MySQL: If an Active node has issues with DB connection and then another node takes over the active state, after the connection is restored the previously Active node remains in set to status Standby and after the failover delay passes it becomes Unavailable again instead of remaining is Standby status (Active since is not updated after restoring the connection).
Steps to reproduce:
- Set up a High availability cluster with 1 active node and 1 standby node
- Turn of the network for the node that is in Active status
- Wait for the failover period to pass
- Observe that the previously active node is now in status Unknown and that the standby node is now active
- Restore the connection to the node in status unavailable
- Observe that The Unavailable node is now in status Standby, but the Last access is not updated every 5 sec
- Wait for Failover delay to pass
Result: even though the connection is restored, the standby node becomes unavaiable again, last access is not updated.
Expected: Last access should be updated each 5 sec and node should be in Standby mode after the connection is restored.