Replicate the Error
Do a DR Test and change the IP address of the server. Without waiting for the IP to completely go into effect, do a Force Start of the Windows Failover Cluster Manager.
The cluster will come up but the Roles will unable to be restarted. The nodes will be running. The databases will in a “Not Synchronizing / Recovery Pending” state.
Avoiding this Error
Wait until the IP address change goes into affect before restarting the cluster. Do NOT use the Force Start option of Windows Failover Cluster.
To fully avoid this issue with Always On during DR Test, make sure to wait long enough for the new IP address to go into effect before re-enabling the cluster in Windows Failover Cluster.
Fix Option 1
Delete the Always On AG group and recreate it.
Fix Option 2
Do a force failover from the SQL Server side instead…
ALTER AVAILABILITY GROUP xxxxxxxx FORCE_FAILOVER_ALLOW_DATA_LOSS;
Comments