I'm having an issue with a new failover cluster that I just created where I can kill the clustered service 4 times through the task manager (or simulate failure) and the service and role are then marked as failed rather than failing over.
Why is this? It seems that this is new behavior in 2012 R2. I have multiple clusters in the same setup running in 2012 R1 and this does not happen. If I kill it 4 times the entire role fails over to the second node.
The only difference I can see in the cluster itself is that the other cluster on R1 is using Node Majority for the quorum settings and that is not the default chosen in the wizard setup in R2, it is Disk witness. I don't see an option to use node majority but either way it should work. I can also move the role between nodes successfully so that isn't really an issue.
Why is the 4th service failure within 15 minutes not causing a failover? The policies are pretty much default:
--If resource fails, attempt to restart on current node (15 min, 3 max, 1 second delay)
--If restart unsuccessful, fail over all resources in this role
--If all the restart attempts fail.....is checked