How to manage rooms, workers etc when using multiple server/node clustering?

We have some nice thorough conversation regarding this here as well:

This picture especially clears much of the stuff.