[Openstack-operators] /var/lib/nova/instances fs filled up corrupting my Linux instances

Michael Still mikal at stillhq.com
Wed Mar 13 14:53:25 UTC 2013


On Wed, Mar 13, 2013 at 9:42 AM, Conor Glynn <cglynn at sportingindex.com> wrote:
> I ran:
> cd /var/lib/nova/instances; find -name "disk*" | xargs -n1 qemu-img info |
> grep backing | sed -e's/.*file: //' -e 's/ .*//' | sort | uniq > /tmp/inuse
> then I removed some of the old _base files that did not exist in /tmp/inuse

That should be ok, although you should probably note that you can also
just turn on the image cache manager in folsom and it will magically
keep the base directory clean for you. That manual cleanup command
line is only really needed for essex these days.

Note, if you have multiple compute nodes all storing instances to a
shared NFS directory this is not true, and the image cache manager is
dangerous in folsom. This is true for your cleanup script as well.
Were both compute nodes using the same NFS directory?

I went and had a look at my blog about this stuff and realised its not
as clear as I thought it was. I'll work on fixing that ASAP.

Michael



More information about the OpenStack-operators mailing list