SLURM Job Dependencies
You may submit jobs that runs depending on status of the previously submitted jobs or schedule a bunch of jobs to run one after the other.
Once you submit a job, using that job ID, you can submit dependency jobs.
You need to extract the job id “12345” from the output of the “sbatch” command
$ sbatch job.cmd
Submitted batch job 12345
By adding the “–parsable” option to “sbatch command”, only the job ID would be returned and its value can be stored in a shell variable for later use.
$ jobID=$(sbatch --parsable job.cmd)
$ echo ${jobID}
12345
Next, you can submit a job that only runs after successful completion of the first job as follows where we set the “afterok” as the dependency type.
$ sbatch --dependency=afterok:${jobID} second_job.cmd
The format here is
$ sbatch --dependency=type:job_id jobfile
If the job requires more than one job to be completed before it is executed, you can supply all the jobids using , separator
$ sbatch --dependency=type:job_id,job_id,job_id jobfile
You can also set the job to run if any one of the job ids completes successfully using a ? separator
$ sbatch --dependency=type:job_id?job_id?job_id jobfile
The other dependencies that can be used for<type:job_id> are as follows:
| Type | Description |
|---|---|
| after | This job can begin execution after the specified jobs have begun execution |
| afterany | This job can begin execution after the specified jobs have terminated. |
| aftercorr | A task of this job array can begin execution after the corresponding task ID in the specified job has completed successfully |
| afternotok | This job can begin execution after the specified jobs have terminated in some failed state |
| afterok | This job can begin execution after the specified jobs have successfully executed |
| singleton | This job can begin execution after any previously launched jobs sharing the same job name and user have terminated |