This week, a good part of the Fedora Infrastructure team as well as some members from the CentOS Infrastructure team met up in Frederisksburg (Virginia, USA) for a few days of hacking together.
The least I can say is that the week has been really productive.
It started with an effort on trying to refresh the documentation for packagers, trying to update it for the new workflow put in place since pkgdb got deprecated and pagure set on the top of dist-git. You can find our initial effort on the pagure project package-maintainer-docs. It is far from finished but gives us a basis to build on. Hopefully this will help make the documentation easier and simpler to maintain in the future with a single place to edit and a pull-request workflow to propose changes.
We announced and decommissioned two systems. One of them was broken for over a year and the other never got an UI to make it useful. So, we said good bye to darkserver and summershum. Similarly, we officially move the jenkins instance running in the Fedora Infrastructure cloud, over to ci.centos.org where there is a team that has experience and is willing to maintain it.
In addition to this, we have been working on showing in bodhi if an update has been waived or not. This will allow packager to know if their waivers has been recorded and help admin tracking how an update got through. We also changed the message shown to the user to include the tests that are missing, making it clearer what is blocking. Randy also worked on fixing the "Waive" button in bodhi allowing to more easily waive test results in the UI.
Over the week, we also tie some loose ends on the upcoming pagure release (4.0). There still a little work left to do but hopefully we will be able to make the release next week.
This is the work I was the most involved, but we also started looking into deploying AWX in our infrastructure.
We started working on gating rawhide via bodhi. There is still some work upcoming but we have at least a somewhat clear path forward for this.
Finally, we have discussed and started planning to move more of our apps into the openshift instance we run. This may help reducing the load on our sysadmins as well as offer new possibilities of workflow (continuous delivery on commit or auto-deploy on build for examples).
As you can see this was quite a productive week. Many items that were on our todo got crossed and we have some interesting milestones in our roadmap.
I would like to thanks Paul Frields Red Hat and the OSAS team in Red Hat for organizing, supporting this event and allowing us to have a productive and nice time here this week.