[Openstack-operators] Havana / RDO / MultiNode - image deploy issue
robertc at robertcollins.net
Mon Feb 3 05:23:36 UTC 2014
On 2 February 2014 16:15, Tim Chipman <tim.chipman at gmail.com> wrote:
> Hi, a quick question I wonder if any pointers could be offered,
> I am in the process of building out a multi-node Havana Openstack setup (2
> nodes currently, endpoint is 8) and I have an odd problem.
> (storage is on local dedicated LVM on eaach node/ for Ceph and then exported
> via iScsi more or less under the hood)
> - I can deploy new Images from the CirrOS image fine
> - was able to deploy using an ubuntu 10.04 stock cloud qcow2
> - have tried a few others (CentOS 6.X and a newer Ubuntu). I seem to have
> issues with these; the image deploys and after ~1-2 minutes goes to an
> 'error' state in web admin interface; not a lot more detail there.
> In the logs, I see:
> 2014-02-01 22:43:39.612 36406 TRACE cinder.openstack.common.rpc.amqp
> ImageUnacceptable: Image d13eae68-2f1c-41be-87e8-43b4eef6a2f0 is
> unacceptable: Size is 16GB and doesn't fit in a volume of size 1GB.
> which seems a fairly clear concern; except ... I am ? in theory?
> provisioning my new images/VMs with adequately large disk space (m2.small
> has 20gb disk, exceeds 16gb; plus VDA of 20gb being specified)
The error was from cinder, not nova, so this suggests to me that
you're copying an image to a cinder volume, for subsequent
boot-from-volume. The cinder device could be too small... but I
haven't made much use of cinder yet (so don't know where the skeletons
are buried :)).
> Possibly I have been working on this too many hours, so simply can't see
> what is right in front of me.
Whats the output of 'nova flavor-list' and secondly, whats the *exact*
command line you are using to boot your instance?
Robert Collins <rbtcollins at hp.com>
HP Converged Cloud
More information about the OpenStack-operators