We have a process that has a timer going for 30 seconds in a branched flow. If the primary branch completes first it hits a terminate node that's expected to cancel the timer midway through, and that's exactly what it does in every instance, unless it's on our production environment.
On our production environment only, we're seeing the terminate node fire, then the timer finish a little under 30 seconds later, script tasks after it also complete without any problems, and finally and end node not set to terminate. Since that's the only other one in the process model, it completes the process again and we see two outbound flows in the parent process.
I see other issues with sub-processes behaving strangely when the parent process is terminated. This is not what's happening. The terminate node and the timer are on the SAME process model. What in the world could keep the timer active after the process it's in has been terminated?
Discussion posts and replies are publicly visible
Going through Process Nodes I can verify based on completion times that the Timer started first, then Terminate Process node ran, then the timer finished. Then those things which follow the timer all took place, and the last entry is the regular end event about 35 seconds after the Terminate Process.
To my knowledge this is only happening on our production environment.