-
Dear all, I've noticed that some jobs fail with MinorStatus='Received Kill signal' however site admins find that the job is successfull from the batch system point of view. This behavior happens for a small fraction of identical jobs, where the successfull ones complete in several hours, while the failed ones get killed after a few seconds.
Does any one have an explanation of this behavior? |
Beta Was this translation helpful? Give feedback.
Replies: 1 comment
-
The log messages suggest that the job was killed by the DIRAC user, e.g. with dirac-wms-job-kill or Kill button in the Web portal. The kill instruction was sent back to the job via the heartbeat mechanism. Please, check that this is not part of your production management actions or may be a human (faulty) intervention. |
Beta Was this translation helpful? Give feedback.
The log messages suggest that the job was killed by the DIRAC user, e.g. with dirac-wms-job-kill or Kill button in the Web portal. The kill instruction was sent back to the job via the heartbeat mechanism. Please, check that this is not part of your production management actions or may be a human (faulty) intervention.