I think many of us in the Linux community have been observing the changes happening due to Red Hat’s decisions as it impacts the wider Linux community that has been unfolding in the last few weeks. I’ve been thinking to myself how these companies lose touch with their base due to the profit motive over the thing that drew us to the company to begin with. Red Hat was a proponent and exponent of open source in my view, but their recent ownership and decision making is putting a very large warning sign on similar open source companies for advocates of open source like me.
I recently received my Framework laptop and was happy with Fedora until I really got to thinking about the implications of Red Hat’s behavior towards open source. Yesterday I decided to switch off Fedora due to the relationship Red Hat has in influencing the Fedora project.
Personally I do not have a direct issue with Canonical that would prevent me from using Ubuntu at this time, however this action on Red Hat’s part made me think critically about what open source projects, especially those owned by corporations, to support for my personal use. I know that for Framework as a company it must be extremely difficult to balance efforts to provide customer support the wide range of operating system distributions and that you can’t support every edge case. But I had a few questions I was hoping that could be answered:
Can you talk about the difficulties of supporting a distribution like Debian or the variety of other unsupported and community supported distros?
Is there any way for us as active supporters of both Framework and open source to make wider OS support a reality?
Can you discuss what pieces are necessary for Framework to achieve official support for an OS and whether there are efforts to shorten that distance?
I think Framework is in a unique position where many of us in the open source community would like to see a hardware company dedicated to open source succeed. If there is a way to create an operating system support scaffolding, or a (no pun intended) framework for the community to assist in accelerating operating system support it seems intuitive that it would go a long way in widening your customer base and I for one would like to contribute to that effort. I think many of us would like to see our favorite OS work on the Framework laptop and would contribute to the effort if there were a clear way to do so.
We don’t have hard and fast rules for this, but the primary factor is that we’re a small team, so we prioritize providing solid support for a smaller number of distros over spreading more thinly over a broader set.
Currently we prioritize Ubuntu and Fedora because that is what we’re seeing the highest usage of in the community. We run surveys regularly with both purchasers of Framework Laptops and people on our newsletter who haven’t purchased to understand trends. So far, no single Linux distro is anywhere near a majority preference, but Ubuntu and Fedora are consistently the top two. If that changes, that would certainly be a strong indicator to us that we need to reevaluate.
Look around the forum and you’ll find you can run far more than just Ubuntu or Fedora on a Framework laptop! Outside of those two, users help each other out. It’s usually not an issue.
Well, Debian 12 has a newer kernel then the supported Linux distributions (Fedora 37 and Ubuntu 22.04). It also now includes non-free firmware on the installation media.
But as I said - try it and if you have issues come back here. I don’t believe there’s an unofficial support thread for Debian. That may be good (no issues?) or it may not, but there are lots of helpful Linux people here. And although official Framework Linux support (Matt Harley and Loell_Framework) are under no obligation to contribute, they often do.
nrp has shared the Framework perspective. I will now share the perspective from the Linux support team and myself personally.
We focus on Ubuntu and Fedora because they have a solid balance between stability and having access to strong community support. We actively ask users to utilize whatever distro they like, but as outlined here, we will ask users to use officially supported distros for troubleshooting. Community supported distros listed are also tested, but we ask users to do hardware troubleshooting on officially supported distros. This is due to available hours available for our team to test various distros.
As the Linux support team grows, we will be testing previously unvetted distros for community support. We will not, ever, be able to provide official support for rolling releases as they can work great one day, then change overnight. Debian however is a candidate that we would love to bring into supported territory at a later date.
Remember, supported vs unsupported vs community supported only means we’ve can state that we test again and have worked to ensure a solid user expwerience.
We do have folks running Void, Arch and Gentoo all day every day without many issues, but it did take some user configuration.
All of this is the exact opposite of Red Hat preparing to make some new base
for RHEL. Additionally, this model provides a clean path for > Red-Hat-opinionated decisions to differ from those we make from Fedora. Take
BTRFS as an example. Or, the increase in CPU baseline.
Fedora remains a strong distro for Framework and we are thrilled for their continued support from their team and help - they do this as good community citizens.
As nrp points out, we continue to take a pulse on things to see what the latest flavor is and which users prefer. But at this time, we’re focused on the elements above and Ubuntu and Fedora are it at this time.
As the post from then-Fedora lead Matthew Miller makes clear, RH is a sponsor (well, let’s face it: the main sponsor) of the Fedora Project.
That means that - quite apart from the separate question of the extent to which RH influences decisions at Fedora - RH is providing financial support (presumably in the form of infrastructure, employees tasked with working on Fedora, etc.)
RH will only continue to provide such support as long as RH perceives such support to be consistent with its interests. And nobody outside of RH management can say with any certainty how RH will continue to perceive its interests.
Given RH’s actions thus far, would you have confidence in RH’s continued support of the Fedora Project?
And if RH does cease to be a major sponsor, who will step up to fill the gap? I’m not sure a Fedora Project under (say) Amazon stewardship will necessarily be better for the community.
Correction: an earlier version of this post incorrectly stated that RH had laid off Matthew Miller. The person I was thinking of was Ben Cotton, who was RH’s Fedora Program Manager (see: Fedora Program Manager Laid Off As Part Of Red Hat Cuts - Phoronix). I regret the error.
I haven’t found any mentioning on this in any news site or his mastodon account. He’s recently talked about Fedora on mastodon so it doesn’t seem that he left from what I can tell.
Hi Matt, thanks for the reply and expansion there. At this point I see where you’re coming from and all of that makes sense but I think it will help to make some distinctions so we can talk about the challenges a bit differently.
I see that you’re carefully defending the choice to support the OSes that you as a team are officially supporting, but I understand the reasoning why you are supporting the ones that you do already. I think support needs to be more clearly defined for the sake of our discussion. Please know that I’m not here to criticize the support offered but looking for a proactive path forward towards getting more OSes available and working with Framework hardware more broadly.
Support on the business side is definitely important in terms of SLA’s, official responses to queries, etc. and the Framework team is clearly limited in its ability to provide customer support for every distro under the sun, and needs to limit its scope. We completely agree on that.
What I’m referring to is slightly different but of course it intersects with your concerns as a support team. The support I am referring to is for example providing a scaffolding for open source OS projects to adequately integrate and work with Framework hardware, enabling the widest amount of OS projects to work. Now I don’t know what all is involved in that with the work that you do directly with Fedora or Ubuntu for example but those are the types of things I was looking to get clarity on with my questions.
Put another way, if there were a roadmap that included each of the pieces of hardware as it relates to an arbitrary OS, we on the open source community side would have better direction for getting Framework hardware to work on Debian for example. I imagine a simple example would be ensuring there is a repeatable series of tests that would allow us to identify that each of the pieces of hardware works as expected for a given OS. In terms of the customer support perspective, I don’t have an expectation that Framework support would be required to fix issues here, just that we have a clear path to direct the particular OS project’s open source contributions - what works, what doesn’t. I imagine after a certain level of acceptable compatibility (‘support’), you as a customer support team at Framework would feel more comfortable providing the limited support as it relates to answering customer questions, or directing it back to the open source community.
Part of the problem on my end is that I do not know what all is required in that space, which is why I was asking those clarifying questions to begin with. What work, for example, is required on your part to ensure that Framework hardware works with Fedora? And how can that be extrapolated into a system to get more OSes better Framework coverage with open source community effort?
Just want to chime in here: I am running Void Linux (rolling release distro) since 6 months now on the 12th gen intel Framework Laptop 13 and while I had issues in the beginning both with a broken NVME and my HDMI expansion card in combination with my monitor, it is all working now after some kernel and driver updates and a few configuration changes.
While I have to admit that the initial experience could’ve been a lot smoother, I dont regret to have bought the Framework laptop and have put a rolling release distro on it:
I have very capable hardware with hardware acceleration
I have an eGPU setup that allows me to play games (that 1 day a month I actually have time for that)
I get the newest drivers and fixes much earlier than most other users
I can test new things and report issues upstream
It just works now. It’s stable, it’s fast and smooth. Granted, I don’t bring my computer to it’s limits most days, but I do compile or transcode bigger stuff from time to time (1h+ using all cores) and so I can say that it’s great to have power, when needed.
Of course, everything has its upsides and downsides. Using a stable release vs rolling one is one important decision everybody has to make. I took the risk and live at the edge, which is why I took precaution and have an advanced backup system in place.
I really don’t get the intent of this thread, or even what it is based on. Centos Stream delivering nothing different from what Centos did except now instead of being an awkward post production afterthought it is the final pre branded format of the Redhat edition. All Redhat did was stop conveniently packaging and presenting clones with the ability to say that they are exactly what RHEL is and then go out and provide support at half the price while doing none of the work.
If you think Centos was ever Enterprise grade you need to go back and really research it. Centos recevied security updates late, and had no true version control. I could not freeze an install at Centos 7.2 even thouhg Centos was version 7.5. They tried to do that and failed because it was entirely too much work. RHEL you can freeze it at for instance 7.1 and ride that until EOL if need be. Almalinux did not do this, Rockylinux does not do this, Centos did not do this, Debian does not do this, Ubuntu does not do this, the only distro that comes close to doing this is SLES.
So the entire bruhaha over Redhat making 1) a smart business decision, 2) providing essentially the same product Centos was in Centos Stream, and 3) forcing anyone trying to claim that they are ABI compatible and can provide true Enterprise support to actually build their own infra and pay the developers needed to make that distro. They can now build it off Centos Stream but will actually have to do the work and yes contribute upstream, something they did not do previously, or they can wither away.
As to Fedora if this has anything to do with the vigorous discussion surrounding a proposal to add telemetry, that is just silly. It is a transparent public process, with an open and passionate discussion. The proposer is going to have to rewrite it, and compromise to have any chance of getting it put through.
As to Redhat’s continued support…Fedora->Centos Stream->Redhat that is the chain of development, and frankly I dont see that changing anytime soon. The tech industry took some dings and Redhat had to cut some positions, it is not like they did not cut positions within Redhat as well. The system they have in place will actually work better now that Centos Stream is midstream versus the previous Fedora->Redhat->Centos.
Anyway done reacting. Fedora ain’t going anywhere and Redhat is still one of the largest contributors to linux up and down the stack. Framework has a relationship with Fedora, and that makes a big difference.
Alright, think we covered all the bases and all concerns and perspectives shared. Going to button this up. Appreciate the feedback everyone, but I am going to wrap this up so we can focus on other topics.