Controller host vbmc (virtualbmc) went down and generate cluster instablity

Issue

• Sometimes virtualbmc instances are going down:

[root@hostname01 contrail]# vbmc list
+-----------------------------+---------+---------+------+
|         Domain name         |  Status | Address | Port |
+-----------------------------+---------+---------+------+
|   domain01                  | running |    ::   | 643  |
|   domain02                  | running |    ::   | 641  |
|   overcloud-controller-0    | running |    ::   | 650  |
|   overcloud-controller-1    | running |    ::   | 642  |
|   overcloud-controller-2    | Down    |    ::   | 630  |
+-----------------------------+---------+---------+------+

• When the virtualbmc instance goes down, pacemaker notices that state and can no longer use those fencing devices:

Failed Actions: 
* stonith-fence_ipmilan-525400d4fbbb_start_0 on overcloud-controller-1 'unknown error' (1): call=147, status=Timed Out, exitreason='', last-rc-change='Fri Dec 14 14:48:56 2018', queued=0ms, exec=20042ms 
* stonith-fence_ipmilan-525400d4fbbb_start_0 on overcloud-controller-2 'unknown error' (1): call=187, status=Timed Out, exitreason='', last-rc-change='Mon Dec 24 12:14:58 2018', queued=0ms, exec=20006ms 
* galera_monitor_10000 on galera-bundle-1 'unknown error' (1): call=15972, status=complete, exitreason='local node <overcloud-controller-1> is started, but not in primary mode. Unknown state.', last-rc-change='Mon Dec 24 11:45:12 2018', queued=0ms, exec=0ms