[Openstack-operators] [glance] Image enters "killed" state on upload

Liam Haworth liam.haworth at bluereef.com.au
Wed Feb 3 01:57:54 UTC 2016


After searching my configs nothing is out of place and I have even tried
setting the permissions on the images folder to 777 to see if that helps
but nothing has helped and glance api keeps throwing this:
http://paste.openstack.org/show/485794/

On Wed, 3 Feb 2016 at 11:51 Liam Haworth <liam.haworth at bluereef.com.au>
wrote:

> Sooooo, just tired uploading the image via CLI with this command:
>
> glance --os-image-api-version 1 image-create --progress --file
> ./polygraph-dns-services.qcow2 --min-disk 5 --container-format bare
> --disk-format qcow2 --name polygraph-dns-services
>
> Got this error:
>
> 410 Gone: Error in store configuration. Adding images to store is
> disabled. (HTTP N/A)
>
> So I mustn't have something configured properly so I shall go digging in
> my config. What *is* annoying though is that the dashboard said the
> upload went fine where as the CLI tool tells you the hard truth, errrrg.
>
> Thanks for all the help guys! I shall respond with a reply to this thread
> once I have found the afflicting configuration options.
>
> On Wed, 3 Feb 2016 at 10:40 Liam Haworth <liam.haworth at bluereef.com.au>
> wrote:
>
>> In response to Kris,
>>
>> On ctrl1: glance:x:112:120::/var/lib/glance:/bin/false
>> On ctrl2: glance:x:112:119::/var/lib/glance:/bin/false
>>
>> 120 on ctrl2 is nova, so on ctrl1 the perms would be glance:glance and on
>> ctrl2 it would glance:nova. I would suspect this as the issue if the issue
>> has happening on ctrl2 but it is happening on ctrl1
>>
>> On Wed, 3 Feb 2016 at 10:34 Kris G. Lindgren <klindgren at godaddy.com>
>> wrote:
>>
>>> Not related to your issue, but something to keep an eye out for, is that
>>> you need to keep the uid for glance synced across your glances servers when
>>> using an nfsv3 store.  Since nfsv3 stores the uid & gid for the file
>>> perms.  You can run into weird issues if glance is uid/gid 501 on one
>>> glance server and 502 on another.  We had that problem crop up in
>>> production when packages were doing "useradd" without specifying a
>>> uid/gid.  So you could end up with systems with different id's and
>>> permissions that are all screwed up between multiple servers.
>>>
>>> So related to your question .. If I remember correctly you need
>>> read/execute permissions to list the contents/enter a directory under linux.
>>>
>>> ___________________________________________________________________
>>> Kris Lindgren
>>> Senior Linux Systems Engineer
>>> GoDaddy
>>>
>>> From: Liam Haworth <liam.haworth at bluereef.com.au>
>>> Date: Tuesday, February 2, 2016 at 4:25 PM
>>> To: Abel Lopez <alopgeek at gmail.com>
>>> Cc: "openstack-operators at lists.openstack.org" <
>>> openstack-operators at lists.openstack.org>
>>> Subject: Re: [Openstack-operators] [glance] Image enters "killed" state
>>> on upload
>>>
>>> Here is a output from my system instead of my blabbering in a long
>>> winded email
>>>
>>> root at ctrl1:~# uname -a
>>> Linux ctrl1 3.19.0-43-generic #49~14.04.1-Ubuntu SMP Thu Dec 31 15:44:49
>>> UTC 2015 x86_64 x86_64 x86_64 GNU/Linux
>>>
>>> root at ctrl1:~# df -h
>>> Filesystem                                  Size        Used
>>> Avail     Use%         Mounted on
>>> udev                                           7.9G       4.0K
>>> 7.9G         1%          /dev
>>> tmpfs                                          1.6G       724K
>>> 1.6G         1%          /run
>>> /dev/mapper/ctrl1--vg-root        396G      6.4G        370G         2%
>>>         /
>>> none                                           4.0K
>>> 0               4.0K         0%         /sys/fs/cgroup
>>> none                                           5.0M       0
>>>               5.0M         0%        /run/lock
>>> none                                           7.9G
>>> 0              7.9G          0%        /run/shm
>>> none                                           100M      0
>>> 100M         0%        /run/user
>>> /dev/sdc1                                   236M      38M
>>> 186M       17%        /boot
>>> 10.16.16.30:/srv/glance              739G     97G          604G
>>> 14%        /var/lib/glance/images
>>>
>>> And to save from massed output from a LS, ever file in
>>> /var/lib/glance/images is: -rw-r----- 1 glance glance
>>>
>>> No apparmour installed or configured
>>>
>>> On Wed, 3 Feb 2016 at 10:17 Abel Lopez <alopgeek at gmail.com> wrote:
>>>
>>>> Ok, with file store, some of the silly things that crop up are around
>>>> directory permissions, disk space, SELinux/apparmour.
>>>>
>>>> Make sure the glance user and group have ownership (recursively) of the
>>>> /var/lib/glance directory, make sure you're not low on space, if you have
>>>> SELinux set to enforcing, test setting it to permissive (if that is the
>>>> issue, resolve the contexts)
>>>>
>>>> On Feb 2, 2016, at 3:13 PM, Liam Haworth <liam.haworth at bluereef.com.au>
>>>> wrote:
>>>>
>>>> Glance is configured to use file store to /var/lib/glance/images
>>>>
>>>> On Wed, 3 Feb 2016 at 10:12 Abel Lopez <alopgeek at gmail.com> wrote:
>>>>
>>>>> I ran into a similar issue in Havana, but that was because we were
>>>>> doing some 'behind-the-scenes' modification of the image (format
>>>>> conversion)
>>>>> Once we stopped that, the issue went away.
>>>>>
>>>>> What is your glance store configured as?
>>>>>
>>>>> On Feb 2, 2016, at 3:05 PM, Liam Haworth <liam.haworth at bluereef.com.au>
>>>>> wrote:
>>>>>
>>>>> Hey All,
>>>>>
>>>>> This sounds like an old bug after trying to google it but everything I
>>>>> found doesn't really seem to help. I'm trying to upload a 2.5GB QCOW2 image
>>>>> to glance to be used by users, the upload goes fine and in the glance
>>>>> registry logs I can see that it has successfully saved the image but then
>>>>> it does this
>>>>>
>>>>> 2016-02-03 09:51:49.607 2826 DEBUG glance.registry.api.v1.images
>>>>> [req-5ba18ea3-5777-4023-9f85-040aca48dfa7 --trunced-- - - -] Updating image
>>>>> 03a920ce-7979-4439-ab71-bc3dd34df3d3 with metadata: {u'status': u'killed'}
>>>>> update /usr/lib/python2.7/dist-packages/glance/registry/api/v1/images.py:470
>>>>>
>>>>> What reasons are their for it to do this to an image that just
>>>>> successfully uploaded?
>>>>>
>>>>> Thanks,
>>>>>
>>>>> Liam Haworth.
>>>>> --
>>>>> *Liam Haworth* | Junior Software Engineer | www.bluereef.com.au
>>>>> *_________________________________________________________________*
>>>>> *T*: +61 3 9898 8000 | *F*: +61 3 9898 8055
>>>>>
>>>>>
>>>>> _______________________________________________
>>>>> OpenStack-operators mailing list
>>>>> OpenStack-operators at lists.openstack.org
>>>>> http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-operators
>>>>>
>>>>>
>>>>> --
>>>> *Liam Haworth* | Junior Software Engineer | www.bluereef.com.au
>>>> *_________________________________________________________________*
>>>> *T*: +61 3 9898 8000 | *F*: +61 3 9898 8055
>>>>
>>>>
>>>>
>>>> --
>>> *Liam Haworth* | Junior Software Engineer | www.bluereef.com.au
>>> *_________________________________________________________________*
>>> *T*: +61 3 9898 8000 | *F*: +61 3 9898 8055
>>>
>>>
>>> --
>> *Liam Haworth* | Junior Software Engineer | www.bluereef.com.au
>> *_________________________________________________________________*
>> *T*: +61 3 9898 8000 | *F*: +61 3 9898 8055
>>
>>
>> --
> *Liam Haworth* | Junior Software Engineer | www.bluereef.com.au
> *_________________________________________________________________*
> *T*: +61 3 9898 8000 | *F*: +61 3 9898 8055
>
>
> --
*Liam Haworth* | Junior Software Engineer | www.bluereef.com.au
*_________________________________________________________________*
*T*: +61 3 9898 8000 | *F*: +61 3 9898 8055
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstack.org/pipermail/openstack-operators/attachments/20160203/ee37340e/attachment.html>


More information about the OpenStack-operators mailing list