Migrating node-by-node

Node-by-node migration is supported only when you migrate from System Automation for Multiplatforms V2.3 or higher. To migrate the nodes of a domain one by one has the advantage that System Automation for Multiplatforms remains available during the migration.

For more information about how to minimize downtime, see Checking prerequisites.

Run a node-by-node migration:

  1. Exclude the node from automation to ensure that resources that must be kept available are moved to another node in the peer domain: 

    samctrl -u a <node>
    Note: The command can run for a considerable amount of time until all move operations are complete.
  2. Stop the node from another node in the domain, and verify that it is stopped: 
    stoprpnode <node>; lsrpnode

  3. To upgrade the node, run the script ./installSAM from the installation directory on the product CD or from the extracted electronic deliverable. For more information about the installSAM script, see Running the installation.

  4. Start the node: 
    startrpnode <node>
  5. Include the upgraded node in automation again: 
    samctrl -u d <node>

  6. The upgraded node can now join the existing domain. Use the lssrc –ls IBM.RecoveryRM command (see the sample in Verifying the active and installed version number) to display the installed version and the active version of the product. The new code features are not activated until the active System Automation for Multiplatforms version number is equal to the highest System Automation for Multiplatforms version number that is installed within the cluster. You cannot fully use these new code features until all the nodes are upgraded.

  7. Repeat the steps 1-6 for other nodes within the cluster.

  8. To activate the new version, continue with Completing the migration.