Ask Your Question
2

Juno + Centos 7 - BuildAbortException: Build of instance aborted: Failed to allocate the network(s), not rescheduling.

asked 2014-12-30 16:30:06 -0500

mazur_7 gravatar image

updated 2014-12-31 12:45:22 -0500

smaffulli gravatar image

I have a setup with Ubunt AIO 14.04 running Juno, seems to be working fine, after two months I decided to play around with adding new compute node but running on Centos 7.

Problem is that I start any instance on Centos host, every creation fails with Exception:

 BuildAbortException: Build of instance aborted: Failed to allocate the network(s), not rescheduling.

Right before this Exception there is another one which should not influence instance creation based on nova/compute/manager.py:

instance.save(expected_task_state=[None])
(...)
RemoteError: Remote error: OperationalError (OperationalError) (1048, "Column 'instance_uuid' cannot be null") 'UPDATE instance_extra SET updated_at=%s, instance_uuid=%s WHERE instance_extra.id = %s' (datetime.datetime(2014, 12, 30, 21, 34, 4, 684113), None, 150L)

Adding in nova.confg did not help (it was the same issue on Ubuntu before and following lines helped in that case):

vif_plugging_is_fatal=False 
vif_plugging_timeout=0

I am pretty sure that vif creation is successful but it happens few seconds after I start my instance whereas instance creation fails instantaneously (without waiting for vif creation), the following logs on controller node (Ubuntu) proves that such event is sent (I know here it is error but it was error due to my typo on controller node in neutron.conf):

2014-12-30 22:23:21.567 3710 ERROR neutron.notifiers.nova [-] Failed to notify nova on events: [{'status': 'completed', 'tag': u'60f129c9-02f6-43ac-b15c-29bdd47005e1', 'name': 'network-vif-plugged', 'server_uuid': u'09bbb54c-68c3-4c64-bdce-4a63aa78a699'}, {'status': 'completed', 'tag': u'ef56d76f-79c9-41f0-8f41-659bcfa8527d', 'name': 'network-vif-plugged', 'server_uuid': u'83c11b41-1420-4693-ba04-faf4afbbee06'}]

I assume that as soon as this event is received on compute node, neutron agent there instructs OVS to create appropriate interfaces with appropriate IP addressing.

Does anybody have a clue how to troubleshoot this issue ? Or are there any caveats on adding new compute nodes ?

edit retag flag offensive close merge delete

Comments

Seems that the same problem is described here for Ubuntu only case;

https://ask.openstack.org/en/question...

mazur_7 gravatar imagemazur_7 ( 2015-01-08 04:58:15 -0500 )edit

3 answers

Sort by ยป oldest newest most voted
1

answered 2015-01-08 05:37:00 -0500

mazur_7 gravatar image

updated 2015-01-08 05:38:19 -0500

Problem solved with the following EPEL packages:

openstack-nova-common-2014.2-2.el7.centos.noarch.rpm

openstack-nova-compute-2014.2-2.el7.centos.noarch.rpm

python-nova-2014.2-2.el7.centos.noarch.rpm

I had to forcefully install them using rpm -ivh --force

Then I restarted openstack-nova-compute service

After this instances spawned correctly, yet I could not reach them (iptables rules missing, so I disabled iptables).

DHCP addresses got distributed to the instances correctly and I could reach internet from them via my Ubuntu-AIO which acts as a gateway to the internet as well.

edit flag offensive delete link more
0

answered 2015-09-15 04:33:40 -0500

I encountered the same error and fixed by putting

vif_plugging_timeout=10
vif_plugging_is_fatal=False

in the nova.conf file on the compute node.

edit flag offensive delete link more
-1

answered 2015-12-20 23:36:59 -0500

Satender Yadav gravatar image

open this file /etc/nova/nova.conf and add below line in [DEFAULT]

vif_plugging_timeout=10 vif_plugging_is_fatal=False

edit flag offensive delete link more

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

3 followers

Stats

Asked: 2014-12-30 16:30:06 -0500

Seen: 3,895 times

Last updated: Sep 15 '15