fuel-dev team mailing list archive

[Dmitry Borodaenko <dborodaenko@xxxxxxxxxxxx>]

On Wed, Feb 12, 2014 at 7:41 AM, Vladimir Kuklin <vkuklin@xxxxxxxxxxxx> wrote:
> Vladimir,
>
> Please clarify current status for these bugs:
>
> https://bugs.launchpad.net/fuel/+bug/1271549
> https://bugs.launchpad.net/fuel/+bug/1277151
> https://bugs.launchpad.net/fuel/+bug/1263648
>
>
>
> AFAIU, https://bugs.launchpad.net/fuel/+bug/1277151 is the main bug and
> there are change requests
>
> https://review.openstack.org/#/c/71836/ and
> https://review.openstack.org/#/c/71885/
>
> waiting for it. The main problem is that it breaks kvm-based tests as we
> need to patch devops correspondingly.

The change requests above are designed to fall back to by-path
specifically to remain compatible with kvm virtio devices. I've done a
test deployment in a devops-driven kvm based test environment with
both patches applied and it worked fine. I've approved both patches
and they are now merged to fuel-web and fuel-library master branches.

> Bug https://bugs.launchpad.net/fuel/+bug/1263648 should be fixed by writing
> patch for nailgun-agent which uses block device id before it falls back to
> by-path or </dev/sda|/dev/sdb> definition.
>
> Bug https://bugs.launchpad.net/fuel/+bug/1271549 is duplicate of the main
> bug.

I think it's the other way around: #1263648 is the duplicate, and
#1271549 is a similar bug that may be fixed by the above two change
requests.

> Thus, we need to only merge the patches for the main bug, fix devops library
> for our tests and fix nailgun agent for Ubuntu. Then we can write better
> methods of disks naming for 5.0 release. Please, correct me if I am wrong:

With the above fixes merged, I think our disk device identification
should now be reliable enough, is there a plan for further changes in
5.0?

-- 
Dmitry Borodaenko