Commands for managing jobs

Contributors

Jobs are placed into a job queue and run in the background when resources are available. If a job is consuming too many cluster resources, you can stop it or pause it until there is less demand on the cluster. You can also monitor and restart jobs.

When you enter a command that invokes a job, typically, the command informs you that the job has been queued and then returns to the CLI command prompt. However, some commands instead report job progress and do not return to the CLI command prompt until the job has been completed. In these cases, you can press Ctrl-C to move the job to the background.

If you want to…​ Use this command…​

Display information about all jobs

job show

Display information about jobs on a per-node basis

job show bynode

Display information about cluster-affiliated jobs

job show-cluster

Display information about completed jobs

job show-completed

Display information about job history

job history show

Up to 25,000 job records are stored for each node in the cluster. Consequently, attempting to display the full job history could take a long time. To avoid potentially long wait times, you should display jobs by node, storage virtual machine (SVM), or record ID.

Display the list of private jobs

job private show (advanced privilege level)

Display information about completed private jobs

job private show-completed (advanced privilege level)

Display information about the initialization state for job managers

job initstate show (advanced privilege level)

Monitor the progress of a job

job watch-progress

Monitor the progress of a private job

job private watch-progress (advanced privilege level)

Pause a job

job pause

Pause a private job

job private pause (advanced privilege level)

Resume a paused job

job resume

Resume a paused private job

job private resume (advanced privilege level)

Stop a job

job stop

Stop a private job

job private stop (advanced privilege level)

Delete a job

job delete

Delete a private job

job private delete (advanced privilege level)

Disassociate a cluster-affiliated job with an unavailable node that owns it, so that another node can take ownership of that job

job unclaim (advanced privilege level)

Note

You can use the event log show command to determine the outcome of a completed job.

Related information