Htc cheat sheet
List of online workernodes
condor_status -server
To put a node in draining
condor_drain <name_of_workernode>
Add the "-cancel" option to stop the draining.
Summary of the users and their jobs
condor_status -submitters
Get a list of jobs and where they are running
condor_q -run
Note: you must be on the CE to get the workernodes
Get the list of jobs on a given worker
List of jobs on node19-7.wn.iihe.ac.be:
condor_q -constraint 'regexp(".*@node19-7.wn.iihe.ac.be", RemoteHost)'
Get the list of jobs in a given status
If you want to see jobs in run status (JobStatus == 2):
condor_q -constraint "JobStatus == 2" -nobatch
(-nobatch is to split the list when you submit a batch of jobs with queue > 1)
Here is the list of job statuses:
0 | Unexpanded |
1 | Idle |
2 | Running |
3 | Removed |
4 | Completed |
5 | Held |
6 | Submission err |
View details about a job
condor_q -l <jobid>
Get info on a finished job
condor_history [-long] <jobid>
Example:
condor_history -constraint 'Owner == "sgerard" && JobStatus ==4' -limit 10
->you'll get the list, limited to 10, of completed jobs of user sgerard
Why a job is on hold?
condor_q -hold -af HoldReason <jobid>
Why a job stays in idle status?
condor_q -better-analyze <jobid>
condor_q -better-analyze -reverse <jobid>