Implementing a clustered NFS environment on Linux

In addition to the traditional exporting of GPFS™ file systems using the Network File System (NFS) protocol, GPFS allows you to configure a subset of the nodes in the cluster to provide a highly-available solution for exporting GPFS file systems using NFS.

Note: This feature is available with IBM Spectrum Scale™ Standard Edition or higher.

The participating nodes are designated as Cluster NFS (CNFS) member nodes and the entire setup is frequently referred to as CNFS or a CNFS cluster.

In this solution, all CNFS nodes export the same file systems to the NFS clients. When one of the CNFS nodes fails, the NFS serving load moves from the failing node to another node in the CNFS cluster. Failover is done using recovery groups to help choose the preferred node for takeover. For the NFS client node to experience a seamless failover, hard mounts must be used. The use of soft mounts will likely result in stale NFS file handle conditions when a server experiences a problem, even though CNFS failover will still be done.

Currently, CNFS is supported only in the Linux environment. For an up-to-date list of supported operating systems, specific distributions, and other dependencies, refer to the IBM Spectrum Scale FAQ in IBM® Knowledge Center (www.ibm.com/support/knowledgecenter/STXKQY/gpfsclustersfaq.html).