I would like to use HornetQ a certain way, and I would appreciate some input on whether anyone else has done this before, whether its a good idea, and, if so, how to best make it work.
We have a large number, about 60, "input -> process -> output" nodes that communicate using HornetQ. What I'm thinking of doing is this:
- Setup a "mirror" of our production broker, probably using a core bridge, where each incoming message to one of these nodes is copied.
- Use a transformer to set the time-to-live on the copied messages to something relatively small, like a few hours.
- Test a given node (or subgraph) in our staging environment by tapping into the mirror at the appropriate place.
It looks like we will need a <bridge/> entry in our hornetq-configuration.xml file for each queue. Is there a way to do this with less configuration overhead? I'm also curious to know if someone thinks this is a bad idea from a performance point-of-view. It seems like this would be similar to the way a cluster works, so I don't expect it to be a problem. But it would be great to hear from someone who knows more about HornetQ than I do.
Thanks,
Sam Bishop
