Storytelling is one of the most powerful means to influence, teach, and inspire the people around us. To celebrate OpenStack’s 10th anniversary, we are spotlighting stories from the individuals in various roles from the community who have helped to make OpenStack and the global Open Infrastructure community successful.
Here, we’re talking to Thomas Goirand from Infomaniak. He tells the community about how he got started with OpenStack and his favorite memory from the last 10 years of OpenStack
How did you get started working with OpenStack and what are you doing now?
I was running my own hosting company: GPLHost (which still exists but I am not the person running it since 2014). GPLHost was a moderately successful VPS business. Though a bit after public cloud was a thing, I quickly understood that the cloud would be the future of server hosting. However, there wasn’t any viable free software solution … until the OpenStack project started. As I’ve always run Debian and became a Debian Developer in June 2010 (with upload rights in the archive), it was obvious I had to upload OpenStack in Debian before I could get going and use it for GPLHost. I never used OpenStack at GPLHost. Instead, maintaining OpenStack in Debian took more and more of my time, to the point where it became my full-time activity. I was first contracted by eNovance (later bought by RedHat), and then by Mirantis.
These days, I’m maintaining my own OpenStack installer, which is also fully in Debian, and which I also presented at the 2020 Open Infrastructure summit.
I’m also maintaining multiple deployments of OpenStack in production for my employer, Infomaniak: 4 moderately large swift cluster, and 3 full-features (with many types of nodes: compute, volume, network, ceph, etc.) OpenStack clouds. These are very demanding, and my team cannot afford any failure, which is a great live test for both the Debian packages and my OpenStack-cluster-installer.
What is your favorite memory from the last 10 years of OpenStack?
The summits of the early days of OpenStack were memorable, with a week full of evening parties (my first one was the one in Portland, and went to all of them until the Barcelona Summit). There were so many parties that in a single evening, there could be 3 of them running at once, with all the companies spending a lot of money on them. I definitively miss this, and the fun time socializing with everyone. Virtual events are nice, but they will never replace in-person meeting. I really hope the COVID time will be over soon, so we can see each other again.
How did you contribute to the OpenStack community?
In Debian, Debian Developers can only upload new stuff to Debian Unstable (AKA Sid). This means that I’ve been constantly maintaining OpenStack in Debian Unstable for the last 10 years. And Debian Unstable, as its name describes, is a moving target (it doesn’t mean it’s full of bugs, but that it’s constantly changing). This means we always get components updated earlier than in any other distribution. As a result, maintaining OpenStack in Debian is a challenge, because we get to test OpenStack with the very latest version of everything before anyone else. For example, we get the latest version of the Python interpreter first or the latest version of Django, SQLAlchemy, etc.
As a consequence, I’ve been doing a lot of bug reports on upstream OpenStack bug tracker, in the hope that someone from each individual project can fix it. And fixing many bugs myself, when I could (when time permits AND when it is easy enough). Obviously, a single person can’t be in charge of fixing all of OpenStack to make it work with a newer version of Python, but I am happy that I could do at least what I could, and bring even a small contribution to OpenStack, together with folks from other distributions (like Canonical, to name one: hi Corey, James, and the others!).
I also contributed many patches to the puppet-OpenStack project, to make it work for Debian, as my installer uses it to deploy OpenStack.
All together, Stackalytics.com shows a total of 113 patches so far, on many different projects. This is probably not a lot, but for me, it represents a lot. I wish I could do more, and contribute some features, but maintaining all of the (currently) 541 Debian packages of the OpenStack Debian team, plus my installer, plus maintaining the deployments in production for my employer Infomaniak are keeping me really busy.
What advice do you have for the Stacker community and other growing open source communities based on your experience with OpenStack?
It’d be great if all the stackers working on individual projects (like core reviewers) tried to keep in mind that contributors may only have a limited amount of time to spend on each individual patch they propose. Sometimes, asking for more tests and nit-picking on every aspect of a patch can be really counter-productive. I gave up on many patches because of it. If possible, core reviewers should take over the patches and enrich them directly, if they think it needs more tests, for example, rather than asking contributors to do it.
If I were to ask you in 2030, what do you think the OpenStack update will be?
In 2030, hopefully, I’ll still be there maintaining OpenStack in Debian and packaging my 40th OpenStack release. Many people will be using my deployment tool, using Puppet 9, Python 3.14, running on Debian 17. 🙂
I also still have hope that one day, the OpenStack project will contain all of what’s needed to run a public cloud, including invoicing, billing, and payment solutions. If this was to happen, I’m sure this would give a huge boost to the project, with so many smaller hosting company being empowered to start an OpenStack business.
If you’re interested in being featured to tell your stories about OpenStack, please email [email protected].