Top
Symfoware Server V12.0.0 Cluster Operation Guide
FUJITSU Software

3.6.2 Mutual Switch Maintenance

This section explains the procedure to perform maintenance on the standby node while running the active node at the same time.

3.6.2.1 Stopping the Cluster Applications on the Standby Node for Node Maintenance

Follow the steps shown below to stop the cluster application on the standby node only:

  1. Edit the /opt/SMAW/SMAWRrms/bin/hvenv.local file, and add "export HV_RCSTART=0".

    If the hvenv.local file does not exist, then create it.

    # vi /opt/SMAW/SMAWRrms/bin/hvenv.local
    Add export HV_RCSTART=0
  2. On the RMS tree in the [Cluster Admin] window, right-click the cluster application on the standby node to be stopped, and select [Offline] from the pop-up menu to stop it.

  3. On the RMS tree in the [Cluster Admin] window, right-click the system node (where the cluster application on the standby node that was stopped exists), and select [Shutdown RMS]. At this stage, select [Keep local Apps] as the option, and stop RMS.

Note

  • Ensure that the steps up to step 3 have been completed properly before performing the maintenance tasks. If the active node goes down at the point when steps 1 and 2 have been completed, Symfoware will be started on the supposedly stopped standby node, and it becomes the active node instead. As a result, this will compete with the maintenance tasks and unexpected behavior may occur.

  • If applying the Symfoware Server update patches, RMS on the node (onto which the patches are applied) must be stopped. This is because the Symfoware monitoring process that runs when linked with PRIMECLUSTER must be stopped.

  • By stopping RMS, all cluster applications on the system node where RMS is stopped will stop. When you have multiple cluster applications configured, ensure that stopping all other cluster applications will not cause any problems, before stopping RMS.

3.6.2.2 Standby Node Maintenance

Perform maintenance tasks such as patch application and hardware parts replacement.

3.6.2.3 Restarting the Cluster Applications Stopped on the Standby Node for Node Maintenance

Follow the steps shown below to restart the cluster application on the standby node only (that was stopped for maintenance):

  1. Edit the /opt/SMAW/SMAWRrms/bin/hvenv.local file, and delete "export HV_RCSTART=0".

    # vi /opt/SMAW/SMAWRrms/bin/hvenv.local
    Delete export HV_RCSTART=0
  2. Select [Tools]-[Start RMS] in the [Cluster Admin] window, and select the node to be started and click [Ok].

  3. On the RMS tree in the [Cluster Admin] window, right-click the cluster application to be started as the standby node, and select [Standby] from the pop-up menu to start it. If the state is [Fault], then select [Clear fault].

3.6.2.4 Mutual Switch

To perform mutual switch between the active and standby nodes, perform the steps shown below:

  1. Select and right-click the cluster application on the active node, and select [Offline] from the pop-up menu to stop it.

  2. Select and right-click the cluster application on the standby node, and select [Online] from the pop-up menu to start it.

  3. Select and right-click the cluster application that was previously on the active node, and select [Standby] from the pop-up menu to start it as standby.

3.6.2.5 Standby Node Maintenance after the Mutual Switch

To perform maintenance tasks on the new standby node after mutual switch, perform the steps from "3.6.2.1 Stopping the Cluster Applications on the Standby Node for Node Maintenance" to "3.6.2.3 Restarting the Cluster Applications Stopped on the Standby Node for Node Maintenance".