Complete procedure for switching a domain manager

This section summarizes the steps required to replace a running domain manager with its backup and to complete the procedure by restoring the original domain manager to its function. Follow these steps to make sure that no overlapping problems arise with obsolete versions of the Symphony file. You can also follow these steps to switch a master domain manager or a dynamic domain manager. The steps are documented for four scenarios:
Planned outage
The domain manager is replaced with its backup for planned maintenance work (for example, an upgrade of the operating system).
Unplanned outage
The domain manager is replaced with its backup because of an unexpected failure or malfunction.
Short-term
The domain manager is expected to return to service before the next new production period turnover (run of the JnextPlan job).
Long-term
The domain manager is not expected to return to service before the next new production period turnover (run of the JnextPlan job).
Table 1. Complete procedure for switching a domain manager in case of a planned outage.
Planned outage
Short-term Long-term
1. Switch the domain manager to a backup workstation. Use either the conman switchmgr command or the Dynamic Workload Console. For more information about both methods, see Switching a master domain manager or dynamic domain manager.

2. Check that the message boxes for the domain manager undergoing maintenance are large enough not to fill up before it is restored. Increase their size if necessary.

 1. Switch the domain manager to a backup workstation. Use either the conman switchmgr command or the Dynamic Workload Console. For more information about both methods, see Switching a master domain manager or dynamic domain manager.

2. Check that the message boxes for the domain manager undergoing maintenance are large enough not to fill up before it is restored. Increase their size if necessary.
3. Shut down HCL Workload Automation processing on the domain manager undergoing maintenance. 3. Shut down HCL Workload Automation processing on the original domain manager undergoing maintenance.
4. In the HCL Workload Automation database, assign the role of domain manager to the backup workstation, using either the composer mod cpu <workstation_name> command or the Dynamic Workload Console. This is done by changing the workstation type in the database from MANAGER to FTA on the original domain manager and from FTA to MANAGER on the backup.
5. Set the workstation running the original domain manager to ignore, using either the composer mod cpu <workstation_name> command or the Dynamic Workload Console.
6. Run JnextPlan to generate the new production plan so that the backup master domain manager is removed from the plan.
When ready to restore the ownership of the domain to the original domain manager: When ready to restore the ownership of the domain to the original domain manager:
7. Remove the ignore flag from the workstation running the original domain manager.
8. Run JnextPlan to generate the new production plan so that the backup master domain manager is reinserted in the plan.
4. Switch from the backup workstation to the domain manager using one of the methods indicated in step 1. 9. Reassign ownership of the domain to the original domain manager in the HCL Workload Automation database, using either the composer mod cpu <workstation_name> command or the Dynamic Workload Console. This is done by changing the workstation type in the database from MANAGER to FTA on the original backup and from FTA to MANAGER on the original domain manager.

10. Optionally, remove in the original domain manager the conman start command from the init procedure and delete any existing copies of the Symphony, Sinfonia, and message box files. This step is recommended to avoid that any outdated symphony present in the computer is automatically triggered at the first startup. You can add conman start again later.
5. Link the domain manager from the master to download a fresh version of the Symphony file.
11. Switch from the backup workstation to the domain manager using one of the methods indicated in step 1.
12. Link the domain manager from the master to download a fresh version of the Symphony file.
Table 2. Complete procedure for switching a domain manager after an unplanned outage.
Unplanned outage
Short-term Long-term
1. Switch the domain manager to a backup workstation. Use either the conman switchmgr command or the Dynamic Workload Console. For more information about both methods, see switchmgr.

2. Check that the message boxes for the failing domain manager are large enough not to fill up before it is restored. Increase their size if necessary.

 1. Switch the domain manager to a backup workstation. Use either the conman switchmgr command or the Dynamic Workload Console. For more information about both methods, see switchmgr.

2. Check that the message boxes for the failing domain manager are large enough not to fill up before it is restored. Increase their size if necessary.
3. In the HCL Workload Automation database assign the role of domain manager to the backup workstation, using either the composer mod cpu <workstation_name> command or the Dynamic Workload Console. This is done by changing the workstation type in the database from MANAGER to FTA on the original domain manager and from FTA to MANAGER on the backup.
4. Set the workstation running the failing domain manager to ignore, using either the composer mod cpu <workstation_name> command or the Dynamic Workload Console.
5. Run JnextPlan to generate the new production plan so that the backup master domain manager is removed from the plan.
When ready to restore the ownership of the domain to the original domain manager: When ready to restore the ownership of the domain to the original domain manager:
3. Optional:
  • A. In the original domain manager, remove the conman start command from the init procedure and delete any existing copies of the Symphony, Sinfonia, and message box files.
    Note: This step is recommended to avoid that any outdated symphony present in the computer is automatically triggered at the first startup. You can add conman start again later.
    B. After an unplanned outage, the fault-tolerant agent needs a new Symphony file. Perform the following steps on the current master domain manager:
    • i. Verify that the current master domain manager is linked to all agents except the old master domain manager.
    • ii. Shut down all HCL Workload Automation processes (unlink from all agents).
    • iii. Rename Sinfonia as Sinfonia.orig.
    • iv. Copy Symphony to Sinfonia.

      You now have identical Symphony and Sinfonia files.

 6. Remove the ignore flag from the workstation running the original domain manager.
4. Switch from the backup workstation to the domain manager using one of the methods indicated in step 1. 7. Reassign ownership of the domain to the original domain manager in the HCL Workload Automation database, using either the composer mod cpu <workstation_name> command or the Dynamic Workload Console. This is done by changing the workstation type in the database from MANAGER to FTA on the original domain manager and from FTA to MANAGER on the backup.
8. Optional: In the original domain manager, remove the conman start command from the init procedure and delete any existing copies of the Symphony, Sinfonia, and message box files.
Note: This step is recommended to avoid that any outdated symphony present in the computer is automatically triggered at the first startup. You can add conman start again later.
5. Link the domain manager from the master to download a fresh version of the Symphony file.
9. Switch from the backup workstation to the domain manager using one of the methods indicated in step 1.
10. Link the domain manager from the master to download a fresh version of the Symphony file.
11. Run JnextPlan to generate the new production plan so that the backup master domain manager is reinserted in the plan.