After revert the snapshot 'prepare_ha_nova', pacemaker got a split brain

[Dennis Dmitriev]

Reproduced on CI: http://jenkins-product.srt.mirantis.net:8080/view/6.1_swarm/job/6.1.system_test.ubuntu.thread_5/160/ ; snapshot prepared in test 'prepare_ha_nova' , issue appeared in test 'ha_nova_destroy_controllers'

        Scenario:
            1. Create cluster (Ubuntu/IBP , Nova-network/flatDHCP)
            2. Add 3 nodes with controller roles
            3. Add 2 nodes with compute roles
            4. Deploy the cluster
            5. Make snapshot
            6. Revert the snapshot
            7. Monitor pacemaker status on primary controller for the next 10-20 minutes.

Expected results: After reverting, pacemaker status is not changed.

Actual result: On primary controller (node-1 in the attached snapshot), 1-2 minutes after reverting the snapshot, pacemaker lost connection to the other controllers and stopped all resources on the primary controller: http://paste.openstack.org/show/304670/

That failed the test "Destroy two controllers and check pacemaker status is correct" with error http://paste.openstack.org/show/304679/, when pacemaker was re-assembling rabbitmq cluster.

5 minutes later, pacemaker on the node-1 came back: http://paste.openstack.org/show/304671/

Can be related to https://bugs.launchpad.net/fuel/+bug/1458828 .

[Tatyanka (tatyana-leontovich)]

move to 9.0 according to we do not fix medium issues in 8.0 anymore

[Tatyanka (tatyana-leontovich)]

actually move to invalid according to we do not have cases for nova anymmore