IBM SG24-5131-00 Stopping Cluster Services on a Node, Automatically Restarting Cluster Services

are started in sequential order - not in parallel. The output of the command run on the remote node is returned to the originating node. Because the command is executed remotely, there can be a delay before the command output is returned.

8.2.2.1 Automatically Restarting Cluster Services

You can optionally have cluster services start whenever the system is rebooted. If you specify the -Rflag to the rc.cluster command, or specify restart or both in the Start Cluster Services SMIT screen, the rc.cluster script adds the following line to the /etc/inittab file.

hacmp:2:wait:/usr/sbin/cluster/etc/rc.cluster -boot> /dev/console 2>&1

# Bring up Cluster

At system boot, this entry causes AIX to execute the /usr/sbin/cluster/etc/rc.cluster script to start HACMP Cluster Services.

Note

Be aware that if the cluster services are set to restart automatically at boot time, you may face problems with node integration after a power failure and restoration, or you may want to test a node after doing maintenance work before having it rejoin the cluster.

8.2.2.2Starting Cluster Services with IP Address Takeover Enabled If IP address takeover is enabled, the /usr/sbin/cluster/etc/rc.cluster script calls the /etc/rc.net script to configure and start the TCP/IP interfaces and to set the required network options.

8.2.3Stopping Cluster Services on a Node

You stop cluster services on a node by executing the HACMP /usr/sbin/cluster/etc/clstop script. Use the HACMP for AIX Stop Cluster Services SMIT screen to build and execute this command. The clstop script stops an HACMP daemon or daemons. The clstop script starts all the cluster services or individual cluster services by calling the SRC command stopsrc.

Using the C-SPOC utility, you can stop cluster services on a single node or on all nodes in a cluster by executing the C-SPOC /usr/sbin/cluster/utilities/cl_clstop command on a single node. The C-SPOCcl_clstop command performs some cluster-wide verification and then calls the clstop command to stop cluster services on the specified nodes. The nodes are stopped in sequential order—not in parallel. The output of the command that is run on the remote node is returned to the originating

Cluster Management and Administration 157