrong timeline. You'd need to start it as a new follower.
Ok, how can we perform manual failover and failback?
Promote the follower with SELECT pg_promote(); Or run pg_ctl promote Create a standby.signal file on the current primary and restart it an run pg_rewind. You'd need to adjust your backup tool as the WAL files from a promoted secondary aren't and can't be used for ongoing point in time backups. Or learn and use at least repmgr when you can't use Patroni.
Обсуждают сегодня