CentOS Connect is a free mini-conference focusing on CentOS Stream, the CentOS SIGs, and the entire Enterprise Linux ecosystem. CentOS Connect at FOSDEM happens February 3, 2023, the day before FOSDEM.
Connect in person
DoubleTree Brussels City Center
Pagoda room, eighth floor
Rue Gineste 3, 1210 Bruxelles, Belgium
Connect online
Live streamed on YouTube
Virtual hallway track on Google Meet
Questions relayed to speakers
Code of Conduct
CentOS is committed to providing an inclusive and harassment-free experience for participants at all of our events. All participants, whether in-person or virtual, are expected to follow the CentOS Code of Conduct. To report violations or for any concerns, contact either the Community Architect Shaun McCance or the Board President Amy Marrich.
Badge Stickers
Our badges have space for stickers so people can show what projects they work on. If your project is in the CentOS ecosystem, feel free to bring some stickers and we'll put them out. The stickers must fit inside a one-inch square. We used these die-cut sheet labels at 1"x1" from Sticker Mule, but you can use any vendor, as long as the stickers fit.
Schedule
All times are UTC+1, Brussels local time.
Going to FOSDEM? Come see us Sunday in the Distributions Devroom.
Speakers
Adrian Reber
Adrian is a Senior Principal Software Engineer at Red Hat and is migrating processes at least since 2010. He started to migrate processes in a high-performance computing environment and at some point, he migrated so many processes that he got a PhD for that. Most of the time he is now migrating containers but occasionally he still migrates single processes.
Davide Cavalca
Davide is one of the founding members of the Hyperscale SIG, which he currently co-chairs, and also serves as a director on the CentOS Board. In his day job, Davide is a Production Engineer on the Linux Userspace team at Meta, which is responsible for the CentOS deployment on the production fleet.
Adam Samalik
Adam is a principal software engineer with Red Hat mostly contributing to Fedora ELN, CentOS Stream, and RHEL.
Peter Georg
Works for the Physics Department at the University of Regensburg, Germany. Chair of the CentOS Kmods SIG since June 2021.
Fernando Fernandez Mancera
Fernando is a free software enthusiast focused on computer networking. He is an active contributor of several projects like the Netfilter subsystem, NetworkManager and Nmstate. Currently he works as a Senior Software Engineer at Red Hat focused on Network Management tools like Nispor, Nmstate or NetworkManager.
In addition, Fernando contributed a lot to the SUGUS GNU/Linux user group in Sevilla, Spain.
Neil Hanlon
Neil Hanlon is a Linux developer with over ten years of experience. He is the infrastructure team lead for Rocky Linux and a member of the Release Engineering team, where he works on the development and maintenance of Peridot, an open source build system. Neil has a deep understanding of Enterprise Linux, networking, systems administration, and architecture. In his free time, he contributes to open source projects such as OpenStack-Ansible and engages with the wider Linux community and is passionate about sharing his knowledge and expertise with others.
Aaron Williams
Aaron is the Community Director for the Ampere Developer Community. He started his career as a Java developer and began his developer advocacy at SAP as the the Global Director of SAP's internal maker and community spaces program called the d-shop. And has been a developer advocate/community manager for projects in the ASF and LF.
Sessions
Offered CentOS Infra services for SIGs
In this talk, we'll do a quick recap about which kind of services (from git hosting to building and cdn delivery, as well as CI testing) the CentOS Infra team is offering and maintaining for the Special Interest Groups
MirrorManager and CentOS Stream 9
Fedora relies on MirrorManager since 2008 and with CentOS Stream 9 CentOS mirrors are now also managed by the MirrorManager instance.
For the CentOS community I want to use this session to give an overview how MirrorManager works. I want to give an introduction about all the different parts that are necessary to make MirrorManager work as well as how the Fedora instance is set up. I also want to highlight how MirrorManager differs from the traditional CentOS mirror infrastructure.
In addition to the introduction for the CentOS community I also want to present what has changed in the last 6 years and how and why we rewrote core components in Rust.
One year on: Experiences using ebranch to bring over Fedora packages to EPEL
At this event last year, I described a WIP tool called ebranch (https://pagure.io/epel/ebranch) that is meant to simplify the workflow of branching a specific package for an EPEL release, together with all the missing dependencies needed to build it.
One year on, this tool has been used for bringing over various sets of new packages to EPEL, in different programming language stacks (from Python to Perl to Rust); this talk discusses the current state of the tool, how features are added to address specific needs, the experiences gained in writing and using the tool, and the pros and cons of how different language stacks are managed in Fedora when it comes to branching to EPEL.
Hyperscale SIG update
Update on what the Hyperscale SIG has been working on, what deliverables are available and how to use them, and what's coming up next.
Introducing CentOS Stream CoreOS and OKD Streams
Christian Glombek and Alessandro Di Stefano
CentOS Stream CoreOS (SCOS) is a Linux distribution built from CentOS Stream RPM packages, and focused on running container-based workloads with Kubernetes. It is part of the SCOS Stream of OKD, the Kubernetes community distribution of OpenShift, co-maintained by the CentOS Cloud SIG and the OKD Working Group.
In this presentation, we'll present the technologies and methodologies driving the CentOS Stream CoreOS (SCOS) release engineering, and the Cloud-Native architecture we leverage to package the operating system that runs Kubernetes/OKD.
We'll show how this framework, powered by Tekton pipelines and operated via GitOps, can enable, thanks to rpm-ostree, the CoreOS Assembler and the Layering model, delivery scenarios for different OSes beyond the Cloud-Native ones: IoT, multimedia, automotive, thin-client-based environments. Users can derive their own purpose-driven variants by maintaining a common multi-arch base OS, distributed as a bootable Open Container Image (OCI).
CentOS Stream: RHEL development in public
CentOS Stream is where RHEL development happens in public. You can preview content coming to RHEL, test your things on top of it, and even participate! We'll show you how it works, highlight the key differences between Fedora ELN, CentOS Stream and RHEL, and see where it's all happening.
Kmods SIG Update
Update on what the Kmods SIG has been working on with a particular emphasis on automation of rebuilding kABI tracking kernel modules if required.
Network management in Enterprise Linux: present and future
The talk will explore the current state of network management in Enterprise Linux systems and discuss potential future developments in the field. The presentation will cover topics such as network configuration and troubleshooting, with a focus on the NetworkManager and Nmstate tools. The aim of the talk is to provide a comprehensive overview of network management in Enterprise Linux and to discuss the challenges and opportunities that lie ahead for enterprise Linux users.
Introduction to Rocky Linux and Peridot: Maintaining a downstream fork of Red Hat Enterprise Linux
Learn about Peridot, a new open source build system created and used by Rocky Linux to simplify the process of maintaining a downstream fork of Red Hat Enterprise Linux. Discover how Peridot can be used to patch and rebuild RPMs, modify upstream RPMs, or package your own software, and how Rocky Linux uses it to manage the rebuilding of all packages in Enterprise Linux and help drive upstream contributions while enabling Rocky's unique special interests. This presentation is ideal for users of RHEL-like operating systems who want to improve their Linux deployment and management processes. Join us to learn about the benefits of Peridot and Rocky Linux and how you can use them to optimize your Linux system and streamline your development workflow.
From code to cloud - the journey of Openstack package
OpenStack is a huge cloud computing project, which does not provide packaging for platforms – RDO does it for rpm-based Linux distros. Delivering packages for such project with new release every half a year is a challenging task. In this talk I’d like to bring closer to audience our continous-delivery approach to package delivery – starting from creating and adding new packages, through updates and managing all packages already delivered, share our practice in automation and tips how to not get drown in dependencies of dependencies. The journey will also have quick stop in building tools we are using in RDO and continuous integration to ensure stability and compatibility, to finally reach the point of having new OpenStack release. This talk is intended not only for those, whose daily duties are connected with cloud or continous-delivery technologies, but also for anyone interested in topic of delivering packages at great scale in open source cloud project, or would like to contribute to RDO.
A year in review 2023 - CentOS Automotive SIG
A review of what's going on in our CentOS Automotive SIG, our AutoSD image, how to run an AutoSD VM to try AutoSD, PREEMPT_RT kernel. Similar in ilk to "Fedora: The Vehicle for Automotive Linux" presented by Stephen Smoogen and Allison King at "Nest with Fedora 2022".
Running Cloud Native Applications on CentOS on a Cloud Native Processor; Setting up and running a Mastodon Server on Arm servers in the cloud
In recent months, Mastodon has garnered a lot of attention, and seen a huge influx of new users. Mastodon is a social network built on ActivityPub, a protocol for federated social media. In early December, the network broke 8 million users, and had 2.5M active daily users in one week.
That influx of new users and interest has led to many new Mastodon instances being added, some with a very broad appeal, and others targeting smaller groups and niche interests. It has also led to some of the more popular instances of Mastodon struggling to scale with the new demand.
In this talk, we will walk you through how Mastodon’s federated architecture is designed for the cloud and how easy Mastodon is to set up and run on a CentOS instance on AArch64 cloud instances for free. And since Mastodon’s backend is written in Ruby on Rails, using Redis and PostgreSQL, we will show how easily they run on an AArch64 processor.
In addition, we will look at how well the Ampere Altra processor handles cloud native workloads on CentOS. We will show you not only how to run Mastodon on AArch64, but how to do it for free, without having to worry about getting a large cloud bill. Recent events at Twitter gave us the fun idea of how to combine all of this: create and run a Mastodon server on Oracle Cloud’s (OCI) Always Free tier using Ampere A1 and CentOS.
We will also talk about some of the scaling issues that Mastodon runs into, and how Ampere cores designed for cloud native workloads like Mastodon are uniquely able to give you predictable throughput and scaling as your server grows in popularity. All while doing this on a processor that is more efficient (i.e. greener) than other processors out there.