Meanwhile I’m over here like…
> doc@gigabizzle:~$ uname -r
> 6.15.6-1-liquorix-amd64
> doc@gigabizzle:~$ sudo zfs --version
> zfs-2.3.99-461_gd6dcae316
> zfs-kmod-2.3.99-461_gd6dcae316
> doc@gigabizzle:~$
3 Likes
How could you think that a liquorix kernel and zfs is a desirable state to be in?
My impression is that they are hi-tech fringe experimental stuff.
Correct me if you have a different view .
1 Like
The only issue I’m having here is the fact that I had to compile the ZFS kernel because of Debian’s “issues” with ZFS, so it’s kind of a bleeding edge version. Doesn’t give me the warm fuzzies, lets put it that way.
Not ideal, but until I can get my pool backed up in order to fix it …it is what it is. I’m not running ZFS on the root drive either, just on the storage drive.
Liquorix is a custom kernel build, mostly based on zen kernel patches, aimed at responsiveness and desktop interactivity. It’s often used by low-latency audio folks, gamers, or just people who want snappier UI behavior. It’s not experimental in the sense of “held together by duct tape,” but it does include more aggressive scheduler tweaks and features that haven’t hit mainline yet.
So basically..
I’m on Debian, so using zfs-dkms.
The issue isn’t Liquorix. It’s Debian stable pushing a newer ZFS version that:
Requires newer kernels than Debian ships by default.
Might mismatch the DKMS build during a kernel update, but that’s a ZFS-on-Linux packaging issue, not Liquorix’s fault.
Issues:
DKMS needing to be rebuilt whenever I update kernels — but that happens regardless of whether you’re on Liquorix, mainline, or a kernel you rolled yourself.
1 Like
A reliable filesystem is rather basic. You know what you are doing and you will make it work… maybe Debian testing and get a newer kernel.
Is this a server? I thought servers tended to be old stable stuff.?
I hate the thought of a filesystem misbehaving on me.
There is some sort of move towards shifting filesystems out of the kernel into user space (like FUSE). I saw some discussions in the NetBSD world.
2 Likes
In case anyone is interested, these are the release notes for “Trixie” aka Debian 13.
My own personal preference using Debian 12 as my daily driver is to hold fire until at the very least the first point release and probably later but will more than likely do a vanilla install to a new PC
1 Like
I have Debian 11. I am not going to try an upgrade. I shall wait, like you suggest, then do a fresh install on a new partition beside the Debian 11.
That gives me opportunity to transfer software and deal with any issues.
3 Likes
In that case, you may also find this resource helpful
3 Likes
nah this is my workstation. Nowadays my servers tend to be either HardenedBSD or FreeBSD.
1 Like
I dont think I would choose ZFS for a workstation… not a multiboot one anyway. It tends to want to ‘own’ the disks.
2 Likes
i dont have zfs on the boot disks, I just have ZFS on a third storage disk, shared between the two OSes.
1 Like
That makes sense, especially if you have lots of data and it is dynamic.
I have a shared data disk like that, but it is ext4. I rsync it to another disk nightly… sort of like a raid hand driven.
3 Likes
yeah. I got into ZFS via my servers, which have more of a use case for it since you have RAID Z1 Z2 … but it’s also good on single drives as well if for nothing else than the dataset/snapshot features.
But yeah, usually I have ZFS spread across 4 drives in a RAID 10 like setup. lol
2 Likes
Actually it is (was) August 9, not 19th. And so far the RC2→ release path has been a smooth one (for me at least). ZFS is not a major pain point for me, so I’ll list the “enjoy points” one by one, I admit it’s fairly informal testing (over 2+ months of running RC2):
- KDE Plasma 6.3
- Speed of reboots noticeably improved over nearly every other distro I’ve tried
- Secure Boot totally works OTB, like OpenSUSE
- More stable kernel, not bleeding edge as usual for Debian, but net effect is much faster updates and fewer reboots, compared to OpenSUSE and Fedora
Everybody needs to make their own decisions about their “Daily Driver,” but I think I’ve found mine in Debian 13 Trixie!
Most probably, I will reuse my MX partition for something else, seems to be little advantage to the 25 version over “mainline” Debian Trixie.
2 Likes
I’ve been running Debian 12 as my daily driver for well over a year now (~18 months) and love it to bits after moving on from Ubuntu 22.04. I like its clean looks (I’m a GNOME fan) and now don’t miss the Dock taking up space on the left of the screen preferring to use the Meta key to select applications either from the pinned apps or by just typing the first letters of the app name in to the search box.
Even with Ubuntu, I preferred vanilla installs of new versions (to clear down the inevitably accumulated crap) and would wait until at least the first point release before venturing in to the new one as my daily driver.
I shall probably do the same with Debian 13 as that will also give time for 3rd party applications to catch up and avoid disappointments. No doubt I’ll source a new mini PC for this and then keep the existing one as my scratch unit.
2 Likes
That is a good move. A bit of redundancy takes away all the stress of doing installs.
3 Likes
Normally I would agree, but I took the time and trouble to install Trixie RC2 “bare metal” and fidget with all the “obsolete-ware” (usually named -ESR) and the upgrade process, as mentioned before, was surreally smooth (under 5 mins., included a reboot). Soooo, I feel pretty comfortable proclaiming Trixie as my DD. I have disabled Discover updates on all KDE instances, and share config with OpenSUSE Slowroll and Fedora Plasma spin, so far haven’t seen any ill effects (OK, it’s been a whole 2 days of “triple booting”).
1 Like
You just need to manage it carefully. Decide which Linux is going to control grub and always do update-grub in that Linux whenever you change anything.
If you dont know which Linux is in charge of grub, it will be the one in which you last did grub-install
3 Likes
Used to favor grub, now there is just too much contention between btrfs, secure boot, sysvinit, systemd, old/tired nvram, etc., etc., that grub no longer handles “gracefully” so I’ve pretty much switched to rEFInd as my boot mgr.
It’s just a wack-a-mole contest between different OSes (yes, I include Windows as the prime “repeat offender” although more than one Linux flavor does this nvram write or declares ownership of the grub universe without prior permission) that want to write to nvram and/or control the boot loader sequence.
2 Likes
I am really locked into grub. It would be an effort to change.
3 Likes
I’m sorely tempted to source this from Amazon.fr as my new DD for a vanilla Debian 13 install. If it all goes wrong, I can at least take it to our local computer guy to sort out for me but the absence of any OS on it is quite alluring… I also have an ext4 reformatted 1TB HD left over from an upgrade to my Synology NAS which I can use for additional internal storage. All looking quite attractive…
1 Like