Article ID: 1445 - Last Modified: October 25, 2013
I have a job that is stranded. What does that mean? How do I get rid of it?
Stranded jobs are those with which Job Control has lost contact. This may be due to a prolonged network outage, or a reboot of the machine where the job is running. The job might have failed, or it might be still running. Since the actual status of the job cannot be determined, you should first try to recover the job:
$SCHRODINGER/jobcontrol -recover JobId
where JobId is the actual Schrodinger JobId for that job. If Job Control makes connection with the job and determines that the job is still running, the status changes to
running; otherwise, it changes to
completed:finished, depending on what Job Control can determine about the job. If Job Control still cannot connect to the job, the status remains at
Stranded jobs can be killed by Job Control, which means that Job Control abandons the job, and will not attempt to recover it (even if you discover later that it is still actually running). Its status is set to 'killed'. No cleanup can be done a job that is killed while stranded, because it cannot be contacted, so its files remain on the execution host (unless it is cleaned up by the host's cleanup mechanisms).
Keywords: stranded, recover
Type the words or phrases on which you would like to search, or click here to view a list of all
Knowledge Base articles