LTS kernels are no longer supported for 6 years because no one used them | Hacker News Reader

LTS kernels are no longer supported for 6 years because no one used them

lunduke.locals.com

86 pointsjosteink62 comments

Did anyone think to actually ask the developer who is maintaining the LTS kernel versions why he made that change (back in February?), i.e. me?

{sigh}

No, I guess that would take too much effort, and wouldn't result in such a click-bait headline "LTS kernels are no longer supported for 6 years because it turns out no one used them." doesn't have that same fun sound...

It's an article from Bryan Lunduke. What can be expected other than controversial and click bait articles?

It was extremely frustrating to see several articles saying that make the rounds on HN...

> LTS kernels are no longer supported for 6 years because it turns out no one used them

I assumed companies use a lot of LTS software for servers to avoid frequent upgrades that always carry some chance of breaking stuff. Could you please elaborate more or link me to some place where I can read more about this?

If for-profit companies need to use older kernels, they can hire engineers using the profits?

Yes, they can. But it's different from saying that no-one uses them.

I'm sure the nuance is not lost on the people that made the decision.

I'm just wondering if it is actually how you said, i.e. they don't want to provide free support to companies vs how he write, i.e. not enough users.

Don't distro kernels use LTS? Debian stable never changes it's kernel through its life cycle, and there is also a LTS of Debian

The one does not discount the other. In other words: disregarding the lts misunderstanding it still appears true that the Linux foundation is abandoning Linux kernel development. At least financially. Worth a look I guess.

Also: thank you for all your contributions to Linux for all these years!

The amount of resources and other stuff that the LF provides to the Linux kernel community has increased over the years, including last year. Just because new people are brought in with new projects (that the LF member companies want to host) does not mean that somehow less is being given to the kernel community at all. It is not a zero-sum game here at all, that's not how the LF works in any way.

Again, this would have been easy to verify if someone just asked us.

So to repeat, no "abandonment" is happening here at all, the opposite is happening, just like it has for the entirety of the LF's existence, support has grown every year.

you cant keep up with the fire-hose of nonsense that is the internet. Don't try, its such a waste of your skills and appreciated efforts.

If it gets enough attention just put a simple refutation or clarification on the foundations front page and move on.

> LTS kernels are no longer supported for 6 years because it turns out no one used them.

Looking quickly at Debian, old old stable (10) uses 4.19, old stable (11) 5.10, and stable (12) 6.1. All these versions are LTS kernels.

I don't think Debian uses LTS kernel branches as is, instead cherry picking patches, but it seems they do leverage LTS branches, at least to ease the back-porting I presume.

I'm probably missing something, and looking through the Debian mailing lists, I've not seen any discussion about the end of long term LTS kernels, but it seemed they did rely on it to a degree.

Maybe it was more of a nice to have, but from the outside (I'm neither a kernel or Debian maintainer), it leaves me a bit confused.

Well, we can at least change the title to that above. I've done so now.

> it turns out no one used them

Can you explain what this means? E.g. Debian very much sticks to LTS kernel versions in their stable releases. Were they not working with upstream, you? If they'd be doing their own maintenance, there's no need for them to stick to LTS versions, but clearly they didn't tell you. Any idea what's going on there?

Bryan Lunduke is the Russia Today or Newsmax of open source...

Might be worth looking at the Linux Foundation's stated mission and aims before fetching the pitchforks

https://www.linuxfoundation.org/about

The Linux Foundation Method

A proven and repeatable way to scale project communities via a comprehensive portfolio of support programs for aspiring industry leading projects:

Neutral home for code and collaboration - We aim to democratize code and scale adoption, for all projects.

Ecosystem curation and community building - We strive to create new technology categories by identifying trends, accelerating the growth of nascent technologies, and removing barriers to adoption.

Enterprise ready, the OSS way - We provide turnkey technology and support programs for developer enablement, business operations, training & certification, marketing and events, and membership development to help projects scale fast.

Project insights and management tools - We help projects streamline operations and boost community engagement with cloud-based, collaborative tooling, contributor and participation analytics, and infrastructure management.

Lunduke has been picking so many fights, I'm a bit lost as to whether I should pay attention or not. Sure, it doesn't seem great, but the Linux foundation putting money in other areas of the ecosystem is hardly a good point. Show us a real breakdown, instead of a gratuitous Linux kernel expenditure vs not-Linux kernel expenditure.

I'm not familiar with the author, but seems like they have quite identical post from 2021 as well: https://lunduke.substack.com/p/linux-foundation-spends-just-...

But the author and many others aren't wrong in highlighting this and have covered the LF issues for years.

Given that is from a guy that cares a lot about how the Linux Kernel project is funded, that is very concerning.

The first step in solving a problem is to identify that there is an inherent issue, which the author has done, but goes ignored by those who do not wish to care about the project.

Last time I saw a post of his linked here, someone posted this link: https://old.reddit.com/r/linux/comments/muc18q/whats_the_dea...

I approve of this fight-picking. It’s what journalists are meant to do.

The Linux Foundatiom should an engineers-first organization.

The leadership team breakdown shows me it’s full of MBAs specializing in “on a computer”.

> The Linux Foundatiom should an engineers-first organization.

why?

what do you feel it's purpose is? do you think the people who give it gobs of money agree with you?

Do you see any value in MBAs that are able to 'sell' other MBAs on the importance and value of Linux and OSS?

Are you saying MBAs and sales only provide value if they’re in the position of leadership?

I agree, Journalists should pick fights, but I would prefer if they built up their arguments a bit better than this.

I am one who is not familiar with the author previously, and from what he brings up I am highly inclined to agree with him, but I have to agree I became more sceptical to his arguments when all he brings up to support it is one very sparing data point (that graph feels very misleading). Give me data on how the kernel support has changed over time. Give me data on how the foundations income has changed over time. Not showing me any of this makes me very sceptical. Would showing this data not support the authors point?

There's journalism that challenges the status quo, and there's muckraking, which is the degenerate form. This seems to be the degenerate form.

I remember when I discovered the Bad Voltage podcast he was a co host. I knew nothing about him at the time, but I remember thinking good Lord why is this guy allways a contrarian asshole. Since then he's left the podcast and it got so much better and I've learned who he is and decided I really don't like the guy.

He is a troll, picking fights about things he doesn't understand. This is not the first

This is weird. LF has spent less on Linux than everything else. LF has to deal with lots more than LF. Even based on the numbers this article mentions, LF spends close to a million a year on Linux. I don't understand what the aim is here.

Right? Linux is much more than just the kernel.

a million is what, 10 engineers full time?

more like 3-5 given people who can manage this kind of code base are of high value.

I would say even less because you have to take into account all the costs of hosting, dev machines, laptops, testing, travel etc. $1 million for such huge project is peanuts this days.

How many engineers do you think are actually managed by LF? From my understanding, the majority of engineers contributing to Linux are on the payroll of other companies like Google.

On the other hand, many companies have people on their payroll to work on Linux. I would argue Linux barely needs anything from the LF.

> many companies have people on their payroll to work on Linux

... for that company's own interest.

> I would argue Linux barely needs anything from the LF.

Linux? No. It's a computer program. Computer programs have no needs (yet).

Users, real natural persons, YES! Remeber when Linux couldn't play audio without hiccups? Remember Con Lolivas patches?

Have a look at linux kernel maintainers. How may systems in there are "supported" vs. just "maintained".

I think this is really just a naming problem. The Apache Software Foundation was founded for the Apache webserver, but today they have hundreds of projects unrelated to that. The same is true for the Linux Foundation.

Good point. In a sense, foundations have to appear much more interesting than they are nowadays. Developing an OS kernel or a web server isn't appealing to newcomers, so I guess there should be alternative pathways to that. There should still be alignment in behaviour to the identity the entity claims to be. Or we just could have the Whatever Foundation. So, are these other projects in any way related to the development of Linux kernel or not?

I would call this "foundation bloat".

I always find a kind of ironic that Linux Foundation is the home of Zephyr, an embedded OS that has nothing to do with Linux, other than being competition against Linux distributions for the embedded market.

It isn't only Zephyr, also being the home of several other projects completly unrelated to the Linux kernel.

> Zephyr, an embedded OS that has nothing to do with Linux, other than being competition against Linux distributions for the embedded market.

Zephr isn't "competition against Linux". It primarily targets microcontrollers that Linux would have no hope of running on (e.g. parts with <1MB of memory).

It certainly is, given that it isn't the only target, if you bother to actually watch their talks, specially its license makes it much more appeling to OEMs than a GPL tainted kernel (from their point of view).

Does anyone know if the Linux foundation makes their financial statements available?

My cursory glance turned up the slick annual presentation type report, but I am looking for the 3 basics, income statement, cash flow, and balance sheet, audited or unaudited.

An aside, I think Linux is a true modern marvel of humanity, it really is people at their best. It's a bummer when in-fighting and drama pop-up, but I guess it comes with the territory and its part of the sausage making.

you can get their 990 which has some details. Latest available is 2021 https://projects.propublica.org/nonprofits/organizations/460...

The 2022 report (which was not linke even though it was the source of this article) doesn't look so dramatic. The listed grievances look like small contributions alongside their main projects:

https://project.linuxfoundation.org/hubfs/LF%20Research/2022...

Sigh. Nobody cares about the upstream LTS kernel. All the LTS distros maintain their own downstream LTS kernel. Android too.

There is no reason to duplicate that work upstream and then nobody uses it anyways.

And by the way, LF founs the Civil infrastructure project which mantains LTS kernels for 10 years.[1]

[1] https://wiki.linuxfoundation.org/civilinfrastructureplatform...

> All the LTS distros maintain their own downstream LTS kernel.

Debian definitely uses the upstream LTS for their base for their kernel on stable. They do add their own patches, but I don't think it's fair to say that they duplicate work here.

I guess IBM was a major driver and sponsor for the Linux Foundation and had reason to push standardization and enterprise things such as LSB (alignment and extension effort for POSIX/SUS) but now owning RH has of course totally shifted the perspective. Also, commercial Linux seems more about a few cloud providers and tool vendors - standardization has no standing in that environment.

It's a real shame that Bryan so frequently leans on pissing people off to make a point. He has a wonderful voice, a keen, obvious and sustained interest in the community, but interacts with it like a shock-jock.

I guess it's working for him in the way he needs it to, but it feels like way too much drama.

On one hand, LF's website doesn't inspire confidence with its "Decentralized innovation" buzzword pitch. On the other hand, a bunch of other useful projects like OpenJS are part of LF. I wish there was a breakdown of what they spend on what...

This article just complains about The Linux Foundaiton decision of reducing the LTS releases from 6 to 2 years but doesn't cite or analyse their reasoning. Why they did that? It may be worth complaining but the author should do better.

Then again when did just throwing more money at the problem solve more problems? There could be other bottlenecks that the foundation felt are not solvable by money alone.

LF does fund a decent number of core kernel people and some infrastructure, but they could do so so much more. The organisation is so corporate it's kinda gross, I really wish more of that got reinvested into the kernel

Alas, that is the state of things. It's not even really up for debate. It just is what it is.

Its not up for debate apparently

is this for real? very concerning, and i'm curious as to why all these years they've decided to abandon it.

The generous reading is that they see more value in focusing on projects that spreads and popularises Linux and makes it more ubiquitous in more areas, rather than on core Linux kernel development.

Reply on news.ycombinator.com