I presented at the IBM MQ French User Group in Paris on the topic of What's new in MQ. I covered both what was new in IBM MQ 9.3 LTS and what was new in the latest IBM MQ 9.3.1 CD release.
We’ve just talked about Native HA, that’s all about making sure the data (the messages) is as available as possible.
But that’s not everything. You need to consider if it’s enough to have a single queue manager that is very highly available. If you want a system to always be available, think about an active-active, setup with multiple independent queue managers. Each can be highly available in its own right (e.g. with Native HA).
Let’s focus on the active-active setup
Here the objective is that applications treat the set of queue managers as one entity, able to connect without caring which queue manager they are connected to, with each of them able to produce and consume messages efficiently
On z/OS, you’d use a queue sharing group, made possible thanks to the capabilities of the z platform. On distributed, we don’t have that luxury, instead you have the Uniform Cluster…