MCS Problem when nfs hangs

[Sergio Crespo]

Hi
we've been having some problems with our 2 node cluster. When nfs hangs, the package stops responding, and the only way, and as far as I know, to solve the situation, is to reboot both nodes.
Does anyone knows a way to recover from this situation without rebooting?
We have 2 N machines, with MCS A.11.12, with Gigabit lan cards, a crossover to the heartbet, with Nodetimeout 6s and heartbeat timeout 1s. Application(SAP).

[Rich Wright]

Your need to fix the NFS hang problem. First, make sure that automounter is disabled and not used.
Then check to be sure that the network path from NFS Host to NFS Client is good. Verify that speed and auto-negotiate flag are set the same on the servers and the router or hub.
Also, you probably already have it, but with a 2-node cluster, a lock disk is required.

[Robert Gamble]

This suggestion will not immediately solve your problem, but it needs to be addressed: It appears NFS is a single point of failure, and needs to removed from this high availiabilty (HA)environment. If NFS is still needed, then try providing NFS from an HA system or package.