Ask Your Question

vxlan peers not being created (mitaka openvswitch agent native driver)

asked 2016-05-26 22:53:21 -0600

don gravatar image

updated 2019-01-07 15:10:00 -0600

My vxlan peers are not coming up (not being created). Below is the config of a compute node. There are no errors, it just doesn't create the vxlan tunnel interface. Can you suggest where to look?

(This is on Ubunto 16.04 with Mitaka)

If i look in the database:

mysql> select * from ml2_vxlan_endpoints;
| ip_address    | udp_port | host    |
| |     4789 | nubo-19 |
| |     4789 | nubo-20 |
| |     4789 | nubo-21 |
| |     4789 | nubo-23 |
|  |     4789 | nubo-5  |
|  |     4789 | nubo-6  |
|  |     4789 | nubo-8  |
|  |     4789 | nubo-9  |
|  |     4789 | nubo-16 |
| |     4789 | nubo-4  |
10 rows in set (0.00 sec)

the node w/ the config below is nubo-23 (they are all the same).

I don't see any errors in the logs. Not a 1.

i'm expecting to see an entry in br-tun like:

Bridge br-tun Port "vxlan-0a640a97" Interface "vxlan-0a640a97" type: vxlan options: {df_default="true", in_key=flow, local_ip="xx.xx.xx.xx", out_key=flow, remote_ip=""}

but i have no entries.

This is on Ubuntu 16.04.

mysql> select * from ml2_vxlan_allocations where allocated = 1;
| vxlan_vni | allocated |
|        63 |         1 |
1 row in set (0.00 sec)


type_drivers = vxlan
tenant_network_types = vxlan
mechanism_drivers = openvswitch

flat_networks = *

vni_ranges = 1:32000
#vxlan_group =

enable_security_group = false


arp_responder = True
l2_population = True
minimize_polling = True
ovsdb_monitor_respawn_interval = 30
polling_interval = 5
prevent_arp_spoofing = False
tunnel_types = vxlan

bridge_mappings = physnet1:br-ex
local_ip =
of_interface = native
ovsdb_connection = tcp:
ovsdb_interface = native
tenant_network_type = vxlan
tunnel_id_ranges = 1:32000
tunnel_type = vxlan

firewall_driver = neutron.agent.firewall.NoopFirewallDriver

enable_security_group = False

ovs-vsctl output::

    Manager "ptcp:6640:"
        is_connected: true
    Bridge br-int
        Controller "tcp:"
            is_connected: true
        fail_mode: secure
        Port br-int
            Interface br-int
                type: internal
        Port int-br-ex
            Interface int-br-ex
                type: patch
                options: {peer=phy-br-ex}
        Port patch-tun
            Interface patch-tun
                type: patch
                options: {peer=patch-int}
    Bridge br-ex
        Controller "tcp:"
            is_connected: true
        Port phy-br-ex
            Interface phy-br-ex
                type: patch
                options: {peer=int-br-ex}
        Port "enp3s0f1"
            Interface "enp3s0f1"
        Port br-ex
            Interface br-ex
                type: internal
    Bridge br-tun
        Controller "tcp:"
            is_connected: true
        fail_mode: secure
        Port br-tun
            Interface br-tun
                type: internal
        Port patch-int
            Interface patch-int
                type: patch
                options: {peer=patch-tun}
    ovs_version: "2.5.0"


br-ex     Link encap:Ethernet  HWaddr d0:43:1e:d6:03:35 
          inet addr:  Bcast:  Mask:
          RX packets:295129 errors:0 dropped:4 overruns:0 frame:0
          TX packets:45436 errors:0 dropped:0 overruns:0 carrier:0
          collisions:0 txqueuelen:1
          RX bytes:27115572 (27.1 MB)  TX bytes:9312921 (9.3 MB)

br-int    Link encap:Ethernet  HWaddr b2:e7:33:0a:8c:4f 
          UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric ...
edit retag flag offensive close merge delete


the tunnels interfaces are created when needed. Did you try starting a vm? There are no vm ports in your br-int.

darragh-oreilly gravatar imagedarragh-oreilly ( 2016-05-27 08:18:12 -0600 )edit

Good point. I just tried spin up instance, now i have tap#### in the br-int, but still no vxlan tunnel in the br-tun.

don gravatar imagedon ( 2016-05-27 16:51:32 -0600 )edit

3 answers

Sort by ยป oldest newest most voted

answered 2016-05-29 14:51:08 -0600

don gravatar image

OK so the problem was kind of stupid :) but when neutron-server was started, it was missing --config=...ml2 --config=...openvswitch. E.g. it had the /etc/neutron/neutron.conf, but not the ml2/openvswitch ini files. This in turn causes no errors to occur anywhere, but it doesn't know what to do so it cowardly gives up :)

edit flag offensive delete link more


Hi,Don,I'm using(mitaka with ovs and vxlan ,2 physical servers,controller+network+compute on one server,and compute on another server),booting VM on the first server was perfect,but on the compute node it has no ip.When I executed the "ovs-vsctl show" there's no vxlan peers, do you finally solved?

feyman2009 gravatar imagefeyman2009 ( 2017-02-03 20:20:42 -0600 )edit

answered 2016-05-27 02:42:24 -0600

MIB gravatar image

Try below conifg : to enable l2 population.



type_drivers = vxlan

tenant_network_types = vxlan

mechanism_drivers = openvswitch,l2population

edit flag offensive delete link more


OK i tried that, but no difference.

don gravatar imagedon ( 2016-05-27 07:18:03 -0600 )edit

Hope you tired below command after making changes in /etc/..../ml2_config.ini

su -s /bin/sh -c "neutron-db-manage --config-file /etc/neutron/neutron.conf \ --config-file /etc/neutron/plugins/ml2/ml2_conf.ini upgrade head" neutron

MIB gravatar imageMIB ( 2016-05-30 02:24:37 -0600 )edit

answered 2016-05-27 00:39:41 -0600

dbaxps gravatar image

updated 2019-01-07 15:11:14 -0600

don gravatar image

You wrote :- Bridge br-tun Port "vxlan-0a640a97" Interface "vxlan-0a640a97" type: vxlan options: {df_default="true", in_key=flow, local_ip="xx.xx.xx.xx", out_key=flow, remote_ip=""}

VTEPs across all landscape usually are located on the same network say If for instance eth1 are
supposed to support VTEPs, corresponding IPs are supposed to be assigned before deployment starts ( at least on RDO)
edit flag offensive delete link more


This example is from my working system where the system has 2 IP, and its picking the other for the endpoint. But that system is working. There is no requirement for vxlan to be on a single subnet.

don gravatar imagedon ( 2016-05-27 07:16:27 -0600 )edit

Your Answer

Please start posting anonymously - your entry will be published after you log in or create a new account.

Add Answer

Get to know Ask OpenStack

Resources for moderators

Question Tools

1 follower


Asked: 2016-05-26 22:53:21 -0600

Seen: 850 times

Last updated: Jan 07