When I ran the Openstack tempest scenario test, I got failure for the case
tempest.scenario.test_network_basic_ops.TestNetworkBasicOps.test_update_instance_port_admin_state.

This case boots a VM server and then tries to ping the VM instance outside to the VM with floating IP, and then sets port's attribute "admin_state_up" to False, and then check the connectivity again, the expected result is that the VM should be unreachable, but the actual test result is the VM could still be reachable.

I roughly checked codes of OpenDaylight Netvirt project, there are some codes to handle the openstack command to change the "admin_state_up" attribute from True to False and then store the value to MD-SAL
configuration datastore. However, there is no such codes to handle the Data Change to generate the flow rules or configurations for the OVS.

I also noticed that there is no codes to set the Interface's admin_state column in the OVSDB table, due to there is no such attributes for the termination endpoint. Even through I used the command " sudo ovsdb-client -v transact '["Open_vSwitch", {"op" : "update", "table" : "Interface", "where": [["_uuid", "==", ["uuid", "b1d9c977-3ec7-48cc-8072-ae8d888b7ef8"]]], "row":{"admin_state": "down"}}]' " to set the admin_state to down for the port, the OVS still forwards packet for that port [This might be a bug for the OVS, too. But I still think that once we set up "admin_state_up" to False, we should set the "admin_state" to down for the interface related to the neutron port.]