Controllernode connection retries can run too fast

MCSProcessManager._wait_for_controllernode has a retry mechanism if exception is raised, but it doesn't sleep between the attempts, so the controllernode process doesn't get a second chance to start