Article ID: 559 - Last Modified: August 10, 2015
How do I stop a multi-processor docking job so that I can use the RESTART mechanism at a point in the future?
Currently, the -RESTART mechanism applies only to distributed jobs, and the restarting is done at the level of subjobs only. That is, if several subjobs have completed, they won't be rerun, but any incomplete subjobs will be started again from the beginning. Resuming an individual subjob (or a serial Glide job) is not yet supported. If your large docking job is split into many subjobs (i.e., many more subjobs than processors used), restarting won't have to redock very many ligands (in those incomplete subjobs).
In this situation, you can just kill the main job from the Monitor panel in Maestro, or from the command line with
$SCHRODINGER/jobcontrol -kill JobId
The job can be restarted by running from the command line
$SCHRODINGER/glide -RESTART jobname.inp
If you are running large subjobs (as might be the case if you are docking a large database with the same number of subjobs as processors), all of your subjobs might be incomplete, causing the -RESTART mechanism to essentially rerun the entire job from the beginning. In this situation, the only way to stop the job and continue later would be to kill the current job, run new jobs to dock the ligands not yet docked, and then combine all the results together at the end with the 'glide_sort' and 'glide_merge' utilities.
Keywords: Glide, subjobs, termination, retrieve, lost, -RESTART mechanism, launch directory, kill subjob
Type the words or phrases on which you would like to search, or click here to view a list of all
Knowledge Base articles