CPE Weekly 2020-04-04

Saturday, 4, April 2020 Rich Bowen Uncategorized No Comments


The Community Platform Engineering group is the Red Hat team combining
IT and release engineering from Fedora and CentOS. Check out our teams
info here https://docs.fedoraproject.org/en-US/cpe/

GitForge Updates

*There has been a lot of discussion this week on the devel and infra
lists about the decision to move to GitLab in the near future.
Firstly, let us apologize again to the communities for our drop in
communication between the requirement collecting phase and the
decision making phase. As we have said before, it was in no way, shape
or form an intentional lapse of communications. However we do
recognize that it was still nonetheless a decision that was not made
in public, and for that we can only now offer our apologies for this
mistake and learn a hard lesson from it.

We do want to let you know that we deeply appreciate the requirements
you have given us and would like to ask you to continue engaging with
us while we are moving through our next steps with GitLab.
While the discussions on the lists are deeply emotional, they are
still incredibly valuable to us to truly comprehend the importance of
our next steps in ensuring we make the right choices in the coming

Now more than ever, your guidance is needed to make sure we achieve
the best possible result for you and our team from this decision.
CPE management and I, our team's product owner, are also actively
engaging with the Fedora Council and soon the CentOS Board to make
sure that ALL of the developments and progress between us and GitLab
are publicly available.

We have a long way to go in this process and your feedback on our
progress will be vital to make sure we remain on course.

We hope in time you can understand our decision was made in good
intent for the betterment of both our team and the communities we
serve, and we hope to still be able to rely on you all as peers and
friends for feedback and guidance during this journey.*


* CentOS CI is stable with 0 downtime!

* RHEL 7.8 released this week too!

CentOS Stream

* qt5, go-toolest (module), container tools (module) are updated to
their 8.2 versions

* We are also looking at easier automation of koji tags

Fedora Updates

* Final Freeze starts 7th April 2020 @ 1400 UTC

* Pagure 5.9.1 release pushed to both staging and pagure.io

* the-new-hotness configuration was updated

* Michal Konečný has been working on mapping the fedora infrastructure
applications, his project, (which sounds really cool and useful!) can
be found here https://github.com/Zlopez/fedora-infra-map

Data Centre Move

* Please note Communishift will be down from 13th April - 8th May to
facilitate the first shipment wave of our datacenter

* We are also still on track to switch to a reduced Fedora offering
from 25th May until est. 1st July\*.

* For a list of services we are planning to have available during this
window, please see mail thread in archive

* We will not have staging available so we will not have capacity to
review or deploy new or upgraded features and applications during this

* As always, please view our public schedule here for more a more
detailed overview

* We found a password, we do not know whose it is, but we have turned
it into the lost and found.

AAA Replacement

* First development phase complete & the team worked through 57 tickets in total

* The codebase was sent to our team first for demo and we will be
using feedback to develop the portal further

* During phase two we would like to change some codebases in existing
apps, and write documentation on how to upgrade applications to
redirect to the new API

* We would like to roll this request for feedback out to some
community maintainers during this phase too for another iteration on
the service and documentation

* Our work is publicly tracked here
https://github.com/orgs/fedora-infra/projects/6 so please stop by and
check out the progress we are making, and what we are looking at
working on next


* Monitor-gating is still running in production and giving us some
data about the health of the packager workflow:

* For example, these are the statistics between Monday and Wednesday:
39 messages retrieved
prod.monitor-gating.multi-build.end.failed -- 7
prod.monitor-gating.multi-build.end.succeeded -- 2
prod.monitor-gating.multi-build.start -- 10
prod.monitor-gating.single-build.end.failed -- 3
prod.monitor-gating.single-build.end.succeeded -- 7
prod.monitor-gating.single-build.start -- 10

* rpmautospec 0.0.1 through 0.0.10 have been released and deployed in staging

* We got two builds to go through fine, from the same commit,
getting two different NEVR and an auto-generated changelog

* However, for this to happen, we had to tweak a couple of things
on the builder which is not really ideal/acceptable, so we moved a
part of the processing inside the chroot where the SRPM will be built,
which solved the main issues we faced

* To be done: Make tagging (latest) existing builds a separate
operation executed outside the build root, to avoid having to talk to
Koji inside it.

* In other words: Stay put, we're getting there!

Sustaining Team

* Mbbox

* The team have made some progress on koji-builder CRD

* The team are also working on Bodhi 5.3 with few improvements and bug
fixes from 5.2.

* The team had a reboot and Update cycle this week.

* The team are also discussing ideas around a releng bot to help
process the tickets that require manual interventions.

* They also add more updates to compose-tracker

* And started playing with odcs on new

As always, feedback is welcome, and we will continue to look at ways
to improve the delivery and readability of this weekly report.

Have a great weekend!


Source: https://hackmd.io/gSci385uRoeNVEuoNUS1pg?view

Leave a Reply

Your email address will not be published. Required fields are marked *