Description
During auto-failover we perform some safety checks for services and kv before we hit the quorum check. That in itself is fine, but the errors that are logged imply that an auto-failover might otherwise have happened, so this requires knowledge that these checks are performed before the quorum check. We should consider moving these checks after the quorum check, so that we can more clearly see why an auto-failover is not possible.
Attachments
Issue Links
- relates to
-
MB-58753 Auto-failed over node attempting to failover the healthy node in the cluster
- Closed