I'm testing the qsample in Tuxedo 11g. It uses 1 queue space with 3 queues: STRING (for the server app), RPLYQ and ERRQUE. The maximum number of concurrent messages in the qspace is 3. I wanted to confirm that if queue STRING is full, no request will be processed until there is a free slot to enqueue in the RPLYQ.
What's happening is that I'm running the command `client & client & client &` which run 3 instances of the client in parallel. The STRING queue is filled inmediatly and after a while, all processes fail saying that there's no more space in the queue space.
I guess it's trying to enqueue in the RPLYQ before dequeue from the RPLYQ and since it can't do it, it's enqueueing the error in the ERRQUE which still reach the limit of messages and finally all requests fail leaving the ERRQUE full (because the client does not dequeue from the ERRQUE) and forcing me to manually empty it.
Is my guessing right?