Comment 3 for bug 1398402

http://jenkins-product.srt.mirantis.net:8080/job/6.1.system_test.centos.thread_5/99/consoleFull

I reverted prepare nova snapshot and execute steps manually:
1. destroy one controller
2. check pacemaker status as in tests
3. Check that galera is ok
4. run ostf ha (wait while it pass)
5. run ostf sanity
6. run ostf smoke
and in this order cluster succesfully recovered
than I do rever again (btw we have active nodes at this moment :) And would be better to pause them before revert)
and then as in tests I do next actions:
1. destroy one controller
2. check pacemaker status as in tests
3. Check that galera is ok
4. run ostf ha (ha,smoke,sanity) and fail the same as in test, because keystone and memchached need so time for recovery after revert of snapshot
I wait for 3 minutes, run ostf again and all the ostf tests are passed, so we need to run ostf as in first variant