`rocker rerun` should probably use a more gentle way to stop and delete containers

[mreithub]

Currently all rocker does is to force-delete the container.
This will kill the processes running in that container which in most cases won't cause trouble (as all the container data is thrown away anyway).

But persistent data (data volume containers/host-based volumes) might be in an irreparable state when the processes get killed.

To avoid that it's most likely better to issue a docker stop call before removing a container.