condor_now — HTCondor Manual
Start a job now.
condor_now [-name] [-debug] now-job vacate-job
condor_now tries to run the now-job now. The vacate-job is immediately vacated; after it terminates, if the schedd still has the claim to the vacated job's slot - and it usually will - the schedd will immediately start the now-job on that slot.
You must specify each job using both the cluster and proc IDs.
Print a usage reminder.
Print debugging output. Control the verbosity with the environment variables _CONDOR_TOOL_DEBUG, as usual.
- -name **
Specify the scheduler('s name) and (optionally) the pool to find it in.
The now-job and the vacated-job must have the same owner; if you are not the queue super-user, you must own both jobs. The jobs must be on the same schedd, and both jobs must be in the vanilla universe. The now-job must be idle and the vacated-job must be running.
To begin running job 17.3 as soon as possible using job 4.2's slot:
condor_now 17.3 4.2
To try to figure out why that doesn't work for the 'magic' scheduler in the 'gandalf' pool, set the environment variable _CONDOR_TOOL_DEBUG to 'D_FULLDEBUG' and then:
condor_now -debug -schedd magic -pool gandalf 17.3 4.2
condor_now will exit with a status value of 0 (zero) if the schedd accepts its request to vacate the vacate-job and start the now-job in its place. It does not wait for the now-job to have started running.
1990-2021, Center for High Throughput Computing, Computer Sciences Department, University of Wisconsin-Madison, Madison, WI, US. Licensed under the Apache License, Version 2.0.