Test that setting the maintenance mode for a offline node (node3) manually fails, then re-power it
(node3) and enable mainteancne mode for an online node (node2). Now trigger a manual migrate of the
for maintenance moved service. Disable maitenance mode again (node2) and currently the HA stack will
migrate the service that got moved twice, once automatically for maintenance and once manually, back
due to the maitenance fail-back logic.

If that is the right behavior, or if a manual migration should clear any maintenance fallback might
be up for debate, this test just encodes the actual current behavior to ensure its not changed by
mistake.
