RedHat Cluster Suite - GFS shared mount point freezing
 

I configured a cluster of two Red Hat 5 nodes using Red Hat Cluster suite. Both nodes are connected to a shared MSA storage via mini SAS cables.

The storage is configured as a logical volume and mounted as GFS on both nodes at the same time.

Everything is running smoothly on the cluster, but each 5-7 days the GFS storage will freeze and it will not be accessible. everything else on the servers continue to run normally, and no error messages are shown in /var/log/messages.

The storage is accessible only if I reboot or fence one of the nodes (maybe the issue is related to the GFS locking??!)

Any idea what could be the problem, please ask about any useful info that can be provided from the servers.

Hello Guys,

Any comment? :)

Although I'm happily using OCFS2 for clustered filesystems, I am very interested in the outcome / resolution to your problem.

If there is no GFS-related chatter getting logged, and the GFS utilities (are there any?) and clustered LVM tools (are you using CLVM?) aren't able to tell you anything useful, I don't have good suggestions for troubleshooting.

Contact Red Hat - that Red Hat Cluster channel subscription ain't cheap.