NFS failover

As part of GPFS recovery, the CNFS cluster failover mechanism is invoked. It transfers the NFS serving load that was served by the failing node to another node in the CNFS cluster. Failover is done using recovery groups to help choose the preferred node for takeover.

The failover mechanism is based on IP address failover. The CNFS IP address is moved from the failing node to a healthy node in the CNFS cluster. In addition, it guarantees NFS lock (NLM) recovery.

Failover processing may involve rebooting of the problem node. To minimize the effects of the reboot, it is recommended that the CNFS nodes be dedicated to that purpose and are not used to run other critical processes. CNFS node rebooting should not be disabled or the failover reliability will be severely impacted.