I can't remember what I was looking at, but within the past week I ran across a comment that said Arch Linux only supports constant upgrades such as that, and any delays that result in skipping a version are what risk causing breakages. The commenter was very surprised that they were even thinking about supporting a version jump on whatever the thread was about (pretty sure it was somewhere here on HN).
There's also no need to do that, since nobody prevents users from installing newer versions alongside old ones, and invoking them directly.
Even not considering the fact that different GCC versions can coexist, complaining about this breakage in absolute terms doesn't make any sense. Releases changing default compiler versions inevitably cause more breakages with packages who haven't been updated to be compatible with newer GCC versions. So ultimately it's a matter of choosing the distributions with the appropriate release model, not a matter of Ubuntu/Debian breaking stuff.
Libav has also been deprecated in Ubuntu long ago, so it's not clear what you refer to. If you happen to refer to the transition from and back to ffmpeg, that's very old history.
I'm referring to the libav* libraries which are part of the ffmpeg project (not the horribly-named libav fork) - and external debian repos providing updated version of those (due to better codec support in media players, etc) such as debian-multimedia
ndiswrapper was the main cause back in the day, shockingly giving Windows drivers access to the Linux kernel can cause problems. The most recent time was when VDPAU was new and I was trying to get HD video playback working on a mini-PC with an nVidia Ion GPU by running a version of the nVidia driver much newer than Ubuntu packaged. Now that I think about it that must have been around a full decade ago.
The only two major distributions that I used for a sufficient amount of time are Ubuntu and Arch. Arch "unstableness" is exactly what I want most of the time on my personal computer as it's my to-go Petri dish. "Stability" would mean that it's harder for me to break it apart, and make a Frankenstein out of it. That's exactly what I have been experiencing with Ubuntu LTS releases --- stability.
Most of the time, I want to have all the available LLVM versions alongside with all the GCC versions, with all the available binutils (Qemu, Docker, Oracle VBox, etc.) versions on the latest kernel full of my monkey patched printk's. When I finally get to break its back I dive the Wiki for few hours to restore it.
I can imagine a non-office, hacking desktop OS that follows the Arch packaging strategy being highly successful.
I also maintain a few compute servers for 10-20 people. They are on Ubuntu LTS. The packages that I need there are always the ones that just work and don't let anyone do anything "cutting edge".
Arch is rolling release and you take the good with the bad. The ones who try to defend arch as some paragon of stability miss the point that Arch's model is inherently unstable, but it comes with other benefits.
I'm the one who kicked off this entire conversation pointing out that arch is unstable, and it cracks me up watching silly people scramble to try and defend Arch as being some paragon of stability.
No, it's not. That's baked into its identity.
I also agree that you shouldn't run your production database on Arch Linux. It isn't made for workloads like that. But personally I find maintaining Arch Linux+"custom packages"(with AUR) easier then Debian+"latest packages"+"custom packages".
Back here in reality, rolling release is less stable because more bugs in the software get through. And this is a reasonable expectation and not some magical fairyland where bugs never get written so being right up against the dev branch is as stable as being on the stable branch.
We really live in two different software worlds. Every software I'm using has its number of bugs a purely decreasing function of time, especially in the "main" paths and use cases.
If it were true, it means there wouldn't be bugs in the first place because they wouldn't have gotten written. The very fact that the bugs got written implies new bugs can, and will, be introduced.
There were rough patches along the way but, all distros were having the same problem in someway or another (vdpau, fglrx, multi-gpu support, ndiswrapper & wireless stuff, etc.) but it's a set it and forget it affair for a very very long time.
...and obligatory xkcd: https://xkcd.com/963/
To counter with anecdata: I've been running Arch Linux testing repositories for over a year now with no issues whatsoever.
That said, just because you had good luck doesn't mean that it's stable.
Here's the most trivial way I can think to explain this:
Check out Arch News.[0] Ctrl-F (Find) 'manual intervention'. Six years of results on the first page ; 13 instances of 'manual intervention required'.
Reliability != stability. Stability usually implies a platform on which one can use and develop for without expecting common major changes, if ever.
As for the anecdote of using the [Testing] repositories, a users experience with such things really depends on their use of new and currently developing software.
A simple computer with simple peripherals that is used to run emacs all day isn't likely to be broken by the Testing repositories.
Personally , my anecdote : the second someone starts using Arch for something new-fringe (hi-dpi, touch, tablets, SPDIF, SLI, NVRAM, new window managers, new x-windows replacements, prototype schedulers, filesystems, or kernels, PCI pass-through, exotic RAID configurations..) [Testing] repository is an act of masochism. It's only a matter of time before something drops out.
I found that my best bet was chicken sacrifice and Opus Dei-style self flagellation before each [Testing] 'pacman -Syu' . At least then I had a 50/50 chance of the next boot. But, granted, I use a lot of weird or fringe hardware.
All that said : other distributions don't have better [Testing] repositories. It's just that [Testing] is for .. testing. It's unstable by its' very nature.
13 instances of "manual intervention required" over 6 years seems awesome.
I think I did 2 manual interventions during my time using Archlinux, and each time it took maybe 2 minutes, it was just a matter of copy-pasting the commands in Arch News.
Meanwhile you can use RHEL or CentOS and basically leave the thing alone for 5 years.
Distros targeting the same desktop needs are Ubuntu, Mint, Debian-flavors, Fedora, Gentoo, ...
In corporate environment for critical software you would get a contract with an on-call duty and you would only have a limited choice fo supported distros.
There is no defending it, it's a fact of life. You get bleeding edge at the cost of stability.
I probably wouldn't recommend Arch for corporate environments, but at this point I feel I should point out I wouldn't recommend Linux either if you're working in a windows shop.
I've had various non-rolling distros fall apart on all sides within 6 months. With Arch I've had one manual intervention(aka one copy-pasted command) in over a year and one minor issue where I needed to restart a service. That's amazing and something I'll gladly take in exchange for painless and straightforward setup of pretty much anything.
Not to mention with regular backups(that should be done either way) and delaying updates in critical time periods you can easily minimize the risk.
That's the minimum that you should be doing regardless of the OS. And thus, that's not where the risk is at.
It's probably more difficult to understand if you get paid regardless of whether you can get meaningful work done.
But for those of us who live on the other side, losing an unplanned half a day due to an OS issue that actively causes us to lose money, it matters.
Arch is bleeding edge, this means software bugs are far more likely on it.
I had Arch on a laptop I used for my freelance business once and took notes of the time spent on maintenance. So I actually have the numbers and know how much money I wasted compared to the time I used an Ubuntu LTS. Guess how many hours I wasted on repairing breakage on Ubuntu LTS. Exactly zero.
However, I come from a business perspective here. You really don't need an operating system that introduces new versions on a whim in the middle of operations. There's a reason businesses run RHEL/Centos or macOS and whatever has LTS in Windows land: cost of maintenance and reliably wide windows of no possible breakage.
As a business, it's kind of a dodgy position to be at relying on "yeah, randomly upgrading versions worked for me so far. Fingers crossed, lol".
So I'm interested in what people use their Arch Linux for and what didn't work for them on any other stable platform. My personal observations point towards the Arch Linux users being the tinkerers who like to use cutting edge and hose their Debians trying to wedge some newer version into it. While the ones that ran screaming away from Arch pretty much never had the desire to change the underlying system and were happy with a security backport now and then.
You don't want your PHP version to endlessly get updated during your operation when you build your app at a specific version.
It would mean not only you have to take care of distro update gotchas but you have to go through all the breaking changes the language introduces.
Stick with LTS. Rolling is for enthusiasts, though Arch does have put decent efforts on their wiki but I did find errors on minor pages that didn't make it work as written.
I think we'll figure out how to upgrade in the 2-3 years of support PHP has for each version. Trying to compare this to arch's rolling release style is bananas.
100% my experience as well, down to the freelance. I made another comment pointing out that the ones who think it's just a minor issue with Arch get paid whether they're getting meaningful work done or not.
> It is the user who is ultimately responsible for the stability of their own rolling release system. The user decides when to upgrade, and merges necessary changes when required. If the user reaches out to the community for help, it is often provided in a timely manner. The difference between Arch and other distributions in this regard is that Arch is truly a 'do-it-yourself' distribution; __complaints of breakage are misguided and unproductive, since upstream changes are not the responsibility of Arch devs__.
Even their wiki is very clear to inform you that you're at the whims of the package upgrades.
But I also want to point out that "it hasn't broke in a year" is cute, but I ran arch for probably 10+ years until I finally got tired of it and moved to Ubuntu. The last straw was me sitting down to get paying work done and spending half my day trying to recover my work environment. In the 2-3 years since I moved to Ubuntu I've never once sat down at my PC and had something stop working that was working before.
This defense of your favorite distro is especially silly when you think about it logically. Of course a rolling release system is going to have more breakages. The sensible response isn't "it never breaks!", but is instead "that's the nature of rolling release, you opt into it when you choose Arch".
For a long time before using arch I thought too that rolling release might be more unstable, but I have come to the conclusion that quite the opposite might be true.
This assertion makes no sense. Debian Stable has always the paragon of stability to the point that it's actually criticised for it. The Debian Testing release is even famous for being more solid than other distro's stable releases.
Breaking changes in Debian are practically only remotely possible in Debian Unstable or if you purposely install packages from backports or PPA repositories that have no assurance of stability or even maintenance.
Maybe you meant debconf? When I was a DD it was an RC bug if your package used debconf to do "magic", i.e. using it as a registry for an application, or not properly re-seeding answers from the configuration and assuming the debconf db was the source of truth.
That's a fist-full of caveats. Arch is willing to upgrade the kernel, maybe you should include a dist-upgrade in the comparison.
I can plan for it? You can see if you're upgrading your kernel. Nothing breaks if you decide not to run an upgrade.
This nonsense is why stable and LTS releases were introduced in the first place.
If I'm faced with a major upgrade of an LTS system, I usually choose to reinstall from scratch, shedding packages I haven't needed for a long time with it. But that happens every couple years or so.
If you're okay with hitting the forums now and then to find out why your desktop has been behaving oddly the last two weeks, rolling is your thing. You won't have to break in your new dist-upgraded or freshly installed system for like two weeks.
If you don't want to play lottery every time you hit enter for that update or if you're a business where you can't afford possibly breaking all your laptops for some security update, stable is your only option.
It's "choose the best tool for the job" I guess.
This goes back to my original point, which we've drifted a little from.
Ubuntu has more unexpected behaving when upgrading than Debian.
A working dist-upgrade should not be something that your OS struggles to provide. Asking people to reinstall their OS isn't an acceptable answer.
Upgrading should not be a lottery.
This argument that you MUST dist-upgrade as soon as the next version is available, and completely ignore the entire idea around LTS, is silly.
This idea that you never want to upgrade from one LTS to the next is equally ridiculous. Upgrading the operating system is part of the package manager's job.
I like Linux, I don't dislike dealing with the OS. But I want to deal with it on my time. Expecting to get work done and then realizing something is broken and being forced to deal with it is a completely different proposition from picking out a weekend to reinstall your work PC with the newest LTS version.
I went through some hell when I ran Gentoo unstable(~) but for the past several years I've been on Gentoo stable and have very few issues. I've even used it on work laptops at three different companies.
It is definitely more of a do-it-yourself distro than Arch for sure, but I enjoy working with it.
That never seemed acceptable to me and I ended up moving over to Arch Linux for a good 10+ years before the instability of it finally got to me and I moved to Ubuntu.
Having said all of that, I'm glad to hear that Gentoo is stable for you. I really enjoyed my time with Gentoo and I got the feeling that after he left there were mistakes made in terms of everything working reasonably.
Your description of Arch reminds me of an experience I had with linode and arch years back. Due to an internal hardware issue, they basically lost my server. When I went to recreate, I discovered that you couldn't just update the world because it would immediately result in a broken system. You instead had to reconfigure a few things before updating the world.
I no longer use Linode.
At some point the vendor has to take responsibility. That's a large part of why I like Ubuntu, they do. It works every day without me having to worry about it.
I actually appreciate that Gentoo is set up to allow me to run that command if I want to.
But I did move off of Gentoo due to constant dependency conflicts. They took forever to deal with and they were pretty well guaranteed to happen whenever you didn't hold to a strict and frequent update schedule.
Well, no. The point of a rolling release system is that you upgrade by little increments, so you never have a moment when a 'major upgrade' happens and breaks your system or forces a reinstall.
Maybe don't pacman -Syu before work or if you're working the next day? It's not strictly necessary to update every day.
And no one said Arch never breaks, I just provided anecdotal counterevidence that it's really not that bad.
Why be condescending? The above user has a good point, and your view is clearly against the majority of Arch Linux users.
As I understand it, Arch Linux is like having a pet. It requires constant care and feeding to keep it alive but can be very rewarding (allegedly)
And the reason for this is because you realize the latest update killed virtualbox and you require it for your work so you need to be able to downgrade and pin the vbox application to a specific version.
Meanwhile, Ubuntu catches the problem and refuses to update the package until it's fixed. Therefore I never even realize there was a breaking update to vbox somewhere.
The difference being that Arch tells me when to tinker because it blew up again, Ubuntu allows me to choose when to tinker because it doesn't.
In my experience it happens very rarely and is often quite easily resolved.
>happens very rarely
>is often quite
I just gave a link from their FAQ that clearly states the system will break when packages break and they aren't responsible for it. As another user pointed out, you can go do a search for 'manual intervention' and find years of results on their front page.
Arch Linux users are ok with having their system randomly stop working. I used to be ok with it, until the umpteenth day that I lost half my working day getting the system back up and running.
Also, personally I've stuck with Arch Linux for 6-7 years now because it's the first distribution I found that consistently works for me, so I don't know that anecdotes will get us very far either way.
So, your comment is 180 degrees off the mark. Arch exists for people who don't want breakages to happen.
(On the other hand, if you're conditioned by Windows to reinstall the system every year, then Ubuntu might work fine for you.)
Or, for people who don't want breakages but also want/need current software versions.
Debian-based stuff can be pretty nice and stable as well, but on Arch installing and setting up anything usually "just works" without missing dependencies, renamed packages etc.
I think I'll take the planned upgrade, thanks.
PPA breaking is apparently not Ubuntu's fault.
I think the main issue/benefit with arch linux is that it wont try to do anything for you, except give you extra instructions for when a package change requires "manual intervention".
I put that in quotes because I don't think of that as breakage, I think of that as normal upkeep. We'll, as long as it's listed.
Anyway, the implications are that things can be stable longer on arch because there is no magic under the hood. But there will be times (though I haven't experienced it myself), where something unexpectedly breaks, and it'll be hard to recover unless you know what your doing... though maybe it's actually easy with the rolling back packages? The package manage does keep copies of old versions of packages, I've just never needed them.
As an example if this, my co worker and I have very similar thinkpad laptops. His graphics start doing wierd things that hes has to go fix after every update (Fedora or Ubuntu, can't remember). I took 4+ hours to get mine working how I want, with research and experimentation, but nothing has broke it since.
More UX anecdote; I use Arch (btw) but trying Manjaro just now for something so I grabed an image from osboxes.org; the system update fails after hitting enter on the default options https://i.imgur.com/OTNRMrH.png
'y' for the last would have continued things, but "ugh, I have to think about something" is breakage to some.
The thing is, my experience with Ubuntu is that there will be times when something unexpectedly breaks, and because of the added complexity, it will be nearly impossible to recover even if you know what you're doing.