During the night, you receive an alert concerning your SQL Server failover cluster or your availability groups and you’re in panic mode because the message displayed is : a failover has occured .. see the log for more details …
So keep quiet and after connecting to your environment you are not able to find anything … What was happen? Maybe someone has triggered manually a failover and you are not aware of it. I’m sure by reading the previous sentences many of you will recognize but the real question is: is it possible with a Windows failover cluster to disting a manual failover from an automatic failover?
The answer is yes and one way to find out the response is to take a look at the cluster.log. In fact, you have an record entry that clearly identifies a manual failover of resources:
[RCM] rcm::RcmApi::MoveGroup: (, 1, 0, MoveType::Manual )
As a reminder, this is the resource control monitor [RCM] that is responsible to perform actions according to the state of a resource. In fact, when you trigger a manual failover the MoveGroup API is called with a identified parameter MoveType::Manual
Let’s me know if you find a other way to discover a manual failover 🙂
Happy failover (or not) !
By David Barbarin