[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Condor-users] condor_rm on lots of jobs



On Thu, 29 Apr 2010, Dan Bradley wrote:

The DAGMan experts may correct me, but I believe the best practice for removing DAGs is to condor_rm the DAGMan scheduler universe job and let DAGMan condor_rm the jobs in the DAG. Example:

condor_rm -constraint 'JobUniverse == 7'

Yes, the best way to remove a DAG is to just condor_rm the DAGMan job itself. DAGMan will then remove all of the corresponding node jobs.

However, that could still result in a lot of POST scripts running at the same time. When you submit a dag, there is an option -maxpost which can be used to specify the maximum number of POST scripts that will be run at the same time in that DAG.

Actually, this should not result in any POST scripts being run; in fact, any running POST scripts should be killed. Years ago someone claimed that they ran into a case where many POST scripts ran when removing a DAG, but we were never able to duplicate the problem. If you do condor_rm the DAG, and see POST scripts run at that point, we'd be very interested to see the dagman.out file...

Kent Wenger
Condor Team