Open Stack

On Fri, Jun 2, 2017 at 11:42 AM, Attila Darazs <adarazs at redhat.com> wrote:

> If the topics below interest you and you want to contribute to the
> discussion, feel free to join the next meeting:
>
> Time: Thursdays, 14:30-15:30 UTC
> Place: https://bluejeans.com/4113567798/
>
> Full minutes: https://etherpad.openstack.org/p/tripleo-ci-squad-meeting
>
> = CI Promotion problems =
>
> The last promoted DLRN hash is from 21st of May, so now it's 12 day old.
> This is mostly due to not being able to thoroughly gate everything that
> consists of TripleO and we're right in the middle of the cycle where most
> work happens and a lot of code gets merged into every project.
>
> However we should still try our best to improve the situation. If you're
> in any position to help solve our blocker problems (the bugs are announced
> on #tripleo regularly), please lend a hand!
>
> = Smaller topics =
>
> * We also had a couple of issues due to trying to bump Ansible from 2.2 to
> version 2.3 in Quickstart. This uncovered a couple of gaps in our gating,
> and we decided to revert until we fix them.
>
> * We're on track with transitioning some OVB jobs to RDO Cloud, now we
> need to create our infrastructure there and add the cloud definition to
> openstack-infra/project-config.
>
> * We have RDO containers built on the CentOS CI system[1]. We should
> eventually integrate them into the promotion pipeline. Maybe use them as
> the basis for upstream CI runs eventually?
>

Thanks for sending this out Attila..

So after some discussion with David and others I wanted to spell out a bit
of a nuance that may cause this to take a little bit more time and effort.

The original plan was to build and test containers as part of the rdo
master pipeline [1].  We were on track to complete this work in the next
couple days.
However what we realized was that rdo has to feed tripleo container builds
for tripleo promotions, and tripleo promotions are always done on a random
new delorean hash.
There is no way to determine which hash tripleo will pick up, and therefore
no way to ensure the containers and rpms are at the exact same versions.
It's critical that rpms and containers are built using the exact same repos
afaik.

It is also good form and upstream policy for the tools, jobs and build
artifacts to be created upstream.

So the new plan is to build and test containers in the tripleo periodic
jobs that are used for the tripleo promotions.
When the containers pass a build they will be uploaded to the container
registry in rdo with a tag, e.g. current-tripleo.

The main point of this email is to levelset expectations that it will take
a little more time to get this done upstream.
I am very open to hearing suggestions, comments and critques of the new
high level plan.

Thank you!

[1]
https://ci.centos.org/view/rdo/view/promotion-pipeline/job/rdo_trunk-promote-master-current-tripleo/



>
> * Our periodic tempest jobs are getting good results on both Ocata and
> Master, Arx keeps ironing out the remaining failures. See the current
> status here: [2].
>
> * The featureset discussion is coming to an end, we have a good idea how
> what should go in which config files, now the cores should document that to
> help contributors make the right calls when creating new config files or
> modifying existing ones.
>
> Thank you for reading the summary. Have a great weekend!
>
> Best regards,
> Attila
>
> [1] https://ci.centos.org/job/rdo-tripleo-containers-build/
> [2] http://status.openstack.org/openstack-health/#/g/project/ope
> nstack-infra~2Ftripleo-ci?searchJob=
>
> __________________________________________________________________________
> 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
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstack.org/pipermail/openstack-dev/attachments/20170602/b7a89c5d/attachment.html>