HSG80 ACS Solution Software V8.6 for IBM AIM Installation and Configuration Guide

Table Of Contents
424 HSG80 ACS Solution Software Version 8.6 for IBM AIX Installation and Configuration Guide
Normal Operation of the Agent on an HACMP Cluster
Once the Agent has been installed onto each node of the cluster and the Client(s) has been
installed and configured, the normal operating mode of the Agent in the HACMP
environment may be activated by the following procedure.
1. Bring up the HACMP Cluster
Installation, configuration and operation of an HACMP cluster is described in the IBM
publication: High Availability Cluster Multi-Processing for AIX Installation Guide
Version 4.3 (SC23-4278-00). Examples of typical HACMP configurations is described
in IBM Redbook: SG24-4498 -00 HACMP/6000 Customization Examples which is
available directly from IBM or on the internet at:
http://www.redbooks.ibm.com/abstracts/sg244498.html
2. Configure the Agent Within HACMP
After the Agent has been installed and the HACMP cluster software is running, the
Agent should be configured on each node and started on one node of the cluster. This
is achieved using the unique features of the HACMP utility script:
/usr/stgwks2/steam/bin/HACMP_config.sh.
3. Start the Agent Within HACMP
Use the Config.sh utility to start the Agent on the local node. The Config.sh
utility automatically calls the HACMP_config.sh script containing the necessary
functions. These functions are used to control the operation of the Agent.
There are five new menu options designed to support the operation of the Agent in the
HACMP environment. The implementation details of these functions are described in
the section on Agent configuration. Their application within the HACMP environment
is described as follows:
a. Option 21, CHECK ALL Nodes for Running Agent
This function determines if there are any instances of the Agent process running on
any node in the cluster. It automatically polls each node through each known IP
address that is currently active and reports the results back to the local operator. It
is important to note that a node that is temporarily isolated, or is waiting to rejoin
the cluster following a
node_down event, will not be reported by this utility.
b. Option 22, STOP Running Agent on ALL Nodes
This function is used to stop any instances of the Agent process running on any
node in the cluster. Since it uses the same polling mechanism as Option 21 above, it
can only act on nodes that are currently connected to the local node by at least one
known IP address.