As you may know, the Community Platform Engineering (CPE) team that helps manage both infrastructure for both Fedora and CentOS is trying to improve how the different initiatives or requests for changes we receive are planned and prioritized. This effort to improve has led us to look at planning our work in three-month windows. By planning, limiting, and focusing our work every three months, we can dedicate a substantial team around each initiative, driving it to completion before working on the next. Overall this helps us finish and deliver work faster.

The prioritization work involves the Fedora Council, the CentOS Board, a representative of the Red Hat Business Unit, and the CPE team for each three-month window. Initiatives are considered against the team’s mission statement and weighed against each other to ensure that the most valuable initiatives are properly prioritized. This way of working  is quite new for us so any feedback on the overall process is greatly appreciated.

What’s coming up next (April – June 2020*)?

CentOS Stream (Phase 2)

In the next three months the team will be focusing on making CentOS Stream easier to consume and release. This means investing in automation of the packaging and compose processes. 

More information about this effort on our Taiga board.

Fedora Account System (FAS) Replacement (Phase 2)

This project is the second phase of the work that was started at the beginning of this year to replace FAS (which you may have heard referred to as “AAA”: Authentication, Authorization, Accounting). The first three months of the year were focused on building the web portal allowing users to register and manage their account. This phase will focus on the integration with other applications. Most of the work will be done in fasjson and fasjson-client

More information about this effort on our Taiga board.

Fedora Data Center Move

This is a big effort that in fact has already started and will continue for most of the next three months (current estimated end date for this is mid-July). The work will involve deploying critical infrastructure in the new data center so that we can continue to build Fedora while the hardware is shipped from the old data center. Then as the hardware arrives at the new location we will be working on redeploying the services and adding more capacity to that infrastructure.

More information about this effort on our Taiga board.

DNF Counting

This project will update and improve the current solution that gathers anonymous statistics about the number of installed Fedora systems by making use of the new DNF “countme” flag. This will help the project better understand how Fedora’s various offerings are used in the world, and give us better insight into the real-world lifecycle of our releases while taking great care to preserve our user’s privacy. 

More information about this effort on our Taiga board.

CentOS CI Infrastructure Phase 1

This work will allow us to update the current infrastructure used by ci.centos.org from an OpenShift 3.x version to an OpenShift 4.x version. The goal of this work is also to build up more administration knowledge of OpenShift in the team as well as improving the performance and reducing the maintenance effort needed to run the current infrastructure.

More information about this effort on our Taiga board.

Mbbox

The goal of mbbox is to make the current solution used to build rpms and modules for CentOS easier to maintain and update. The core of the work is to deploy koji and MBS using Kubernetes operators in order to manage the upgrade and deployment of new versions.

More information about this effort on our Taiga board.

Getting updates

You will find regular updates on our Taiga board relating to teams progress on each initiative. If you have any questions or if you want to know more about any of these initiatives or would like to contribute, please join our #redhat-cpe channel on Freenode IRC or visit the Taiga links above for more information.

In addition, a dedicated sub-team, the sustaining team, will continue to service the lights-on operation work for CentOS & Fedora.

We recognize that this email/blog post is late, our 3 months window has started for a month already, as we said this is still a new process for us and we’re still in the adjustment phase. Hopefully we’ll do better in June!