tests/python/TestEntities.py -> def test_runMultipleAdds(self):
We added a timer so that this test won't hang up, but we should find a way to fix this.

More info in #30

Msync increases how long PMGD takes. That means the reader writer lock has more timeouts and so more threads wait. That somehow seems to be messing with some queue that starts denying newer threads from even getting connected.

One of the things we need to figure now at the VDMS layer is that timeout based locking is bound to cause more failures since threads don't get to wait however long. Do we introduce internal retries or let the client retry? We should of course try to figure out the right values as we run more experiments but just saying.