When the tempest scenario test case was run (test_network_basic_ops.py), couple of failures were seen mostly around Floating IP's
We have a end to end setup with l2 and l3 pings working.We have VM's created associated to Floating IP's.

BUild used -
Kilo and Liberty
ODL - distribution-karaf-0.4.0-20151020.153447-1732

Initial configs,
neutron net-create test-net
neutron subnet-create test-net 10.10.10.0/24 --name testsubnet
neutron router-create testrouter
neutron router-interface-add testrouter testsubnet
neutron net-create ext_network – --router:external=True
neutron subnet-create --allocation-pool start=192.34.0.20,end=192.34.0.30 ext_network 192.34.0.0/24 – --enable_dhcp=False --name ext_subnet
neutron router-gateway-set "$rid" "$extid"
nova boot --flavor 1 --image "$image_id" --nic net-id=$netw test_vm
neutron floatingip-create ext_network
neutron floatingip-associate "$floating_ip_id" "$vm_port_id"

After the above configs, the floating IP is associated to the VM and accessible via ssh

Failure 1: test_update_instance_port_admin_state
Initially the floating IP is reachable and we can ssh to it.
Now we edit the admin_state attribute of instance port to "DOWN". It is expected that the floating IP should no longer be reachable.
However, in our case it is still reachable.

Failure 2: test_update_router_admin_state
Similar to failure 1..In this case the admin_state attribute of the router is set to "DOWN"..
Even after this, the Floating IP is reachable.

Failure 3: test_network_basic_ops
Once you disassociate a floating-ip from the VM, the test case verifies that the status of the floating IP shows "DOWN".
In our case it shows as "ACTIVE". Not sure if this is the expected behavior and the test case needs some change.
GET http://105.0.1.216:9696/v2.0/floatingips
{
"floatingip" :

{ "id" : "a058801b-2c98-4d4d-b90a-505ae2d295e6", "tenant_id" : "036bcfd5b251416ca35fe2cfa730f433", "floating_network_id" : "7c1a4760-3c64-43b2-a9d4-f604b95a3bf1", "floating_ip_address" : "192.34.0.21", "status" : "ACTIVE" <==== }

}