enable to segment

It has serveral compute nodes in our cloud env, and they are set in the
same segment. We usually pull the latest code or change hardware for test.
Every time when updating, we have to kill all masakari services after
masakari-monitors, so that it wouldn't trigger ha recovery.

Once upon a time, we forgot to kill all masakari services when updating,
and terriable thing happened. It triggered VMs ha, but many instances
evacuated failed. When the compute nodes re-active, the instance can't
be handly recovery, bacause bdm info of the instances was incorrect while
evcuations failed.

So we need a solution to simply and easily close and open ha in specific
case.