Multi datacentre DR node for safety Can't do multi datacentre durable write safely since only one node in distant datacentre