Migrating HA RDQMs

Follow this sequence of steps to upgrade all the RDQM nodes in an HA group and so migrate the replicated data queue managers (RDQMs).

About this task

We should upgrade all the nodes in an HA group in the same sequence to avoid operating with the nodes in the group at different levels.

If we have configured your HA group such that one node acts as a primary for all RDQMs, with the other two nodes as secondaries, we should upgrade the secondary nodes first and leave the primary node until last.

The sequence in which you upgrade, and the nodes that are marked as preferred and second preferred locations for RDQMs, affect where the RDQMs fail over to as you upgrade. During the migration sequence, while nodes are running different levels, the options for failing over are limited. An RDQM running on a lower level node can fail over to a higher level node but, once a queue manager has been started at the new level, it cannot fail over to a lower level node. We should choose an upgrade sequence and use preferred and second preferred locations settings to keep queue managers running on the lower level nodes for as long as possible. We should make changes to preferred and second preferred location settings before you suspend nodes, to ensure that the changes are effective immediately.

If we are also running DR RDQMs on any of the nodes, we should deal with these queue managers at the same time by following the instructions in Migrating DR RDQMs.

Procedure

• Uninstall HA RDQM support and upgrade RDQM and IBM MQ .
  1. Suspend the HA group on the node, by entering the following command:

     rdqmadm -s

  2. Log in as root or switch to superuser using the su command.
  3. Uninstall IBM MQ (this step also uninstalls RDQM):

     rpm -qa | grep MQSeries | xargs yum -y remove

  4. Uninstall Pacemaker:

     rpm -qa | grep linbit | xargs yum -y remove

  5. Uninstall DRBD:

     rpm -qa | grep drbd | xargs yum -y remove

  6. Install the new level of IBM MQ and dependent software, see Installing RDQM (replicated data queue managers).
  7. Resume the HA group on the node by entering the following command:

     rdqmadm -r

     We can now proceed to the next node in the group.

• Repeat the steps for the second node in the HA group. (Use the same path as you did on the first node.)
• Repeat the steps for the third node in the HA group. (Use the same path as you did on the first node.)

Parent topic: Migrating replicated data queue managers

Related information

• rdqmadm (administer replicated data queue manager cluster)