[Openstack-operators] Trying to deploy an instance..running out of disk space on the compute node?

Abel Lopez alopgeek at gmail.com
Thu Jan 17 17:24:10 UTC 2013

Is /var not a separate filesystem? By default, your instances run out of
Make sure that has enough space

On Thursday, January 17, 2013, Zach Easterbrook wrote:

> Foreword, all of this is on essex.
> I'm trying to deploy a virtual machine image I've made myself following a
> procedure similar to the one outlined here:
> http://docs.openstack.org/trunk/openstack-compute/admin/content/manually-creating-qcow2-images.html#d6e6214(I used virt-manager and x11 forwarding instead of VNC to connect to the
> instance)
> However, when I attempt to boot this image using the following command: (6
> is a custom flavor)
> nova boot --poll --flavor 6 --image <img id> --key_name my_key
> --availability_zone=zone1 t1dbvm
> it will say "Error building instance"
> on the compute node I'm seeing this:
> http://paste.debian.net/225727/
> Then a bit further down:
> 2013-01-17 11:58:41 TRACE nova.rpc.amqp IOError: [Errno 28] No space left
> on device
> 2013-01-17 11:58:41 TRACE nova.rpc.amqp
> I'm wondering if my root partition is too small. The image I'm trying to
> deploy is 10GB but I only have ~7.6GB available on root, (but 400GB+ on
> /home).
> Any thoughts? I know this VM image is bootable, I can boot it manually
> using virt-manager.
> PS: I don't know if it's related but in the nova-compute.log I'm also
> seeing "Instance found in database but not known by hypervisor. Setting
> power to NOSTATE."
> Help is appreciated, thanks very much in advance.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstack.org/pipermail/openstack-operators/attachments/20130117/12769c3f/attachment.html>

More information about the OpenStack-operators mailing list