Docker: attempt clean stale pids on restart
Issue #4235 (closed) points to a problem with being unable to properly restart the omnibus containers. One of the cases is in the even of server outage (or unclean stop).
This change ads the functionality of cleaning up PIDs & socket files prior to the start of runsvdir.
This does not address ...: runsv not running
complaints directly.
Closes #4235 (closed)