Comments - Maxscale rejoin after auto-failover
Content reproduced on this site is the property of its respective owners,
and this content is not reviewed in advance by MariaDB. The views, information and opinions
expressed by this content do not necessarily represent those of MariaDB or any other party.
Looks like a transaction was completed on the master but was not replicated to the slaves before the failover took place. This means the master has diverged from the rest of the cluster and should be recreated from a backup.
This is strange, I reproduced it with full durable settings, semi sync (configured with wait_point=after_sync)
It should either be replicated, or not committed. I cannot show it now, but I could see a replication thread got killed in logs
I'm talking about 1-2 commits missed for about 500 writes/sec
Sorry for the late reply. If you can reliably reproduce this, I'd like to ask you to open a bug report on the MariaDB Jira under the MaxScale project if this still happens with the latest 2.3 version of MaxScale. This way we'll be better able to track the progress of this issue.