2 cluster issues with our OES11SP1 2 node file cluster. Both nodes are fully patched VMs with 9 shared file volumes. Each node has 2 CPUs and 16 GB of RAM. The servers do not have local replicas but query other eDir servers. Time is in sync.

1) Migrating volumes between the two nodes more often than not leads to the source node freezing up rather than migrating anything. The cluster unload script completes without error but the volume does not detach and therefore does not migrate. The source node also usually displays a high load value (in top) in the order of 9 to 12 but it's not CPU or memory based. No volume will succesfully move. As mentioned, the cluster unload script output looks fine and there's nothing in /var/log/messages at all. The "fix" is to reboot the node but that's decidedly suboptimal.

2) Sometimes we see issues on a file volume where a user (such as myself) has inherent collective rights to see 4 top level folders and then more direct group membership that allows me to see a further two top level folders. Except every 1 to 2 weeks I lose the ability to see the 2 groupy folders but can see the other 4 folders just fine. Sometimes a migration of the volume fixes the issue in terms of what I can see and less often we have to reboot a node.

These are production servers so the impact is high and like I said I don't see a huge amount of helpful log output.