Hello Friends
Just being curious if we can share a List of Linux distributions where each one offers the two most used environments: Desktop and Server such as:
Where so far:
Thanks for your understanding
People build servers and desktops with Gentoo.
There is no easy iso. You build it from a stage 3 tarball
Same with FreeBSD… there is a basic iso. You build the rest either with ports or with packages.
Stuff of nightmares for those unlucky enough to inherit support for such… I’d shit bricks if I ever came across such…
I worked on a recent “Proof of Concept” (we just call it POC) - basically “appliances” rolled by the vendor (a small “startup” that’s part of Curtin University) for network security… I was dismayed to discover they were “Arch based” and a tad “misconfigured” (the vendor had included a “comma” in place of a “dot” in a hardcoded IP address). I scratched my head, WHY would you choose Arch for something that might require long term support?
From my perspective - there’s little difference between a Linux “server” and Linux “desktop”.
Both Red Hat Enterprise Linux and Oracle Enterprise Linux installers can install a desktop UI (and both default to a GUI for the install). CentOS is (or was?) the same.
On my own desktop Linux environments - I do most of my configuration in the shell (e.g. bash or zsh) - it’s where I’m most comfortable… So from my POV - the differences are very trivial… And you can always, after installation, install a DE on top of a Ubuntu or Debian “headless server” install (same for pretty much every Linux or UNIX platforms, like Solaris, or any of the BSD’s, probably AIX too).
e.g. you could run an Apache website serving hundreds or even thousands of users, on a desktop Linux system, if it had enough resources (CPU, RAM and network throughput).
I’ll append to your list :
and there’s more - as @nevj has mentioned - you could use Gentoo as a server, and there’s Arch based servers, and of course FreeBSD…
I thought a rolling release was eternal LTS.
Yes
I wasn’t referring to the vendor term “LTS” - I was referring to enterprise environments, i.e. corporations, with SOE (standard operating environment) servers - throwing another distro into the mix makes long term support tricky… Occasionally my “Wintel” colleagues might be called upon to step into the “gap” during an absence - and having to support e.g. Arch or Gentoo - when all they know is from running a Ubuntu VM at home…
I DETEST environments where different teams, from “other” SME domains, throw in any (their “favourite”? The only one they know? Something someone suggested) distro into the mix, and all of a sudden you have an alphabet soup of Linux distros… It’s horrible…
RHEL is not my personal preference for a server platform (I’d prefer Ubuntu - but I’m not on the “architecture” team - so I can’t make decisions like this - many of these “architects” know bugger all about Linux or UNIX anyway) - but - it’s better to have just one - and that makes it much easier to develop a SOE / MOE - e.g. develop a VMware template, or an AWS EC2 AMI - with the base O/S that’s going to be your SOE - and deploy ALL servers from that “base”…
I get it… you want economy of scale .
Too many varieties of server leads to support headaches.
I keep forgetting you guys deal with many machines.
I haven’t counted - but yeah - I have “machines” of various distros and releases (even including Solaris) - I’m expected to support - I reckon well over 2,000 - probably closer to 6,000 (taking into account bare metal Solaris, Solaris virtualised via LDOMs, and Solaris further virtualised in “Zones” [kinda like Docker or FreeBSD jails]). It’s kinda “hellish”…
I don’t mind Solaris - but one environment I do after hours support for relies on Solaris for their external DNS stuff…
Thankfully most of it is RHEL8 or 9, but there’s still legacy stuff out there running on RHEL5 of 6 or 7, and even (FFS!) Debian 3!
I stupidly checked my work emails today - and - some PM (project manager - some of the most unrealistic expectation humans I’ve ever encountered) wants us to do a full failover / DR exercise - within a budget of about 40 hours - before “go live”… Yeah - dream on! This is to replace a product we deployed with 3 people in 2020 - now there’s like 15 or 20 people involved! Too many chiefs, not enough Indians…
My first main memory of “project managers” dates back to over a decade ago - maybe a quarter century - ringing me on my mobile just before beer o’clock “do you know we’re upgrading product X today and need you from 3:30 pm and all Friday evening”… On the same Friday evening? What were they managing? Timelines? Fantasy?
You can remove CentOS from that list - it’s basically dead now. It has been replaced by Rocky Linux.
I still use it - from time to time (i.e. “daily”)… And my main SSH jumphost is CentOS 7… I want to replace it (and patch it) but it has an uptime (after being vmotion’d across various infrastructure) of over 7 years - so I don’t want to touch out of purely sentimental reasons 
- I have a sentimental fondness for uptimes…
Do you use kexec to change kernels?
If so, I’d love to hear how that works.
Dont mention them to me. They ruined CSIRO .
There is nothing worse than people who are “managers” but have no expertese in any activity they are attempting to manage.
Science used to be managed by scientists.
Thanks a lot for the replies … Yes, CentOS and Rocky is other history
Very interesting read your experiences.
Thanks again!