White Paper

b) In the nfs2.conf file, set the PACKAGE_NAME to SG-NFS2.

c) The PACKAGE_TYPE should be set to the default value (FAILOVER).

d) Set the NODE_NAME variables for each node that can run the package. The

first NODE_NAME must specify the primary node, followed by adoptive

node(s) in the order in which they will be tried:

NODE_NAME thyme

NODE_NAME basil

e) Set the RUN_SCRIPT and HALT_SCRIPT variables to the full path of the

control script, with no timeout:

RUN_SCRIPT /etc/cmcluster/nfs/nfs1.cntl

RUN_SCRIPT_TIMEOUT NO_TIMEOUT

HALT_SCRIPT /etc/cmcluster/nfs/nfs1.cntl

HALT_SCRIPT_TIMEOUT NO_TIMEOUT

Similarly, set the RUN_SCRIPT and HALT_SCRIPT to

/etc/cmcluster/nfs/nfs2.cntl in nfs2.conf.

f) In nfs1.conf and nfs2.conf, set the DEPENDENCY_NAME, DEPENDENCY

CONDITION, and DEPENDENCY_LOCATION variables so that the failover

package can run only if the export package is already running:

DEPENDENCY_NAME SG-NFS-XP-1-dep

DEPENDENCY_CONDITION SG-NFS-XP-1=UP

DEPENDENCY_LOCATION SAME_NODE

g) To monitor NFS services, set the SERVICE_NAME variable:

SERVICE_NAME nfs1.monitor

h) In the nfs2.conf file, set SERVICE_NAME to nfs2.monitor. The

SERVICE_NAME variable in the nfs.conf scripts must match the

NFS_SERVICE_NAME[0] variable in the hanfs.sh scripts. If you do not want

to monitor NFS services, leave this variable commented out.

i) Set the SUBNET variable in both nfs1.conf and nfs2.conf:

SUBNET 15.13.112.0

Starting a Serviceguard NFS failover package

1. Copy the configuration files to each server in the cluster:

# rcp –p /etc/cmcluster/nfs/* {hostname}:/etc/cmcluster/nfs

2. Verify the cluster and package configuration files on each server:

# cmcheckconf –k –v –C /etc/cmcluster/cluster.conf –P

/etc/cmcluster/nfs/nfs1.conf –P /etc/cmcluster/nfs/nfs2.conf

3. Verify and apply the cluster package configuration files on a single server: