Building a Scalable NFS Server With GPFS

Overview

As server farms continue to get larger the need for a scalable network file system based on the industry standard NFS protocols becomes more apparent. Linux NFS will not handle these large server farms from a single server. More robust storage subsystems are prohibitively expensive and therefore defeat one of the reasons for using Linux in the server farm in the first place. This paper describes a solution to this problem using multiple NFS servers sharing a common file system managed by an underlying GPFS cluster. The solution is easy to design, easy to manage, feature rich and highly reliable.