Managing HP Serviceguard for Linux, Sixth Edition, August 2006

Configuring Packages and Their Services
Creating a Disk Monitor Configuration
Chapter 6 207
NOTE The
SERVICE_CMD
must include the string cmresserviced” or the
cmconfigres will fail to find the disks to be monitored. cmconfigres
looks for SERVICE_CMD entries including the text “cmresserviced” to
determine which disks to monitor.
It is important to set SERVICE_RESTART to an empty string (““).
After setting up the package ASCII files and control scripts, use
cmcheckconf and cmapplyconf to apply the package configuration, as
usual. Then create a monitor configuration file using the cmconfigres
command. This can be done for all nodes in the cluster at one time, or it
can be done a node at a time. The first approach is easiest when first
setting up the cluster, but the second approach may be more appropriate
when modifying the monitoring configuration at a later time.
Configuring All Disks for Monitoring
During initial cluster setup, to create a monitoring configuration that
includes all disks for all packages that contain a monitoring service, use
the following command:
# cmconfigres create
This command builds the $SGCONF/cmresmond_config.xml file on each
node in the cluster. The command operates by examining a node at a
time and scanning all the package control scripts for the packages that
might at some time be running on that node. Note that the content of the
$SGCONF/cmresmond_config.xml files is different from node to node if
the list of packages that can run on each node is different from other
nodes.
The following is a specimen of a cmresmond_config.xml file:
<?xml version="1.0" encoding="UTF-8"?>
<resourceMonitorServerConfig
xmlns:xsi="http://www.w3.org/2001/XMLSchema-instanc
e" xsi:noNamespaceSchemaLocation="ResMonServerConfig.xsd">
<runDirectory>/usr/local/cmcluster/run</runDirectory>
<htmlRoot>/usr/local/cmcluster/htdocs</htmlRoot>
<logFile>/usr/local/cmcluster/log/ResMonServer.log</logFile>
<logLevel>0</logLevel>
<port>3542</port>