Serviceguard NFS Toolkit A.11.11.06, A.11.23.05 and A.11.31.05 Administrator's Guide HP-UX 11i v1, v2, and v3

Figure 1-1 CFS versus Non-CFS (VxFS) Implementation
In a Serviceguard CFS environment, files and filesystems are concurrently accessible on multiple
nodes. When a package fails over, the adoptive systems do not have to mount the disks from the
failed system because they are already mounted. There is a new multi-node package that runs
on each server in the cluster and exports all the cluster filesystems. The exported filesystems do
not have to be reexported when a package fails over. These factors may reduce failover time.
The files and filesystems can be shared and accessed concurrently within the cluster. However,
file sharing and access from outside the cluster will still require NFS — client systems which are
not members of the CFS cluster will use NFS to access the shared filesystems.
Cross mounting (with nfs_xmnt script) is not needed since you can use CFS to share files and
filesystems within the cluster.
Figures 1–2 and 1–3 also show how files and filesystems are accessed differently in a CFS
environment versus a non-CFS environment. In a non-CFS environment, clients must access the
server which exports a specific filesystem. In a CFS environment, clients can access the cluster
via a load balancer or another mechanism such as a DNS round-robin scheme (represented by
the cloud in Figure 3.) Each NFS client can be directed to the server which currently has the most
capacity available. This approach has a limitation that clients are bound to a particular NFS server
when they issue the mount command.
14 Overview of Serviceguard NFS