[Removed rhos-maint@redhat.com as I saw an email sent to it failed and
could not be delivered]

Hi Kashyap,

Thanks a lot for your reply. I do see the fix in my environment.

In /usr/lib/python2.7/site-packages/neutron/common/config.py, I see the
following code snippet at line numbers 89 and 90.

     89     cfg.StrOpt('nova_url',
     90                default='http://127.0.0.1:8774/v2',

In /etc/neutron/neutron.conf, I do have the Nova v2 API version number in 
the URL below. But, instead of 127.0.0.1, I have the IP address of the network
host which is same as CONFIG_NETWORK_HOSTS in my packstack-answers.txt.

# URL for connection to nova (Only supports one nova region currently).
nova_url = http://172.29.68.136:8774/v2

I have the below line in my packstack-answers.txt.

CONFIG_NETWORK_HOSTS=172.29.68.136

Yes, I can reproduce this issue every single time. Am I seeing a new bug ?

Please let me know if there is any work-around, and also if I need to send any
logs/debugs.

Regards,
Vikram Hosakote
OpenStack Software Engineer    |    vhosakot@cisco.com
Cloud and Virtualization Group  |    Cisco Systems
Boxborough MA                            |    Work : 978-936-8799

From: Kashyap Chamarthy <kchamart@redhat.com>
Date: Tuesday, January 6, 2015 at 5:35 AM
To: Vikram Hosakote <vhosakot@cisco.com>
Cc: "rdo-list@redhat.com" <rdo-list@redhat.com>, "rhos-maint@redhat.com" <rhos-maint@redhat.com>
Subject: Re: [Rdo-list] RDO neutron bug 1087589

On Tue, Jan 06, 2015 at 02:34:14AM +0000, Vikram Hosakote (vhosakot) wrote:
Hi,
I have been hitting the RDO bug
https://bugzilla.redhat.com/show_bug.cgi?id=1087589, and not able to
boot a nova instance.

As you might have noticed, that bug is fixed upstream more than 6 months
ago, so it's reasonable to expect to have it in current RDO stable
versions for sure (Juno).

I see that the fix https://review.openstack.org/#/c/92387/ was merged
on May 29th 2014. Is this fix available in the latest RDO ?

I checked the source of python-neutron RPM for for both CentOS (from
latest RDO) and your package version -- the above fix is available in
both.

   -  python-neutron-2014.2.1-1.el7.centos.noarch.rpm (From here
      https://repos.fedorapeople.org/repos/openstack/openstack-juno/epel-7/)
   -  python-neutron-2014.1.2-2.el7ost.noarch.rpm

Below are the package versions of neutron and nova in my RHEL 7
IceHouse setup. How can I check if the fix is available in my packages
below ?

That fix is part of this file:

    usr/lib/python2.7/site-packages/neutron/common/config.py

You can try two things:

  - In this file
    /usr/lib/python2.7/site-packages/neutron/common/config.py, you can
    check if you have the relevant code snippet:

      . . .
      cfg.StrOpt('nova_url',
                     default='http://127.0.0.1:8774/v2',
      . . .                        

  - In your /etc/neutron.conf, ensure the 'nova_url' config parameter
    has the Nova v2 API version number in the URL:

      nova_url = http://127.0.0.1:8774/v2


[. . .]

Below are the errors of the above bug I see when I boot an instance. Please let me know the solution/work-around. Thanks a lot!
In /var/log/nova/nova-compute.log:
==============================
2015-01-05 11:31:15.870 5275 ERROR nova.compute.manager [req-9dee8033-a261-4a89-9477-772141cb747e ca5d4869d4e2464bb6e1e3c12a82c3a2 4dd830d71afb48d6a5e3970249669970] [instance: 228289f2-d3c2-401d-b779-f359121b4e71] Instance failed to spawn
2015-01-05 11:31:15.870 5275 TRACE nova.compute.manager [instance: 228289f2-d3c2-401d-b779-f359121b4e71] Traceback (most recent call last):
2015-01-05 11:31:15.870 5275 TRACE nova.compute.manager [instance: 228289f2-d3c2-401d-b779-f359121b4e71]   File "/usr/lib/python2.7/site-packages/nova/compute/manager.py", line 1738, in _spawn
2015-01-05 11:31:15.870 5275 TRACE nova.compute.manager [instance: 228289f2-d3c2-401d-b779-f359121b4e71]     block_device_info)
2015-01-05 11:31:15.870 5275 TRACE nova.compute.manager [instance: 228289f2-d3c2-401d-b779-f359121b4e71]   File "/usr/lib/python2.7/site-packages/nova/virt/libvirt/driver.py", line 2289, in spawn
2015-01-05 11:31:15.870 5275 TRACE nova.compute.manager [instance: 228289f2-d3c2-401d-b779-f359121b4e71]     block_device_info)
2015-01-05 11:31:15.870 5275 TRACE nova.compute.manager [instance: 228289f2-d3c2-401d-b779-f359121b4e71]   File "/usr/lib/python2.7/site-packages/nova/virt/libvirt/driver.py", line 3710, in _create_domain_and_network
2015-01-05 11:31:15.870 5275 TRACE nova.compute.manager [instance: 228289f2-d3c2-401d-b779-f359121b4e71]     raise exception.VirtualInterfaceCreateException()
2015-01-05 11:31:15.870 5275 TRACE nova.compute.manager [instance: 228289f2-d3c2-401d-b779-f359121b4e71] VirtualInterfaceCreateException: Virtual Interface creation failed

Are you able to consistently reproduce this issue in a different test
environment?


--
/kashyap