Add ability to restart some jobs while others are still running

From @gabrielks (issue 35 in the original repository): "It would be useful to have this feature for long-running workflows containing many jobs that sometimes fail."

We could allow `torc workflows reset-job-status --failed-only` or `torc jobs reset-status` followed by a TBD command `torc workflows rescan`.