Manjaro is my first experience with a derivative Linux distribution. My old distro was openSUSE which is a fully standalone OS: Manjaro by comparison is kept up to date from Arch. I’ve heard only good things of Arch so it’s surely a good decision I salute! Still I’m curious about some implications.
Firstly I should check that my understanding of what being a derivative means is correct; Manjaro’s OS packages are modified mirrors of Arch’s but with slight changes, updated from them but with different settings and software potentially added / removed on top. That’s referring to the official repo, I know AUR is a separate entity and its own thing.
Anyway, as an user who intends to stick with a good distribution forever if possible, this does raise a little concern in the back of my mind: Supposing something were to happen to Arch, or Manjaro had to take a different route for any reason… what would happen to the OS? Could the Manjaro team ever branch off Arch entirely and become fully self reliant if need be? Would it just mean the Manjaro team has to update and build each component manually, and if so how hard is it to do that?
Obviously Arch is well established and reliable, this will likely never happen with so many derivatives depending on them. Still I’m curious what the difficulty level were if this ever had to be considered; It’s an interesting situation to have an OS depend on two different teams which don’t necessarily work together, a setup in which things usually go one way (Manjaro must take decisions based on Arch not the other way around). It’s easy to worry that if something ever happened to either of the two sides, you may be left having to look for a new OS or things could change drastically… even if I don’t imagine Manjaro would ever have the fate of Antegros being so popular and well established by now.
Something might happen to upstream KDE, GNOME, Xfce, Cinnamon, or MATE?
Something might happen to the most prominent Linux kernel developers?
Something might happen to Debian, which will directly affect Ubuntu and Mint?
Something might happen to LibreOffice, which one could argue is currently the most widely used open-source alternative to Microsoft Office?
The relationship between Manjaro and Arch is just one of many, and some have greater ripple effects (such as mainline kernel development) than others.
With open-source and a community that consists of people from across the world, all with shared and vested interests, it’s unlikely that a project like Manjaro would sink like the Titanic and leave users without a working OS. Worst-case scenario? You’d still be able to use your computer for a good while (on whatever packages / updates you last had on your system prior Manjaro disappearing like a rabbit in a magician’s hat), which gives you plenty of time to migrate over to another distro. And that’s the WORST-case scenario.
That’s a good way of looking at it too. Components are easier to replace than an entire build system, but one can also look at Arch as just one dependency of many. Especially since as far as I know, Arch is intended to be more of a base for other distros to begin with… given how it doesn’t even have an installer, I didn’t look at vanilla Arch since a VirtualBox install would have been too complex
We have some overlay packages based on what we want differently or add a fix for something that arch doesn’t have the way we want it. That replies also to this:
So, not always that is the case.
Nothing bad will happen to Arch.
That new guided installer provided will make even more people to take that route and adopt Arch Linux as their main system. Maintaining and updating it is similar, community might be different, but not that much if you stick to technical stuff and not mention other distros.
By the way, welcome to Manjaro community!
Get used to using backups - Timeshift is great. It meant I could ditch Mint (cinnamon) and have a ready Manjaro KDE desktop with only some minor tweaks to the changes in shortcuts (i.e. my Mouse gestures worked okay once I updated a few, qBittorrent started up seeding all the files on my disks - everything was hunky dory).
So if Manjaro closes down tomorrow, you only have to worry that Arch will survive - and the only thing that will kill Arch would be a complete revolution in computing which would mean nobody had any need for it anyway…
So my opinion is, basically, bring it on as quickly as possible!!!
It is extremely easy to convert a Manjaro installation to use everything - including kernels etc from Arch repo.
Serving as a reminder: Doing such conversion will not qualify on Arch BBS - you are still using a system which you are not exactly sure how it is was built - so never ask on Arch BBS unless you are running AFS (Arch From Scratch).
Thanks for all the answers, I did learn a few new things today. Such as the fact that my comment on Arch being a generic Linux base was likely questionable to say the least I got that idea after I tried core Arch in Virtualbox and noticed there isn’t any installer, which made me think “so Arch is a collection of packages for devs to branch off not meant for normal users to install”. It’s nice to hear they have a proper installer now… I have Virtualbox working on Manjaro again so I may try comparing it for fun.
Is that the feature BTRFS uses to create snapshots of OS changes? I tried it in openSUSE after installing my first SSD and it was a disaster: I remember the snapshot creation tool (snapper?) started using 100% CPU soon after install, the system suddenly froze almost entirely and I could barely use it for hours. I had to do a forced restart at some point, after which the OS broke and would no longer boot again, becoming stuck on a start job and complaining of corrupted data. I had to do a full reinstall and this time I did it on ext4 without backups enabled so I wouldn’t experience that again. I didn’t try the snapper service on Manjaro and I’d be afraid to as it would likely do the same.
I can see why it gives off that impression, but Arch Linux used to be my primary desktop distro before I switched over to Mint, openSUSE, and Manjaro. What appealed to me was that Arch Linux does not require compiling code from source (they supply binary packages from their official repos) and that you can “build” your own custom Arch Linux desktop by specifically choosing what packages to install: nothing is “pre-bundled” so to speak.
The only “takeaway” from Arch to Manjaro is that I can troubleshoot Manjaro with a terminal using what I learned from my days on Arch Linux, as they share the same “base”. Other than that, the experience feels much different. Using Manjaro feels more similar to using Linux Mint than it does compared to Arch. Manjaro’s update policy is more tuned to the desktop user (less frequent, more tested), and it provides live ISOs and installers that come pre-installed with a working, polished system: ready to go, ready to install, ready to be human!
The reason I left Arch Linux is because I don’t want to allocate as much of my free time walking over glass with their fast-paced updates nor go through the Leggo blocks procedure of building from scratch on a new PC or laptop.
There’s more to life than geeking out on a Linux distro and bragging to your friends.
I run both, Manjaro and Arch on the same disk. The Arch install is a minimal system, Manjaro includes the full KDE flavour. Performance is very close, only some more disk space required for Manjaro.
Arch helps me to understand what packages in Manjaro won’t be essential and makes it easier to uninstall them and Manjaro is the source of inspiration if I miss some features in Arch. I love both.
I guess I’m still curious about the last part of my question; In theory, how hard would it be to branch off Arch and do its part? I do NOT think this should ever happen in practice, sticking with Arch seems like a really good system honestly. But I’m wondering solely to understand the process better: How complicated it is to maintain an original package basis like Arch does, or how they work with the developers of individual software to know when and how to update and package it.
It is an opinion based question. The only way to know is to try it out yourself. Assume Arch does not exist, and try to rebase your Manjaro installation to any other distro of your choice (e.g. compile apt or yum, remove pacman, and so on).
Oh… it could be rebased on a different distro too? I’m assuming that might be more complicated… it would likely imply large changes to the OS, like no longer using pacman / pamac which would be sad as they’re great tools. I was thinking of the scenario where everything is kept as is, including all software and the packaging system and default configurations, only instead of getting updates from the Arch repo the Manjaro team would manually update all software in Manjaro Untested. That would be more my definition of branching off
Which brings up a little detail I missed: When taking updates from Arch, does Manjaro pull binary packages or just source code packages? Like are the installed applications (those that don’t need to be patched) compiled by Arch then copied whole from it, or does Manjaro only pull source code packages from Arch then compile them internally? I take it it’s the later case… for example if I look in the about section of Firefox it says “Mozilla Firefox for Manjaro Linux” (not “for Arch”).
I don’t have BTRFS. I use Timeshift to backup to my mounted /mnt/T3 disk.
It uses the folder /mnt/T3/timeshift/, which it mounts for it’s own use at /run/timeshift/backup/timeshift/ (you can use BOTH these paths to get to your snapshots).
Using the normal rsync non-btrfs timeshift means I have my snapshots even if my system disk fails or is reinstalled, on /mnt/T3/timeshift
Yes, I just read it works without relying on BTRFS features, such as by using rsync which is a tool I’m familiar with. I’ll take a look at setting it up, sounds like a good idea to have this option available. Even if in case something goes wrong, it’s easy to do a fresh install of Manjaro so I’m not worried.
There’s still something I’d like to understand on this. Again because I’m curious and like to know how the OS I use works
I noticed when it comes to vanilla Arch the “keep it simple” principle applies even to the software included: There’s a lot of stuff that’s unavailable in its repositories but does exist in Manjaro. Is it thus safe to assume that Arch is only used as a base for updating core system components, whereas most non-essential software (themes, games, others) are maintained by the Manjaro team? Note that I did this comparison without AUR enabled, I haven’t turned it on yet as I don’t seem to need it so far.
Also I’m curious about my other question from earlier: When Manjaro rebases itself on Arch with each update… does that mean it copies most software packages whole including compiled binaries, or only source code packages which Manjaro then builds against the libraries in its own repository? I take it the later must be the case as a lot of stuff would break when patched otherwise.
Manjaro is based on Arch, and as such there are of course tons of similarities and compatibilities, but it’s not guaranteed to be similar or compatible in every single case. It’s the same as with Ubuntu vs. Debian for example. Manjaro has some packages which are Manjaro-specific, and it has modified packages and/or modified configurations of some packages, which make them different from the default Arch package. It’s probably quite similar, maybe so simiilar that you won’t recognize the difference, and many packages are maybe the exact same thing, but overall it’s not the same. The Manjaro Team quite explicitly states that while Arch is used as a base, that doesn’t mean that Manjaro is just a “pre-configured Arch”. Just like Ubuntu isn’t a preconfigured Debian. The differences go a bit further than that. But it’s also cool when the derivative doesn’t stray too far from the base, because then there’s greater package compatibility.
