Ask Your Question
1

Ubuntu 14.04 cloud image won't boot [closed]

asked 2014-06-06 08:25:04 -0600

marcantonio gravatar image

updated 2014-06-11 09:51:03 -0600

I have a new Icehouse environment setup. I can't seem to boot any instances from the Ubuntu Server 14.04 cloud image. I am able to bring up CirrOS and 12.04 just fine. There's no error but it never gets past cloud-init:

Cloud-init v. 0.7.5 running 'init' at Thu, 05 Jun 2014 20:51:10 +0000. Up 7.76 seconds.
ci-info: ++++++++++++++++++++++++++Net device info+++++++++++++++++++++++++++
ci-info: +--------+------+--------------+---------------+-------------------+
ci-info: | Device |  Up  |   Address    |      Mask     |     Hw-Address    |
ci-info: +--------+------+--------------+---------------+-------------------+
ci-info: |   lo   | True |  127.0.0.1   |   255.0.0.0   |         .         |
ci-info: |  eth0  | True | 192.168.1.19 | 255.255.255.0 | fa:16:3e:ae:ae:72 |
ci-info: +--------+------+--------------+---------------+-------------------+
ci-info: +++++++++++++++++++++++++++++++Route info++++++++++++++++++++++++++++++++
ci-info: +-------+-------------+-------------+---------------+-----------+-------+
ci-info: | Route | Destination |   Gateway   |    Genmask    | Interface | Flags |
ci-info: +-------+-------------+-------------+---------------+-----------+-------+
ci-info: |   0   |   0.0.0.0   | 192.168.1.1 |    0.0.0.0    |    eth0   |   UG  |
ci-info: |   1   | 192.168.1.0 |   0.0.0.0   | 255.255.255.0 |    eth0   |   U   |
ci-info: +-------+-------------+-------------+---------------+-----------+-------+

Update 1:

I found this question which exihibited the same problem I'm having. However, I believe my metadata service is functioning properly as from a cirros instance I can curl http://169.254.169.254/ properly.

Update 2:

I finally found something! From a cirros instance I can query the metadata service:

$ curl http://169.254.169.254/2009-04-04/meta-data/placement/availability-zone
nova$

However, running the following query hangs forever:

$ curl http://169.254.169.254/openstack/2013-10-17/meta_data.json

Which is where I'm thinking the difference is between 14.04 and 12.04 or cirros. Is that an API that I don't have enabled?

Thanks, Marc

edit retag flag offensive reopen merge delete

Closed for the following reason the question is answered, right answer was accepted by koolhead17
close date 2014-06-17 07:42:54.258109

Comments

What says /var/log/nova/nova-compute.log on Compute Node during boot up ?

dbaxps gravatar imagedbaxps ( 2014-06-06 08:37:39 -0600 )edit

There's a warning, not sure that it's relevant:

2014-06-06 09:44:55.916 2019 WARNING nova.virt.disk.vfs.guestfs [req-fb4c5d8c-09b7-4ad6-9781-bb6b7a2d3f1f b0e6759357334a92bd7a7cb4a7e23e0c 66904b47fddc4aeb95b973596467c596] Failed to close augeas aug_close: call launch before using this function
marcantonio gravatar imagemarcantonio ( 2014-06-06 08:44:27 -0600 )edit

If you scroll to the bottom , nothing else ?

dbaxps gravatar imagedbaxps ( 2014-06-06 08:56:31 -0600 )edit

Please post /var/log/nova/nova-scheduler.log (Controller) . Try load flavor 2.

dbaxps gravatar imagedbaxps ( 2014-06-06 14:29:51 -0600 )edit

I'm not sure why. The instance is being built successfully. As far as OpenStack is concerned everything is good. There are no errors in that log.

It's the VM that won't boot fully.

marcantonio gravatar imagemarcantonio ( 2014-06-06 14:32:59 -0600 )edit

4 answers

Sort by ยป oldest newest most voted
2

answered 2014-06-10 20:49:47 -0600

marcantonio gravatar image

updated 2014-06-11 09:50:49 -0600

I finally found the solution to this. I started another thread and posted the answer there.

In short, all large packets were not making it back to the instances. I found that, because I'm using GRE tunnels with Neutron, I had to lower the MTU on the instances.

Thanks all.

edit flag offensive delete link more

Comments

Thanks a lot! This was driving me crazy. Adding the option to DHCP/DNSmasq did the trick.

ed gravatar imageed ( 2014-06-11 21:11:18 -0600 )edit
-1

answered 2014-06-07 04:48:07 -0600

dbaxps gravatar image
edit flag offensive delete link more

Comments

These do not seem to be about the same issue. In my case the instance spawns successfully even though it fails to resize.

sameer gravatar imagesameer ( 2014-06-07 05:40:47 -0600 )edit

I agree, my instances don't fail as far as nova list is concerned.

marcantonio gravatar imagemarcantonio ( 2014-06-09 09:13:08 -0600 )edit
1

answered 2014-06-06 20:11:59 -0600

kordless gravatar image

updated 2014-06-07 11:03:51 -0600

Note: Editing my 'answer' because the first assumption about flavor size was incorrect. I'll leave the text of that below for completeness as it is relevant to the "won't boot" part of the question.

The spot where your instance stops is clearly involving cloud-init. I've had problems before with instances not being able to contact the metadata server that should be running on 169.254.169.254. Try pinging 169.254.169.254 from the controller node and see if you get a response. Try booting a cirros image and see if you can ping that address as well from the instance.

ubuntu@stackmonkey-va:/var/www/xoviova$ ping 169.254.169.254
PING 169.254.169.254 (169.254.169.254) 56(84) bytes of data.
64 bytes from 169.254.169.254: icmp_req=1 ttl=64 time=0.244 ms
64 bytes from 169.254.169.254: icmp_req=2 ttl=64 time=0.257 ms

Original answer was as follows:

Not sure if it is relevant, but I encountered errors with launching Ubuntu images with an m1.tiny. The issue was a lack of available space on the flavor's disk, which is now a default value of 1G in Icehouse. I increased the disk size by editing the flavor and the problem went away.

edit flag offensive delete link more

Comments

Due to "Attempting claim: memory 4096 MB, disk 40 GB, VCPUs 2" it's m1.medium

DanIzack gravatar imageDanIzack ( 2014-06-06 23:33:30 -0600 )edit

Thanks for pointing that out. I missed it when looking at the log. Thinking it's something to do with cloud-init and/or the lack of being able to connect to 169.254.169.254.

kordless gravatar imagekordless ( 2014-06-07 10:54:03 -0600 )edit

Thanks! This was indeed my issue. The metadata-agent on neutron was configured with the wrong region name. Once fixed my Ubuntu was able to finish booting (cloud-tools completed). I would've marked this as an answer but I'm not the OP :-)

sameer gravatar imagesameer ( 2014-06-08 09:24:18 -0600 )edit

I can't ping it but, from a cirros instance, I can query it with curl. Also, it is an m1.medium. I can't even boot Ubuntu will m1.tiny.

marcantonio gravatar imagemarcantonio ( 2014-06-09 08:55:08 -0600 )edit

@marcantonio, You are not supposed to ping 169.254.169.254, you are supposed to curl http://169.254.169.254/latest/meta-data/ .
Can you $curl http://msn.com ? ( or whatever on the Net )

dbaxps gravatar imagedbaxps ( 2014-06-09 09:43:32 -0600 )edit
1

answered 2014-06-06 13:10:02 -0600

DanIzack gravatar image

updated 2014-06-06 13:11:00 -0600

Download Ubuntu 14.04 qcow2 image here and make sure that it's working fine.

edit flag offensive delete link more

Comments

Confirmed. That is the exact one that I am using: https://cloud-images.ubuntu.com/trusty/current/trusty-server-cloudimg-amd64-disk1.img (https://cloud-images.ubuntu.com/trust...)

marcantonio gravatar imagemarcantonio ( 2014-06-06 13:16:16 -0600 )edit

Please, provide /var/log/nova/nova-compute.log when launching Trusty VM ( booting up protocol ).

DanIzack gravatar imageDanIzack ( 2014-06-06 13:20:48 -0600 )edit

I updated my question with the log. Thanks.

marcantonio gravatar imagemarcantonio ( 2014-06-06 13:35:45 -0600 )edit

Get to know Ask OpenStack

Resources for moderators

Question Tools

3 followers

Stats

Asked: 2014-06-06 08:25:04 -0600

Seen: 4,105 times

Last updated: Jun 11 '14