RDQM DR and HA Combined

By JOHN COLGRAVE posted Wed May 20, 2020 04:39 AM

IBM MQ Advanced 9.1.5 includes fully integrated support for RDQM DR and HA in MQ. It is now possible to set up two different RDQM HA Groups and have a single queue manager configured in both with asynchronous data replication between the HA Groups.

At the moment it is necessary to create a new queue manager to make use of this facility,but you can use the process to migrate an existing queue manager to RDQM with a new RDQM queue manager that is configured for both DR and HA.

When you create a queue manager for both DR and HA you supply both HA-related information and DR-related information on the crtmqm command. You create the queue manager on the node that you want to be the HA preferred location in the primary site. As with the initial RDQM DR support, part of the output of the initial crtmqm command is a slightly different crtmqm command that is to be run on a node in the secondary site. The key difference with this new support is that on the secondary site the queue manager is also configured for HA and the node where the secondary crtmqm command is run becomes the HA preferred location for the queue manager in the secondary site.

The process for doing a DR takeover is the same as with the initial DR support. To test moving the active instance of the queue manager to the secondary site, you make the current DR Primary the DR Secondary, then make the original DR Secondary the DR Primary. There is no need to start the queue manager on the new primary site, because the queue manager is already configured for HA, the HA support in the new primary site will automatically start the queue manager and the data should already be in sync on the three nodes of the new primary site.

If you would like to see a full example of setting up a deployment combining DR and HA there is an example for the Azure public cloud at