Timer based Processes and Exceptions

[bzeuner]

Hi,

I noticed that a process, that has a timer start event (time cycle, cron: */15 * * * * ?) and starts automatically after deploy won't be executed anymore if an exception was thrown within that process.
Only anew deployment will start that cycle again. Does anyone knows that problem?

Furthermore it seems that the process will be executed several times at one cycle. The JOB table only contains one element but the process itself will be executed more than one time. I don't have enough knowledge about the job executor etc. but does anyone have a hint where to look or debug to find out more?

Ben

[frederikherema1]

The job-executor can contain jobs that aren't related to any process-instance, but rather only to a proc-def. So it's normal for a timer-cycle start-event to have only a single job entry.

As we leverage "jobs" for the timer-start event and only a single job exists per process-definition (with a timer-start-event), the job won't run again if the job failed 3 times. So if the retry-count is set to 0, it won't be retried even if the next time-cycle has arrived. This is just the way this is implemented right now…