asked 2016-08-12 00:19:28 -0800

srik1782 gravatar image

Hi, I am trying to deploy an Openstack VxLAN setup (Mitaka Release) with ODL as SDN controller, OVS DPDK installed in compute node. I deployed my setup through Fuel9 community build. My setup is as follows: o One Openstack Controller with ODL controller installed o One Openstack Compute node with OVS installed with DPDK. I upgraded this to Ubuntu 16.04 and re-attached to Openstack Controller, to upgrade libvirt, qemu.

I am able to launch a VM with ‘vhostuser’ as port type. But, I am not able to get IP addresses through DHCP. Following are some command outputs:

root@node-23:~# ovs-vsctl show 8695ab71-97c1-4d2b-a3b3-813adadf1872
    Manager "tcp:"
        is_connected: true
    Bridge br-int
        Controller "tcp:"
        fail_mode: secure
        Port br-int
            Interface br-int
                type: internal
        Port "vxlan-"
            Interface "vxlan-"
                type: vxlan
                options: {key=flow, local_ip="", remote_ip=""}
        Port "vhuc7775dd1-63"
            Interface "vhuc7775dd1-63"
                type: dpdkvhostuser
    ovs_version: "2.5.0"

I have few questions regarding the above scenario:

  1. To which bridge ‘dpdk0’ need to be attached?
  2. By default, VxLAN ports are attaching to ‘br-int’ bridge, which is of ‘netdev’ datapth_type. So, compute node is not reachable to Openstack Network Node. How can a VM get IP address from neutron DHCP?
  3. Do we need to do any manual steps to get ping working from VM – VM which are brought up in the same compute node?

Regards, Srikanth.

edit retag flag offensive close merge delete


Hi Srikanth, I'm also facing the same issue with the exact same set up that you have mentioned (Fuel 9, 1 controller+odl and 1 compute with ovs-dpdk). Were you able to solve this issue. Please share your findings and fixes you have tried. Thanks, -Veena

veena ( 2016-09-11 21:28:31 -0800 )edit