Managing Systems and Workgroups: A Guide for HP-UX System Administrators

Setting Up and Administering an HP-UX NFS Diskless Cluster
NFS Diskless Questions and Answers
Chapter 10942
Answer: The best approach is to run SAM on a system where the resource is
unconfigured, select the resource, and select “Add Unconfigured”. This
gives you the option of adding the cluster-wide resource to just the local
system, or to all systems in the cluster where the resource is not
configured.
Question: I have a cluster-wide resource, but I have made some local changes to the
configuration of that resource on some of the systems in the cluster.
When I add a new system to the cluster, how is the resource configured
on that system?
Answer: Local tasks performed on a cluster-wide resource do not affect the
cluster-wide configuration of that resource. The resource will be
configured on the new client the way it was added, or last modified
cluster-wide.
For example, suppose you do a cluster-wide add of a printer with the
fence priority set to 1. Then you do a cluster-wide task to modify the
fence priority to 4. At that point, if a new system is added to the cluster,
the printer is configured on that system with a fence priority of 4.
Question: If I have removed a cluster-wide resource from a system in a cluster, and
then I add that resource back to that system but with a different local
identifier (that is, a different local mount point for a file system, or a
different local name for a printer), will future cluster-wide tasks affect
this system?
Answer: No, SAM treats the resource on that system as a different resource.
A cluster-wide resource is uniquely identified by two values: the
identifier by which the resource is referred to on systems where it is a
remote resource (for example, the local mount point of an NFS file
system) and the location of the resource (for example,
hostname : mountpoint of an exported file system).
So if you have configured a resource on some systems in a cluster with
one local identifier, and on other systems in the cluster with a different
local identifier, SAM can’t assume that they are the same resource.
Question: What do I do if I have a relatively large cluster (say, 100 systems) and I
only want about half of them to have access to a resource?