Certainly it is a good way to go, but in my experience try the usual work flow, it is way more stable and the cpu burden is share between both nodes -> faster backups