I am running kolla-ansible(pip) with pike, on centos 7, all updated, in a multinode deploy with 2 nodes for ctrl, neutron, storage and compute. All nodes have four IP-cfgd interfaces eth0-3, and neutron nodes have an UN-numbered interface for external-net traffic.

globals.yml contains:

enable_neutron_provider_networks: "yes"
network_interface: "eth1"
kolla_external_vip_interface: "eth0"
api_interface: "eth1"
storage_interface: "eth3"
tunnel_interface: "eth2"
neutron_external_interface: "eth4"

I can create VMs, networks, routers, FIPs, volumes, and all look quite nice, inkluding the provider network for external acccess.

# oo network show PubInet
+---------------------------+--------------------------------------+
| Field                     | Value                                |
+---------------------------+--------------------------------------+
| admin_state_up            | UP                                   |
| availability_zone_hints   |                                      |
| availability_zones        | nova                                 |
| created_at                | 2018-04-11T19:49:51Z                 |
| description               | Public Internet Access               |
| dns_domain                | None                                 |
| id                        | cfb385b8-1908-4fcc-8944-81792481012f |
| ipv4_address_scope        | None                                 |
| ipv6_address_scope        | None                                 |
| is_default                | False                                |
| is_vlan_transparent       | None                                 |
| mtu                       | 1500                                 |
| name                      | PubInet                              |
| port_security_enabled     | True                                 |
| project_id                | e410ae95250046ea8fa123e0231a0b01     |
| provider:network_type     | flat                                 |
| provider:physical_network | physnet1                             |
| provider:segmentation_id  | None                                 |
| qos_policy_id             | None                                 |
| revision_number           | 4                                    |
| router:external           | External                             |
| segments                  | None                                 |
| shared                    | False                                |
| status                    | ACTIVE                               |
| subnets                   | 2d8c854b-5b77-437e-9b9d-33c9f86d2ebc |
| tags                      |                                      |
| updated_at                | 2018-04-11T19:49:55Z                 |
+---------------------------+--------------------------------------+

BUT, the configuration for ovs, configures the routers global/external interfaces to the wrong bridge. The br-ex bridge is properly configured with the chosen eth4 as can be seen:

docker exec -it openvswitch_vswitchd bash
(openvswitch-vswitchd)[root@neutron1 /]# ovs-vsctl show
34ccd256-f498-4c06-aa06-7fa22441c5a9
    Manager "ptcp:6640:127.0.0.1"
        is_connected: true
    Bridge br-ex
        Controller "tcp:127.0.0.1:6633"
            is_connected: true
        fail_mode: secure
        Port phy-br-ex
            Interface phy-br-ex
                type: patch
                options: {peer=int-br-ex}
        Port br-ex
            Interface br-ex
                type: internal
        Port "eth4"
            Interface "eth4"
    Bridge br-int
        Controller "tcp:127.0.0.1:6633"
            is_connected: true
        fail_mode: secure
        Port br-int
            Interface br-int
                type: internal
        Port "qr-28133bb7-98"
            tag: 6
            Interface "qr-28133bb7-98"
                type: internal
        Port int-br-ex
            Interface int-br-ex
                type: patch
                options: {peer=phy-br-ex}
        Port "qg-5debe8d7-06"
            tag: 7
            Interface "qg-5debe8d7-06"
                type: internal
        Port "tap624dda98-5d"
            tag: 6
            Interface "tap624dda98-5d"
                type: internal
        Port patch-tun
            Interface patch-tun
                type: patch
                options: {peer=patch-int}
    Bridge br-tun
        Controller "tcp:127.0.0.1:6633"
            is_connected: true
        fail_mode: secure
        Port patch-int
            Interface patch-int
                type: patch
                options: {peer=patch-tun}
        Port "vxlan-ac146566"
            Interface "vxlan-ac146566"
                type: vxlan
                options: {df_default="true", in_key=flow, local_ip="172.20.101.104", out_key=flow, remote_ip="172.20.101.102"}
        Port br-tun
            Interface br-tun
                type: internal
(openvswitch-vswitchd)[root@neutron1 /]# exit

The l3-agents qg-interface should have been added to the br-ex bridge, since that is were the flat provider network is accessible.

Somewhat out-of-normal, the OVS configuration hasnt got an "openvswitch_agent.ini", but the same configuration (the OVS chapter) is configured and supplied in the neutron_openvswitch_agent container via /etc/kolla/neutron-openvswitch-agent/ml2_conf.ini, where it says:

[ovs]
bridge_mappings = physnet1:br-ex
datapath_type = system
ovsdb_connection = tcp:127.0.0.1:6640
local_ip = 172.20.101.104

So, why is the qg-interface not added correctly? What is missing? My memory tells me that I ran into a similar issue in my RDO packstack attempts, but in those scenarios the OVS-bridge contained additional virtual links, that weirdly got the traffic through anyway. I THINK I solved it then with the "external_network_bridge" parameter to neutron.conf, but :

 external_network_bridge¶
    Type:   string
    Default:    

    Name of bridge used for external network traffic. When this parameter is set, the L3 agent will plug an interface directly ...