3.5.1 Matrix Server Administration Guide
Chapter 8: Configure MxFS-Linux 133
Copyright © 1999-2007 PolyServe, Inc. All rights reserved.
one of the networks already in use in the cluster. The administrator then
specifies both a primary node to host the Virtual NFS Service and an
ordered set of backup nodes to host the Virtual NFS Service in case of
failover. Finally, the administrator selects one Export Group to be
associated with the Virtual NFS Service. This Export Group defines which
exports will be available via NFS for this Virtual NFS Service.
It is possible to create multiple Virtual NFS Services and associate the
same Export Group with each one. This enables “scale-out” NFS service
for the same filesystems across multiple NFS servers in the cluster. It is
further possible to set up a DNS round robin configuration such that a
single IP address (or DNS name) can be used by all clients, yet the client
connections will be evenly distributed among the several Virtual NFS
Services running on separate nodes of the cluster. Other configurations
are possible.
Supported NFS Versions
MxFS-Linux supports versions 2 and 3 of the NFS file sharing protocols
based on UDP and TCP.
Tested Configuration Limits
MxFS-Linux has been tested to the following configuration limits:
• Up to 16 nodes in the matrix.
• 512 independent PSFS filesystems per matrix.
• 128 Virtual NFS Services per matrix.
• 64 Export Groups per matrix.
• 1024 export records per Export Group.
RPC Program Usage
RPC program 300277 (VNSM) supplements the standard Network Status
Monitor RPC program 100024 (NSM). VNSM allows the virtualization of
status monitoring on a per-Virtual NFS Service basis. One status
monitoring process on each cluster node is responsible for a varying
number of logically independent services (Virtual NFS Services). The
VNSM RPC program is used to manage the responsibilities of the status