We have a 2-node cluster which are connected to the same storage.
The resource owner of the cluster lost its power accidently, but it caused another node in the cluster panic.
We have checked the log of the panic node. It said "reservation conflict" before panic.
Does anybody know what is wrong ?
What storage use as shared ?
How many path have every node to this storage ?
Cluster node reserve storage LUN for dedicated access ( storage reservation) but at failover spare nodes can't set reservation for shared devices that cause panic.
Check this settings and recomendation:
Edited by: Nik on 10.08.2012 1:37
In case of maintenance, one of the node will be evacuated from the cluster, hence the cluster will not panic and the conflict error is not expected.
In this scenario what you are facing, it was unexpected that one of the node went down, hence the cluster panic. The error is expected in this scenario because the member node is not available to communicate. When the other node comes up the operation is normal.