Comment 1 for bug 1708613

One possible reason is there are multiple neutron-zvmagent run on the zVM, the guest is revoked by one of the agent, but this revoke operation didn't update switch tale ( switch table is deleted when the whole guest is destroyed), so there is a time gap between the revoke and the guest deploy. During this time, it is possible that another neutron_zvm_agent start and check the vswitch table, to regrant the guest. It will cause the guest is deleted but the still existed in the vswitch's authorized list.