Redeployment and timer start events

Hi, I couldn’t find a description of the intended behavior when re-deploying a process definition that has a timer start event. If you re-deploy a process definition, does the job that was already scheduled for the original definition get cancelled and only the new one with a fresh calculation of duedate get scheduled? Or do both remain active, since both process definitions are still active, even though by default only the newest version would be started?

Hi,

If I remember correctly, the currently scheduled start job is replaced by the new job on re-deployment. I.e. there are never two active timer start jobs at a time in this scenario.

Cheers,
Thorben

1 Like

Thanks Thorben, that seems to make sense too. In one case I saw, there was a repeating timer start event with a very short interval, combined with programmatic redeployment on startup. This lead to some strange behavior where there were two jobs because deleting the original job failed due to it being picked up too quickly by the JobExecutor and there being a conflict. For more “normal” use cases, I’d expect this conflict to be very unlikely and the original job to be deleted.