Test how service recovery works with the 'static' resource scheduling mode.

Expect that the single service always gets recovered to the node with the most
available resources. Also tests that the group priority still takes precedence.
