Nova compute containers stays unhealthy after the compute node gets fenced

Issue

• The nova container is waiting forfence flag to set down.

  Waiting for fence-down flag to be cleared
  Waiting for fence-down flag to be cleared
  Waiting for fence-down flag to be cleared
  Waiting for fence-down flag to be cleared
  

• The compute-unfence-trigger resources stay in FAILED state.

   compute-unfence-trigger    (ocf::pacemaker:Dummy): FAILED <COMPUTE_NAME> (blocked)
  

• Since these resources are stuck in FAILED state, stonith cannot be enabled because it would keep rebooting these compute nodes again and again.

• Ideally, after a reboot of compute node the compute-unfence-trigger resource should become healthy.