(No provocation)
I see these reasons:
- newbie
- lazy (don’t wanna edit config files etc.)
- unique features (like assistant/toolbox, some optimizations like in cachyos)
- wanna check how different systems are set up (that’s rather distrohopping)
Personally, I used manjaro i3 when I was beigginer and wanted to see how tiling WM should be configured (check out ranger config, for example). But after some time, I don’t see reasons why not to just customize pure arch (same with debian and debian-based distros).


Because the PC enables me do do my hobbies (gaming, 3d modelling and printing), the PC itself isn’t my hobby. If I spend more time tinkering with the OS than having the OS run the things I actually want to do, I’ll go do something else.
I think this is a common misconception about Arch, that it requires continuous tinkering. I see that word used so much, too, “tinkering”.
What I’ve been doing for the past decade is just install Arch, set things up the way I like, and then just keep everything up-to-date as I go. Of course, I install and uninstall things as I try new software, but the OS itself? Zero tinkering. I just use it.
Especially if you only game on it and stuff like that, then simple plain Arch is great. Lean system that just works. Install the things you want and enjoy.
I got two kids and way too many hobbies so I can relate to not wanting to fiddle with the OS. I run Arch on my two home desktop PCs, and my two work laptops. 🤷♂️ Zero maintenance.
Enjoy!
I see what you mean. But in order to reach the point where Arch is configured and my machine Just Works, I would have to learn how to install Arch, what packages I need, what are the ups and downs of the various packages for handling the same things, resolve any conflicts I accidentally created, and then I can get to installing the things I actually want. It’s a lot of work and time that isn’t going into something I consider fun.
Arch is great for people who want to build their OS to be precisely what they want it to be. I happen to not be one of those people.
I consider that important knowledge for just using and troubleshooting your system, and one of the great reasons to go through the installation process - knowing what you’re using and why means when you want to change the behavior of something or figure out why it broke, you’ll know what software to look for.
Not just something to do for fun, or to have precisely the right things, but so you understand what it is you truly do have. This isn’t to say your approach is invalid, but for me it’s a reason to recommend Arch for people who want to go through the learning process.
Arch elitists don’t want you to know this but you can just type archinstall and a text installer guide thingy will install the system for you based on what you tell it.
That said, you’re spot on on the last part. I love arch exactly because it is the easiest one to customize the way I want to but it’s not for everyone and not everyone should use it
I didn’t know how until I did it. There’s a very comprehensive guide that you follow. If you can follow the recipe for cookie batter and manage to make the cookies, you’ll be able to install Arch. 🤟
To be fair that’s something you’d have to do regardless of distro, even on Windows. Learning which tools you like takes like half a career worth of time.
Why Arch based distro then? Why not, say, Fedora? Debian. Popos.
In my case - 90% of Linux issues eventually lead to an Arch Wiki article any way. Might as well give it a go, but I’m too lazy and too much a noob to try the real deal.
I started out with Ubuntu and went straight to Arch. I knew absolutely nothing. I followed the installation guide to a T, and it worked. I didn’t understand anything I did. Then I installed it again, in a new computer. I understood a lot more the second time.
You don’t need to know what you’re doing in order to succeed here. There’s a lot of handholding and learning as you go. 🙂
That’s exactly my experience. Now I understand most things I do, and I smile at this ‘installing Arch is difficult.’ No, it’s not. I can install it without any help from the wiki, by memory. As I understand what I’m doing and why. It’s not the difficult part. The difficult part is to make it yours.
I still can’t install from memory 😆, even if I understand every step now. The reason is quite simple for me: I install it so rarely because it’s so stable. I only ever install it on new hardware. Every computer I have has basically only seen a single Arch install. 👌
Fair point! I just did many installs recently (a bit of a long story), and at some point just stopped even following the wiki. But if I can afford it, I simply clone my entire system, and tweak from there. Takes very little time, and I have a complete clone of my perfectly working system.
Also, theoretically, I don’t even need a backup of the system, if I have at least two laptops with mostly same system. I have, one at home with broken keyboard and no battery, which servers as my home computer connected to a display. And another one is for on the go.
The complicated parts of the setup are already done and I still get the benefits of a fast update schedule, without (mostly) having to worry about accidentally breaking something.
And you’re right about running another, less finicky distro. It’s why I’ve got Bazzite on the HTPC and the handheld. Because I want those to be zero finicky, as poking around in them is way more of a pain in the ass.
Opensuse tumbleweed would likely be a good choice.
AUR, etc
As a lazy admin of my own computer, I agree… for the most part. Running Debian allows you to be super lazy, whereas Arch will punish you for that. One update screwed up my GRUB because I didn’t bother reading the news. Totally my fault, learned my lesson.
This means that running Arch comes with some responsibilities that a super lazy Debian admin can simply ignore. Just read the announcements before updating and you’re good. Ignore them at your own peril.
It wasn’t a total disaster though. Just needed to fix my stupid mistake with chroot, and the system was up and running in about half an hour. Debian admins don’t end up with situations like that by being lazy. You would need to be actively trying to break your system to have to pay a price like this.
Other than that, my system has been running smoothly with hardly any interference on my part. The joy of a rolling release…
Yeah, I mean, if I were to admin a fleet of computers or something like that, I would definitely not run Arch on those. 😅
LOL. Same.
Every now and then I find a comment where someone who clearly knows what they’re doing is deploying an Arch server in a work setting. Feeling confident with that decision takes something I don’t have. Maybe it’s experience, knowledge or something.
I believe there are docker images for Arch? But those probably have some form of reproducibility, I should hope. Since you can’t install specific versions of packages declaratively with pacman unless you have physical access to the actual package file, I would not use it for a server. Maybe coupled with Guix or something? I dunno.
Maybe they need an environment with very up-to-date packages, or something along that vein. 🤔
Docker would be the sensible approach.
Imagine if the new version of some package is only compatible with the old version of PHP, but now Arch is using the new one. The admin would have to keep the old version of PHP until the devs of that other package release an update. I recall reading something about pacman not supporting partial upgrades like that.
Exactly, partial upgrades are not supported. Always best to do a pacman -Syu to upgrade the entire system instead of just one package.
At least that’s the case for system packages. More isolated packages don’t need as much care or carry much risk of rendering a system unbootable or similar risks, but I say it’s like signaling your turn when driving even if no one is around — best to just make it a habit. 🙂
Why do people say to keep Arch healthy you must follow the mailing list for needed manual changes for some upgrades? It that a misconception?
That’s not a misconception, but with a slight modification. Stick around:
First of all, you can decide to do an upgrade at your own leisure. There’s no need to run it 10 times a day as some memes will have you believe. I upgrade probably once a week, but I feel like that’s a lot. If nothing is broken, no need to upgrade. Of course, I want to keep up with security updates and browser updates, so when those happen, I tend to upgrade too. So probably once every six weeks should be enough (keep up with the browser schedule).
Anyway, the modification I mentioned is that you don’t need to follow anything. The only thing you should to do is to check archlinux.org for any news items before you upgrade. It’s extremely rare that there are manual actions needed, and when there are, it’s even more seldom for a package that I personally have installed. I think maybe once or twice in a decade I’ve had to actually do anything, and it’s been minor. It always has the exact command or tells you exactly what to do. They never leave you on your own with these things.
A lot less maintenance than Windows back when I was running that, I’ll tell you that much.
Thanks for the response and the clarification.
I haven’t run Windows in 25 years and I no plans to. I’ve actually been using linux for my desktop all that time, including gentoo for several years. Personally, I think people should use whatever os they like and the more choices the better!
Definitely, fully agree. Gentoo I can imagine is more maintenance than Arch, even. At least a lot more waiting around. 😅 But maybe that’s a misconception, too!
But yeah, definitely use whatever you like. I just want to clear up the misconception that Arch is heavy on maintenance. It most definitely is not, unless you want it to be.
I originally switched to Gentoo when I got my first AMD64 workstation. Gentoo was the only distro with full support and optimization for a little while.
For a big build I would kick it off at bed time :)
Ah, neat. I assume it had early support since you compile the packages yourself?
And tell me if you will: did it ever occur that a build had failed when you woke up? 😅 Or maybe builds didn’t really fail? How common was that?
Builds failing was pretty rare for me but I didn’t customize the compile options much beyond AMD64 optimizations.