Compute service job dependencies

Looking to run a job on the RCS compute service, but only after another job has completed? The submission system lets you easily specify job dependencies.

Just take note of your first job’s ID:

$ qsub first_job.sh
16394.pbs

And provide it when invoking your second job:

qsub -W depend=afterok:16394 second_job.sh

This also works with array jobs and can take job failure into account, among other features.

Further resources