Ceph

Sage suggested we might want to have teuthology archive the monitor's data directory, since that information is often useful. We should evaluate how much space that takes up at the end of our teuthology runs (since disk space is a frequent problem with teuthology jobs) and whether or not it's feasible to do so.

If it is, implement a system that either
1) archives all of the monitor data dirs, or
2) archives the directory of the leader, plus anybody who's crashed.