condor_now

Start a job now by vacating a currently running job.

Synopsis

condor_now [-help]

condor_now [-name scheddname] [-pool hostname[:portnumber]] [-debug] now-job vacate-job [vacate-job …]

Description

condor_now tries to run the now-job immediately. 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.

If you specify multiple vacate-job s, each will be immediately vacated; after they all terminate, the schedd will try to coalesce their slots into a single, larger, slot and then use that slot to run the now-job.

You must specify each job using both the cluster and proc IDs (e.g., 123.4).

Options

-help

Display usage information.

-debug

Print debugging output. Control the verbosity with the environment variable _CONDOR_TOOL_DEBUG, as usual.

-name scheddname

Specify the scheduler by name.

-pool hostname[:portnumber]

Specify the pool to find the schedd in.

now-job

The job ID (cluster.proc) of the job to start immediately.

vacate-job

The job ID (cluster.proc) of one or more jobs to vacate.

General Remarks

The now-job and the vacate-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 vacate-job must be running.

condor_now does not guarantee that the now-job will run immediately, as other factors (such as RANK expressions or other jobs from the same user) may affect scheduling decisions.

Examples

Start job 17.3 using job 4.2’s slot:

$ condor_now 17.3 4.2

Debug why a job won’t start on the ‘magic’ scheduler in the ‘gandalf’ pool:

$ export _CONDOR_TOOL_DEBUG=D_FULLDEBUG
$ condor_now -debug -name magic -pool gandalf 17.3 4.2

Start job 100.0 by coalescing slots from jobs 10.0 and 11.0:

$ condor_now 100.0 10.0 11.0

Exit Status

0 - Success (schedd accepted the request to vacate and start the job)

1 - Failure

Note

condor_now returns immediately after the schedd accepts the request. It does not wait for the now-job to actually start running.

See Also

condor_vacate_job, condor_q, condor_prio

Availability

Linux, MacOS, Windows