slicer crashing port issues

[jsnoble]

As the code stands now, when the slicer crashes, it comes back up and then sends a message to cluster master to shutdown everything which is the correct behavior, but it leads to slicer exiting twice. On the node master it tries to reclaim the port every time a slicer exits. Since it exits twice it sends the same port twice back into the port queue which can lead to port already in use errors in the future