Open Stack

Dmitry,
thank you for clarification of a status. In a meeting we had [1], I
misinterpreted a status of a feature is being fully completed, and that is
what I reflected in my FFEs post [2].

I'm not sure if we have definition of done written for Fuel specifically,
but my expectation from the spec [3] was that Ubuntu bootstrap has to be
default option. That means that all related code has to be merged.
One of the patches you've mentioned goes on since November, 27 and I'm not
sure how long would it take to complete it. It is not a simple switch in a
config file as I'd expect to see.
It's not a 100+ LOC changeset in very core either, so I call for Matt as a
maintainer [4] and core reviewers to assess a situation.

My perspective is that we can test it right now without this patch and
steps to do so are known [6], so feature is actually nearly complete except
of additional action in fuel-menu. I hope that we could go an extra mile
here and maintainer / cores could help with the patch.

I don't personally think that we could go on without having a clear
deadline for when it will be enabled. Even if [4] is merged, but ubuntu is
not a default - I see a serious risks here, as majority of our automated
tests and lots of manual tests done by random people will still be using
another image, thus significantly decreasing our chances to see regressions
on time.

I'd give it an exception till end of Tuesday (as of after CentOS 7 expected
merges) for full enablement, and that should have a fix to critical issue
[7].

[1] https://etherpad.openstack.org/p/fuel-8.0-FF-meeting
[2]
http://lists.openstack.org/pipermail/openstack-dev/2015-December/081131.html
[3]
http://specs.fuel-infra.org/fuel-specs-master/specs/8.0/dynamically-build-bootstrap.html
[4] https://review.openstack.org/#/c/250662/
[5] https://github.com/openstack/fuel-menu/blob/master/MAINTAINERS
[6] https://bugs.launchpad.net/fuel/+bug/1522406/comments/3
[7] https://bugs.launchpad.net/fuel/+bug/1522406
<https://bugs.launchpad.net/fuel/+bug/1522406/comments/3>

On Thu, Dec 3, 2015 at 10:59 AM Dmitry Klenov <dklenov at mirantis.com> wrote:

> Hi folks,
>
> Let me clarify the situation with Ubuntu bootstrap feature.
>
> First of all, I would like to highlight that all functional commits for
> this feature were merged. This means that starting from yesterday everyone
> has an ability to switch to Ubuntu-based bootstrap manually and start using
> it. So I do not see the risk in loosing testing cycles in the community.
>
> The item which brought concerns on today status meeting was the enablement
> of the feature by default. To do it, 2 patches have to be merged together:
> * https://review.openstack.org/#/c/250662/ - main switch.
> * https://review.openstack.org/#/c/251908/ - compatibility commit to QA
> suite to comply with new bootstrap config format.
>
> I would like to raise the question if we can have a feature freeze
> exception for these 2 patches?
>
> There are a couple of reasons why I consider it safer to merge these
> patches several days later:
> * There is a bug caught today which will block the tests to pass if we
> switch to Ubuntu bootstrap by default:
> https://bugs.launchpad.net/fuel/+bug/1522406
> * There were concerns that a lot of FF commit merges can bring instability
> to QA suite. So it might be reasonable not to bring one more variable right
> now and to enable ubuntu bootstrap by default when all automated tests are
> stabilized.
>
> I would like to ask engineering and QA leads to express their ideas on
> this.
>
> Thanks,
> Dmitry.
> __________________________________________________________________________
> OpenStack Development Mailing List (not for usage questions)
> Unsubscribe: OpenStack-dev-request at lists.openstack.org?subject:unsubscribe
> http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
>
-- 
Mike Scherbakov
#mihgen
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstack.org/pipermail/openstack-dev/attachments/20151204/4d75964e/attachment.html>