Disable the openstack-slothd collector; fix #6.

For whatever reason, this seems to be the trigger for issue #6 .  We
certainly are running with a ton of VMs, so the Ceilometer history that
the collector grabs for Cloudlab is very detailed... it would not
surprise me if the collector runs around the clock in these experiments.

I no longer think this is really our problem... but I'll leave that bug
open in case I ever have time to investigate.  But this certainly
explains the nondeterministic locks I was seeing in Neutron.
1 job for master in 2 seconds (queued for 2 seconds)
Status Job ID Name Coverage
passed #751