Change a UVMS cluster 6.2 in UVMS failover

This paragraph explains how to modify a UVMS architecture to migrate from a UVMS cluster 6.2 (not supported from version 6.3) to UVMS failover.

The UVMS cluster is installed on two machines. The cluster type is active/passive: so only one UVMS is started at any given time.

UVMS uses an external database.

The architecture must be modified by completing the following steps:

  1. Stop the secondary node of the UVMS cluster and make sure that primary node is started and accessible from the machine of secondary node.
  2. On the secondary node run the command "unidlt FROMCLUSTER", described, and make sure that the secondary node is removed from the list of cluster members (administration mode > Monitoring & audit > Cluster Monitoring).
  3. Stop the primary node of the UVMS cluster (and leave the secondary node stopped)
  4. Upgrade the primary UVMS node to the last version
  5. Start the primary UVMS node
  6. Activate “the check of startup in failover mode” on the primary node (refer to section "Start-up in Failover Mode")
  7. Uninstall the secondary UVMS node
  8. On the secondary node, install UVMS and select the option “Secondary Failover Installation”

Step 2 is important to ensure UVMS data consistency in UVMS after upgrade from 6.2. If this step is not done correctly before upgrade, you may execute the following command on the primary node to repair data inconsistency ("UJCENT" is a keyword and should not be changed):

unidlt NODE -company UJCENT -type MS -offline

This command removes any reference to unnecessary UVMS 6.2.

The primary UVMS is active and now has a failover UVMS on the secondary node.

As of version 6.10.41, documentation new updates are posted on the Broadcom Techdocs Portal.
Look for Dollar Universe.