Hello
Apparently, there is no reason but in order to diagnose such a behavior, it would be good to look into the 'autorep -J <box> -d' output to see if some abnormal sequence of events happened or any manual sendevent command had been executed.
Along the 'autorep -J <box> -d' output, we also have to look into the event_demon log to fully understand the sequence of event and the job definitions
Some other general tips:
- look into the job definitions for the max_exit_success attribute
- check if jobs are really in "success" status and not in "terminate" status
- we also fixed some problems in this area, like this one:
BOX ENDS INCORRECTLY WHEN FINAL ACTIVATED JOB GOES ON_ICE
- I would suggest to open a case at CA support and upload all the above documentation
Thanks and regards
Jean Paul