I have a 2 node OES2sp1-L cluster with 1 resource - an NSS Volume which I'm able to fail over between the nodes with no problem. I'm testing AFP with this volume. But not sure AFP has anything to do with this problem which has happened twice now.

I use iManager to connect to each of the node's and add a user context. I hit the reload button for AFP and then test from a Mac that I can log in using that new context. I then attempt to fail over the resource to the other node to test it and it goes comatose. I attempt to fail it back to the original node and it goes comatose.

Troubleshooting so far - if I look in the clusterresource.load.out file I see an error 20892 trying to activate the pool, which seems to mean it thinks the pool is already active. I go back to the other node and go into nsscon and it says the pool is inactive. Another possible symptom - if I do an ncpcon and type volumes it hangs and returns nothing. All patches are up to date on the servers. Any ideas on how to troubleshoot further?