After you have created and configured the Messenger cluster resources, you can use the following methods to test whether the Messenger agents fail over as expected:
In the HA Management Client, select a node where a Messenger agent is running, then click . The agent should fail over to another node and start running again. This can also be done from the command line using the crm_standby utility.
Using the crm_resource utility, manually migrate a node where a Messenger agent is running.
If you receive the error Unmanaged on node, see Step 4 in Testing Your Messenger Agent Installation on Each Node for troubleshooting assistance.
If you have set up your Heartbeat cluster in a test situation first, you can also use the following methods for testing failover:
Power off the node where a Messenger agent is running. The agent should fail over to another node and start running again.
Simulate the Messenger agent stopping unexpectedly (kill -9 agent_PID). If you have configured the Messenger agent resource for automatic restart, as described in Enabling Automatic GroupWise Agent Restart, the agent should start running again on the same node.
WARNING:These last two tests should not be performed on a production system because they can result in damage to Messenger databases.