Oct 9th, 2012, 07:07 AM
Reusing channel in remote-chunking
I have a question. We are currently evaluating Spring Batch.
I am trying to figure out whether it is possible to use same queue (channel) when remote-chunking if different job instances are running in parallel (i.e. same job but one running with data from customer X and one running with data from customer Y)?
Also is it possible to use same queue for different jobs?
Last edited by blondak; Oct 9th, 2012 at 08:21 AM.
Oct 10th, 2012, 10:02 AM
I'll admit I've never tried this but I don't think it would work. While on the input side, I think things would be fine, the question becomes the return side. How does the listener know who to return the output to (which instance)?
Oct 11th, 2012, 10:54 AM
Thanks for reply.
I am bit suprised because I would say that running same job with different data in parallel is not unreasonable requirement. Actually I think if first one does not work that the second one would not work as well. You are right that there is no problem with the request queue but with the reply queue. I think it would be possible if the messages would be marked with correlation id = job instance id. Then consumers could filter messages using selectors.
But it looks like they are currently consuming all the messages without any filtering.
Oct 11th, 2012, 10:58 AM
For the record, there may be something within Spring Integration that could be used to help with this scenario. You may want to post into that forum as well for some ideas.