Stay on LTS kernel or update to 6.7

I know I already replied. But, after re-reading this - an issue caused by and a bug within the kernel is a special scenario. Very advanced users who can fix their own OS, with or without asking for help. . . don’t even fall into this. You would need to know how to locate the issue within the kernel, fix it and recompile the kernel. Even very advanced users don’t/can’t do this. This is very specialized by folks who are generally kernel devs. There might (and I stress might) be a couple people on this forum that are able to do this. Maybe folks can find work-arounds. . . but when things get fixed in another kernel update - you may or may not have to undo that work around, or create another work around. But to actually fix issues caused by kernel growing pains. . . that’s gotta be very rare.

I’ll stick with 6.7. Normally no lts kernels for me. That’s just how i roll. :laughing:

That’s not actually what I meant by “you can fix your OS without asking for help”.

I meant it in more simpler terms like if the system breaks, you know how to get access back into the system to for example downgrade the kernel, recover a snapshot or whatever. Nothing to do with what you mentioned. :slightly_smiling_face:

There are plenty of users who have no idea how to do any kind of fixing without asking for help. Nothing wrong with that either. Just that, that was my point.

In the “before times” the advice used to be to keep, at least, the last 3 kernels. These days, I’ve always felt that you always install the LTS kernel. If you want/need to run something else (latest-and-greatest, Zen, Real-Time, etc.), fine, but have the LTS kernel also installed as a fallback in case there are issues.

1 Like

That just wrong on many level… i in shock :astonished:

EDit… @fbodymechanic think i number 3 on list :rofl:


You already mentioned this a few hours ago . .

I understand. People who can backtrack or create work arounds. It’s just not really “fixed” as much as it is just damage control until those bug reports are worked on and it’s actually fixed from within the kernel. We definitely needs these folks. If no one tests anything, then no one has any issues. . . . so there’s no way to know bugs exist until released to the masses, and by the time that happens - it’s too late.

You sir, are absolutely 3. You’ve like Apple and WM’s - you seem to love the things that don’t work. . . .

1 Like

Well i say it again…because i don’t think people are fools just because they don’t use the lts kernel. It’s a choice. I don’t get much breakage and even people who are using the lts kernel have their system broken occasionally also. Breakage isn’t always to do with the kernel. I don’t think there’s anything wrong with using an lts kernel but I’d rather use the most current kernel.

Way ahead of you.

1 Like

I wouldn’t use any OS that would force such a specific preference on me.

Besides, the reasoning behind it is completely wrong. No matter how ADvAnCEd of a user you are, snapshots are not a guarantee that you will be able to recover from a system issue like a major kernel regression.

In fact, the very opposite of this “emphasis” is true: if you’re using Btrfs, you should almost certainly be using the LTS kernel. A kernel regression causing hardware damage due to excessive IO on Btrfs is not just a possibility, it has happened before. You can’t snapshot your way out of hardware damage.

Thankfully, issues like this are very rare, especially on ext4, but unless you have hardware that is just a few months old, there is typically no reason to use anything other than the LTS kernel.


I don’t like topics like this, because a lot of tribalistic, uninformed people are just spouting and repeating nonsense, and I don’t want to waste my time arguing with them.

I’d say you decide for yourself. In the meantime you can read what Greg Kroah-Hartman wrote.


I totally agree with that. Every once in a while the newest stable kernels introduce nasty regressions for various drivers. This can affect file systems, graphics cards, usb devices, etc.

Even ext4, which is known for its stability got hit but such a bug in December 2023 with kernel release 6.1.64. Debian even postponed its kernel update for that reason.

btrfs has had several regression in the past years too. Not to mention the issues that can happen with nvidia, zfs, amdgpu, sound cards, etc. with the newest stable kernels.

If you want to be on the safe side you want to stick to the LTS kernel.


On my last PC I used to use the mainline kernel but on this one I switched to the LTS as my PC would not shut down on other kernels (has started working again in the last month and I only noticed when I accidently booted this one). It really depends on hardware and your use case. No one really cares what version of a kernel you are using, what matters is does your system work as expected and thats what matters in the end.

I’ve always thought that If your hardware is compatible with the latest stable kernel, you don’t need to stick with an LTS one.

1 Like

True thats how I used to think but always good to have a fallback

Sure, additional kernel doesn’t take up much space. If something goes wrong, I can always switch to LTS.


exactly even though you may never need to use it always good to have a plan b

1 Like

Considering 6.7 version: everything’s been updated successfully and works fine, but I’ve noticed RAM usage being ~300 MB higher after that. It’s not a big deal for me, but does anybody have the same thing?

Article is a little old but, I was thinking of mentioning that, if a particular distro has a kernel they run with, use that. Even then (and this is just my personal preference), I still like to have the LTS kernel as my fallback. Having said that, I will say that I’ve not experienced a major kernel issue in a very long time.

1 Like

There’s another factor here that’s worth bringing up. I have a computer I “tinker with” and a computer I “depend on”. The former is good for learning on but the latter absolutely, positively, needs to work and work reliably. Therefore, I take more precautions with that PC. Having a backup kernel is part of that.

1 Like

My definition of force in this context, just to be clear, is “out of the box”. I run Tumbleweed on another system and nothing is stopping me from sideloading an additional kernel. I just mean out of the box it’s a win to have snapshots ready to roll.

Of course, it’s not a guarantee, but the hypothetical you’re raising is so rare and unlikely to occur, which is why I don’t worry about it or entertain that kind of thinking. Are we to assume that the team behind SUSE is just stupid and doesn’t know what they’re doing for setting up their distro this way? I mean, it seems to work just fine for them, and they’re rated the #1 and most stable rolling distro to many users? Just saying. They obviously have the option to add an LTS kernel as well, but they don’t because it’s not necessary for how they’ve setup their distro.