VMM Service seems to be very sensitive to all kinds of issues, including very small ones (network glitches, IP from non-existing pool specified for Bare Metal Deployment, etc.) and whenever something like that occurs it just crashes interrupting everything that was running before and not resuming/restarting it afterwards (i.e. cross cluster VM migration, Bare Metal Deployment, VM deployment, etc.). It would be nice if it could handle those issues in a more reliable fashion and in cases when in really needs to crash, resume or restart corresponding activities.