Re: Group replication stopped on one node after rebootClusterFromCompleteOutage

[Frederic Descamps]

This means that you have one of the node that has committed a transaction but one of the nodes part of the cluster didn't get it yet and cannot get it back because for some reasons the binary log entry (should be in the last binary log-1), is not present anymore.

Why ? That's what you need to answer. Possible reasons are that you explicitly removed the binlogs, or that the binlogs are not resilient on your Virtual environment or that you reduced durability (sync_binlog).

But if for some reason after a full outage a recovery using the binary logs is not possible, you have the possibility to recover by yourself (maybe via a backup or a snapshot, or by using another node that was an asynchronous replica, as clone might take some time).