Some items of note regarding the Messaging Server and high availability (applies to both Veritas Cluster and Sun Cluster) installation:
High availability for the Messaging Server is not installed by default; be sure to select High Availability Components from the Custom Installation menu of the Java Enterprise System installer.
When running the installation, make sure that the HA logical host name and associated IP addresses for Messaging Servers are functioning (for example, active). The reason for this is because portions of the installation will make TCP connections using them. Run the installation on the cluster node currently pointed at by the HA logical host name for the messaging server.
Make sure that the msg_svr_base is on the shared file system; otherwise, high availability will not work correctly. For example, after failing over to another node, the servers will no longer see the data accumulated by the servers on the failed node.
When you are asked for the fully-qualified domain name of the Messaging Server host during initial runtime configuration, be sure to specify the fully-qualified HA logical host name for the Messaging Server. During the install, TCP connections using this logical host name will be attempted.
When you are asked for the IP address of Messaging Server when you run ha_ip_config, be sure to specify the IP address associated with the logical host name for Messaging Server. Do not use the IP address for the physical host.
Clustering software needs to be installed before installing and configuring the current version of Messaging Server. Run the installation on the cluster node currently pointed at by the HA logical host name for Messaging Server. Use the cluster alias when prompted for any node names.
When running the Messaging Server Initial Runtime Configuration (see 1.3 Creating the Initial Messaging Server Runtime Configuration) sure to specify the fully-qualified HA logical host name of the cluster of Messaging Server.
Configure Messaging Server using the cluster host name. If you do otherwise, then you’ll need to re-configure a second time using the cluster host name.