Canonical’s announced a major shift in its kernel selection process for future Ubuntu releases. An “aggressive kernel version commitment policy” pivot will see it ship the latest upstream kernel code in development at the time of a new Ubuntu release.
Original announcement: Kernel Version Selection for Ubuntu Releases
Ubuntu was first os I really stuck with for years… It’s weird the shifts they made in the past. The horrid Amazon search in unity shell was their first major misstep… And as much as I understand the snap shift, their implementation was balls. I was forced to jump ship when a work reliant version of Citrix somehow completely would break app armor…
I don’t know what I intended to really say in this post… Just typing out loud I guess.
I am right there with you on all counts. I’m on openSUSE now and happy as a clam.
It’s a great entry point for people too intimidated to try something like Pop!_OS, Mint, Aurora, elementaryOS, etc. But for some reason, the people who wind up in places like this tend to use anything but Ubuntu, it seems…
I use Ubuntu for work and it serves its purpose really well.
It was the first time I really ditched windows and learnt how to use Linux.I have Kubuntu on my personal PC and it feels klunky to me.
So I am not sure why is that, since it uses the same base.
My only issue right now is that I need to split the apps I use between dpkg, apt or snaps and it sucks when I need to uninstall something.
So once my project is done at work, I will try another distro.
Just to offset the predictable groupthink in this thread: Ubuntu is fine. In my experience it is rock solid and has been for years. Doubly true for the LTS versions. Yes there there is the slightly troublesome issue of Snaps and the even smaller one of self-advertising. But IME the installer is very solid and that is a crucially important issue for prospective normie users. Ubuntu is still a flagship distro and IMO it now deserves more love than it is getting.
Ubuntu is the only enterprise distro that I can run both at home and at work that also has reasonably up to date packages. Debian and OpenSuse and CentOS (RIP) all run much older packages that may not support what I want to do at home so then my home experience would not match my professional experience.
Sure there’s fedora but I don’t want to be reinstalling my servers every 8 months or so as a new release comes out
Ubuntu has long support windows and reasonably up to date packages on recent releases, so I can do whatever I want to without too much faffing about but not have to dist-upgrade every 6-24 months if I don’t want to. Plus it’s an easy one to whip out at work for something because it’s a well established enterprise vendor
I’ve been generally happy with Ubuntu. I don’t really care for snaps, but on my headless server that’s not really an issue. I suppose I could have taken the time to uninstall snapd from the server, but I haven’t cared enough to do so.
I ran it on my desktop for quite awhile as well, but there the snap issue was much more present. I hate Firefox as a snap, and while I’m aware of the new Firefox ppa, I decided to switch to fedora since I’ve never used it and wanted to broaden my experience a bit more.
And yes, I’m aware of Red Hat/IBM’s shitty corporate bullshit too. Maybe one of these days I’ll use Arch btw.
Great. Now Linux Mint will have to start providing their own kernels too, as they were following Ubuntu’s way of choosing a kernel version.
Will this be the final nail in the coffin that will make LMDE the main edition, or will they just follow what Canonical is doing in that case? I’m genuinely curious for their response.
I’m not familiar with linux mint, why?
Also they can switch to debian base relatively easily
From my understanding the Mint most ppl use/are familiar with is Ubuntu-based.
Linux Mint LMDE is the Debian base! 😎 love LMDE so far!
from the site: What is LMDE?
LMDE is a Linux Mint project which stands for “Linux Mint Debian Edition”. Its goal is to ensure Linux Mint can continue to deliver the same user experience if Ubuntu was ever to disappear. It allows us to assess how much we depend on Ubuntu and how much work would be involved in such an event. LMDE is also one of our development targets, as such it guarantees the software we develop is compatible outside of Ubuntu.
I’m surprised by this decision, since Ubuntu’s strength is stability and by extension, friendliness to new users. Imo, a better move would be to ship a separate “unstable” release with non-LTS kernels.
Maybe stability is not a frequent issue nowadays, and they need the new kernel to support new hardware more quickly?
E.g. I can imagine a new linux friendly laptop can’t be sold with ubuntu preinstalled because the old kernel is not supporting some parts yet, but it’s already merged upstream. Or something like that.
Ubuntu ships “hardware enablement” kernel updates to LTS versions of their distribution so new hardware can run older releases.
I just read the article and they say exactly what I guessed:
“This approach would guarantee stability on the appointed release day, but was proving unpopular with consumers looking to adopt the latest features and hardware support as well as silicon vendors looking […] to align their Ubuntu support,” Canonical’s Brett Grandbois explains.
But to “provide users with the absolute latest in features and hardware support, Ubuntu will now ship the absolute latest available version of the upstream Linux kernel at the specified Ubuntu release freeze date, even if upstream is still in Release Candidate (RC) status.”
The announcement has some wording on treating LTS releases differently. My guess is they’ll be more aggressive on non-LTS releases and less aggressive in LTS, in order to preserve stability for LTS. Besides, non-LTS releases have been decidedly unstable for a while now, especially after the move to shorten their support lifespan. And it makes sense. They more or less serve as testing releases for the next LTS. Point being that whoever wants stability uses LTS anyway and they’re likely to adjust the new kernel selection process to keep that stability.
How many engineers can Canonical yeet now that they can skip on testing and backporting fixes to their own stable kernel?
Are they also going to tell Joe average user to just submit any bugs to LKML?
Compiling a kernel yourself isn’t a big deal these days, especially with DKMS. Generally the type of people I’ve encountered who care about which kernel version they’re usiyare the type of people who are capable of compiling it themselves…
I’ve found that is shifting a bit as a lot of newer hardware needs kernel support, and as new people with newer devices enter the linux world they can encounter issues. I know I’ve had to wait for feature to make it to the kernel before I got it for some newer hardware. It can be frustating especially if it’s something essential or realky desirable. Even more so if you aren’t tech savvy.
