[QA] The filling_root test shall recover from the red flag on its own

[Bogdan Dobrelya]

    Scenario:
        0. Setup master node
        1. Configuration releases
        2. Bootstrap slave and make snapshot
        3. Revert bootstrapped nodes
        4. Create Fuel Environment
        5. Add nodes to environment
        6. Run network checker
        7. Deploy environment
        8. Run network checker
        9. Run health checker
        10. Load existent environment from snapshot or save it
        11. Get controllers initial status in pacemaker
        12. Filling root filesystem on primary controller
        13. Check stopping pacemaker resources
        14. Run health checker
        15. Fill root more to below rabbit disk free limit
        16. Check stopping pacemaker resources
        17. Run health checker
        18. Clean up space on root filesystem on primary controller
        19. Check starting pacemaker resources
        20. Run health checker Sanity, Smoke and HA

Expected:
After the step 18, resources will be recovered to the healed node automatically
Actual:
By design of a Pacemaker <1.1.15, the red flag shall be lifted off by an operator (or the test case). It will not reset a bad state automatically, see related docs bug https://bugs.launchpad.net/fuel/+bug/1500422 and its duplicates as well

Please fix the test's expectations to meet the actual behavior