Have one quick and basic question.
Do we need to really create set of Proxy Q's (req/rsp) prior in respective Q manager for MQ virtualization?
Or do we have an alternate where proxy Q's can be created in Devtest on the fly while creating MQ virtualization?
As usual, there is no such thing as a quick and basic question with messaging.
In 10.3 there are ways to do this with IBM MQ "dynamic queues", which are simply temporary queues that are given a static name with no wildcard "*" characters in them. However, you still need to reconfigure your application client, and now the queues you need to reconfigure it for do not exist unless the VSE Recorder or VS is currently running. You can also use an actual temporary queue, with a temporary name, on the response side, but that has to be something that's supported by your application client.
I should add that you have some leeway with proxy queues. They do not have to be on the same Queue Manager or even the same MQ server as your live queues. As long as the proxy queues are accessible to both your application client and VSE, they can be wherever you want.
However, under certain conditions you don't need to create any proxy queues:
- You are generating your VS using RR pairs instead of recording a live service.
- You don't need the resulting VS to support the various live invocation modes.
- You can shut down the live service and any other services using those same live queues when the VS is running.
"Proxy Queues" are just the queues that your application client uses to talk to the VS. If the above conditions are met, then your existing "Live Queues" can serve both roles.
Thank you, Kevin.