I inherited a lot of Ubuntu servers at the university, too. But I am not directly responsible which makes life easier; I am just managing it.

Interestingly, they agreed to monthly updates with possible restarts and they are fine with it, because I keep the servers healthy. And: We even plan to move from VMWare hypervisor to Proxmox VE as well, but we can do it in stages without big downtimes.

There is one CentOS server carefully isolated which cannot be updated anymore. Moving it to Rocky would introduce a big downtime and redoing a lot of custom config. Luckily the user-facing server of that cluster is running a current Rocky Linux.

The things, I established so far, are running stable Debian. Nice to see Proxmox VE being based on Debian. (:

It is interesting that you are in a similar boat, but with a different outcome. I hope that your colleguas will reconsider some day.

This is the most accurate chart I have seen so far; I can relate to all of them except NixOS, because I have tested it only briefly in a container.

I first settled on Ubuntu and than moved from that along the chart and I am know in the Debian phase for my main distro after ~10 years with Gentoo.

For others I recommend Fedora or Debian depending on their needs and skill. Arch has no place for me anymore. I can live with Ubuntu on servers, if someone else had it already installed, but I would never use it on a desktop. But Debian is a default on servers/VMs for me, too. And: I can see, why people choose rocky on servers. I am partly responsible for some rocky servers and they seem to behave nicely.

I was a passioned Gentoo user for many years and I also only met 1 other IRL Gentoo user. Ok, there were more once at a Linux conference where a bunch of Gentoo users had a desk showing off Gentoo compiling @system for a BSD system.

Even the binary repository was not enough for me to not go back to Debian after ~10 years Gentoo usage. It was a fun hobby and a great learning experience.

Please don’t be angry with me, but the package manager is called portage; emerge is just one commandline tool to interact with it.

Have you read the btrfs or zfs man pages? They are very well written. You just have to want to understand it. If you don’t, let it handle your tech-buddy or admin, or have them quickly look over the command that is about to be confirmed with pressing ENTER.

Better go one step back to Debian instead.

Yes, but be sure to exclude certain services on servers and monitor the remaining pending upgrades.

Thx, I may need to look closer at noctalia-shell. If it can provide some comfort I am used to from gnome-shell, this could be it.

I considered PaperWM, too. Again, I must say. I have used it years ago. In the past it couldn’t keep up with the Gnome development. I tried to use it in later version by fat-fingering in some files. But there was a time when I could not make it compatible again. I havn’t tried it since there is new momentum in the project.

When niri first came out, I tried it with waybar. I am not sure, what I missed most, the way I can start programs or the ability to change typical settings in the gnome-control-center. The last bit could be an issue with noctalia-shell for me, too.

I hopefully will have time to experiment between the years.

Since 4 months I use Gnome on a 100Hz display at work. At home I am on a 60Hz panel. I cannot justify buying a new display, because the older 60Hz has still better colors and automatic brightness, which isn’t very common, it seems.

But I want higher refresh rates now everywhere; a dilemma!

I want to, but I miss the top panel the way gnome-shell has it implemented.

Is it possible to run niri instead of mutter together with gnome-shell?

I knew it.

Never heared of the other two.

Thx, I will use sPOngEbOBcasE from now on. [=

whatever, but it is not snake or kebab case either.

Let me rephrase: Why WhateverCase?

Why CamelCase? /o\

this is ls -shit. ^^

I game on Debian; it is absolutely up to the task.

It is called the universal operating system for a reason.

Thx, I decided to not use raid for shipping.

this is scientific data.

Funfact, I recently did a scrub on my offline backup drive of my work PC. It correct around 250 errors. I wouldn’t have noticed any problems if I had used ext4 instead of btrfs.

I agree with both of you. Somehow I don’t worry about the drive in my laptop but 80 TB of scientific data is another thing, and I want to make sure it is the same data when it arrives.

That sounds scary and like I need at least btrfs if I need to ship the data instead of using rsync.