HP-UX HB v13.00 Ch-15 - Serviceguard
HP-UX Handbook – Rev 13.00 Page 42 (of 108)
Chapter 15 Serviceguard
October 29, 2013
WEIGHT_DEFAULT
Change
9
YES
USER_NAME
Change, add
remove
ALL
YES
USER_HOST
USER_ROLE
VOLUME_GROUP
Add/Delete
ALL
YES
Updating the arbitration device (cluster lock VG/PV, LUN or Quorum
Server)
Use the procedures that follow whenever you need to change the device file names of the cluster
lock physical volumes – for example, when you are migrating cluster nodes to the agile
addressing scheme available as of HP-UX 11i v3.
1. Locate or reconstitute the cluster configuration file using:
$ cmgetconf cluster.ASCII
2. Run cmcheckconf to check the existing configuration:
$ cmcheckconf cluster.ASCII
3. Modify the arbitration values or change from one type of arbitration to another (lock
VG/PV to quorum server etc) in the ASCII file.
4. Run cmapplyconf to apply the configuration. (May require cluster halt):
$ cmapplyconf –f –C cluster.ASCII (–f bypasses the modification query)
Reconfiguring a Halted Cluster
All versions of Serviceguard can be reconfigured when the cluster is halted. Depending on the
version of Serviceguard in use, it may be necessary to halt the cluster in order to perform the
cmapplyconf. Perform cmcheckconf on a modified cluster configuration ASCII file to learn
whether it is necessary to halt the cluster (or see above table).
Using -k or -K option with the cmcheckconf and cmapplyconf commands can significantly
reduce the response time.
Deleting Nodes from the Configuration While the Cluster is Running
Whether Serviceguard is running or not, it will allow the user to remove a node from the cluster
if that node is no longer reachable (via hacl-cfg ports). Comment all NODE_NAME references
for the missing node from both package and cluster configuration files, then perform
cmapplyconf on all of them. If the departing node is still attached to the subnet, deactivate hacl-
cfg in /etc/inetd.conf on the departing node and do ‘inetd –k ; inetd’ to restart inetd services.
NOTE Running cmapplyconf on a node that you are trying to remove from the cluster will
generate an error message.
Use the following procedure to delete a node with HP-UX commands. In this example, nodes
Node1, Node2 and nodeC are already configured in a running cluster named cluster1, and you
are deleting node nodeC.