16.2.3 Node switching procedure

The following figure shows the node switching procedure for the host management model:

Figure 16‒2: Node switching procedure

1. Detect the failure.

   If the Administration Agent of the executing node 2 is down due to a failure, the cluster software of the standby node detects the failure of the Administration Agent of the executing node 2.

2. Switch the node.

   The cluster software of the standby node switches the virtual host 2 of the executing node 2 to the virtual host 2 of the standby node. In this case, the shared disk device is switched and the IP address of the virtual host is inherited.

3. Invoke the Administration Agent of the standby node.

   The cluster software on the standby node starts Administration Agent in only the case where a virtual IP address is used for Administration Agent.

4. Invoke Management Server of the standby node.

   The cluster software on the standby node starts Management Server in only the case where a virtual IP address is used for Management Server.

5. Invoke the logical server.

   The Administration Agent invokes the virtual host 2 logical server of the standby node.

The Administration Agent, Management Server, and logical server are invoked by the start server script registered in the cluster system. For details on the start server script, see the OS manual in use.

To Page Top