If you've allowed your Server Person to be a single point of failure out innawoods, that's an organizational problem, not a technological one.

Two is one and one is none.