I’m looking to start 2026 with Trixie and dumping the ubuntus.
I’m thinking of going fully overboard with partitioning my / & /home onto individual btrfs partitions.
Am i inviting a world of pain or do i Keep-It-Simple-Stupid with a single, ext4 format
BTRFS
ext4 KISS
0voters
And do i start with the full debian 13 and start removing and replacing software packages, or start debian-minimal and enema a post installation script with preferred software packages, but might have a bunch of dependency and support issues?
Honestly am I the only one that has ever had major issues with Debian?
Their default sudo configuration is a pain in the ass for some who…uses any other major distro really and somehow, possibly my fault but I’m willing to share, every installation I make has borked repositories where I need to manually disable the installation disc as a main repo for software…
I voted ext4 but it depends, what is this installation for? E.g. on a RPI or N100 nuc running PiHole / HA / other Docker containers, I use ext4. Workstation or NAS would be something different.
Always minimal, purge auto remove is a pain, there will always be some config file left. At least that’s my experience.
Are there good docs of Debian + ZFS. I’ve been meaning to finish up my own Gentoo+ZFS guide. I’ve got several machines on ZFS using ZFSBootMenu with native full disk encryption and it works really well. The ZFS command line tools are way more intuitive than btrfs as well
Knowing that I have to do that…? This procedure makes it a not so reliable distro in mission critical situations when I and other people I work with have to learn new distro specific behaviour. So much for being a super stable reliable distro when the basics operate in a different manner than the vast majority of distros. Even after I learn how to deal with these issues I still can’t rely on other people on the field with less technical knowledge to operate this thing. This is why Ubuntu will keep its market share for a long while.
All distros have quirks. In this case it’s actually more stable since no editing of sudoers is necessary, only a group policy.
These days there are three main Linux “standards” out there¹, RHEL, Arch and Debian. They all do 95% the same way and 5% differently. Debian is stable as a rock and in principle never fights you on your config. That does come with costs though, like outdated packages.
Debian sure ain’t perfect… But neither is any other distro
¹ There are others like Slackware and Gentoo, but those are a sliver of a percentage.
Ubuntu hand-holds by having sudo installed, and adding the default /first user to the sudo group.
Coming from that, it might not be obvious one needs to elevate to root, like with su and then set their user account to be a sudoer or in the sudo group or whatever
It is not the end of the world, but is a surprise first time one encounters it.
On the other hand, some distros don’t seem to have sudo installed at all, presuming administrators would use a root login/su instead of their own login
It was not a major hold back, but was a “huh, okay” moment, when I switched from X’buntu to Debian, presuming it would have more in common
Why is that? Because you may have to read the docs?
Back in the day, every UNIX version from a different vendor used to do things slighyly differently, and you always had to read the docs to know what’s what. Just because Linux kernel and GNU userland are supposed to be compatible across distros doesn’t necessarily mean everything is done the same way. Hence reading the docs is the prudent thing to do.
Honestly, if the docs are missing/incomplete/incorrect/hard to find/something else in that vain, I understand you. If you just can’t be arsed to read them, that’s on you, not the distro.
When choosing to use a distro for production I take into consideration installation and configuration time. A distro that is more complex and has its own specific way of doing things, quite different from the majority I might add, makes it more difficult to work with.
If I need to drive 3 hours to my client, Install 10 machines and configure them, this added complexity is a factor of decision.
If I need the client to install the distro for me and I’m doing remote support, the less complex distro will be the favourite.
All this discussion about just accepting the quirks reveals a lot of you guys don’t have much real work experience with these distros. This is why canonical win hands down with enterprise.
It can also be fun playing with network settings with different spins of different distros using a mix of netplan(some Ubuntu) systemd-networkd and network-manager. As well as systemd-resolved, or not…
Again, it’s not the end of the world, just can be a shift needing to know which place to look for certain things/settings
I do think it’s a PITA that systemd duplicates so many existing, so there’s now two ways to do a lot of things in Debian, the original way and the systemd way
Never had any Debian specific sudo problems, but I recently switched to doas which I’ve been finding much nicer and straightforward to configure. Hopefully it proves to have less CVEs than sudo.