Skip to Content
Former Member
Apr 06, 2011 at 11:00 AM

correlationid in ccBPM in stopped bpm's



We use PI ccBPM to coordinate between several async services.

Each service is called async and a deadline (timeout) is opened waiting for it to callback async. The receiver is configured with a correlation id so that the correct bpm context is restored when the call returns. If the deadline is meet before the callback receiver has been called, we want to report an error, and stop the bpm process.

The process woks fine when there are no timeouts. However, we found that if a process times out, and we exit the bpm instance, but later a callback arrives with a correlation id for that closed process, the whole bpm queue (for ALL existing and future instances) is stuck, until queue is handled manually.

What can be done to avoid this result?