Running commands in jobs

The exec command can be used to execute an arbitrary command within a running job. Usage is of the form:

prominence exec <job id> <command>

Use cases for this could include checking for the existence of output files, or looking at the content of output or log files.

Current limitations:

• Only standard output is returned.
• The command is not executed inside the job’s container.
• The current working directory of the command is the job’s scratch directory.