[Openstack] Spawning Issue with Juno, Neutron, Xen and Libvirt

Anthony PERARD anthony.perard at citrix.com
Wed Jul 22 16:35:49 UTC 2015


On Mon, Jul 20, 2015 at 05:28:00PM -0400, Geoffrey Tran wrote:
> Hello all,

Hi Geoffrey,

> I've been trying to get a single-node installation up and running with
> Juno, neutron, Xen, and libvirt. However, there has been an issue 
> with actually launching a guest.  The error that shows up in 
> /var/log/libvirt/libxl/instance-XXXXXXXX.log  is below:
> 
> > libxl: error: libxl_exec.c:118:libxl_report_child_exitstatus: 
> >    /etc/xen/scripts/vif-bridge online [-1] exited with error status 1
> > libxl: error: libxl_device.c:1085:device_hotplug_child_death_cb: 
> >    script: ip link set vif1.0 name tap20b72258-74 failed
> > libxl: debug: libxl_event.c:618:libxl__ev_xswatch_deregister: watch 
> >    w=0x7f6440009c20: deregister unregistered
> > libxl: error: libxl_create.c:1226:domcreate_attach_vtpms: unable to 
> >    add nic devices
> 
> I am following the steps to get setup as shown here: 
> https://blog.xenproject.org/2014/12/01/using-xen-project-on-openstack-juno-via-libvirt/
> 
> The only difference is that I am running all services on a signle node.
> 
> The following sites mention this error, but so far there are no other 
> solutions other than using nova-networking.
> > https://ask.openstack.org/en/question/64942/nova-libvirt-xen-driver-and-patch-feb-2015-in-upstream-libvirt/
> > http://www.gossamer-threads.com/lists/xen/devel/374061
> > http://lists.xen.org/archives/html/xen-devel/2015-04/msg01723.html
> 
> If I could get any assistance, it would be very much appreciated!

Sorry, I don't have a much better answer. But for some reason, I don't have
this specific issue with recent Nova/Neutron (master, not yet released).

I don't know much of what Neutron is doing with the network interface but
we can maybe try to work arround this issue.
Could you add "exit 0" at the begin of the /etc/xen/scripts/vif-bridge
script? Then check if you've got network access to the guest.

Thanks,

-- 
Anthony PERARD




More information about the Openstack mailing list