"We asked the Open Infrastructure Foundation why it thought CentOS had taken such a dive,"
Well, if you shit on someone from a great height then don't be surprised if they move.
Canonical is cock-a-hoop after Ubuntu snatched first place for OpenStack deployment from the CentOS Linux distribution – but according to some the victory might ring hollow. The finding came from the 2020 OpenStack User Survey, organised by the Open Infrastructure Foundation, which queries respondents on a variety of topics, …
As a big Linux fan, I have to (regretfully) ask if there's just too much choice out there in the Linux world?
Sure...the Windows mono-culture is a pain in the ass.....I left that nightmare twenty years ago. But do we really need umpty-ump desktop distributions?....Umpty-ump server distributions?.....endless arguments about the meaning of the word "free"? .....warfare about the difference between licences (you know GPL2, GPL3, Apache, MIT.......and on and on).
Can't we get real? Maybe four or five desktop distributions (RHEL, Ubuntu, Fedora, Mint, elementary)? A similar (small) selection of server distributions?
I don't even mind if MY FAVOURITES don't make the cut!!
There are three packaging solutons: tgz/git/build from source - Gentoo/Arch/LFS - .deb - Debian and Ubuntu and derivatives - .rpm - Fedora/RHEL and clones, and SUSE.
Apt wins dependency management for packages, I think but the others don't always have as many packages to care about anyway.
Of these for third party packages: .rpm is good for some third party commercial packages too but .deb and the Debian package universe has more or less won everything else.. [See, for example, the relative pain of producing a full bootstrap and rebuild of Rocky/Almalinux onto an ARM system like Raspberry Pi.]
Add in the horror of Red Hat subscription/license entitlement and .rpm is likely to die out in the long term, IMHO.
There is Arch's Pacman which now retrieves ZST archives (a fun thing to find out if you ever try to update an old Arch image which lacks the ZST decompressor which is of course available as a ZST archive). There are a few package managers which exist primarily for embedded, including OPKG which has its own format. Although apt/yum and deb/rpm are the most common, there are several alternatives in modern use.
I've been using Linux for slightly longer, but I don't mind the large number of choices. Niche things that scratch itches seem to be a good thing, overall, and show that the environment is still growing and viable. And cutting out "choice" would likely make things worse - witness the current systemd mess that all of the distros (ie - the decision-makers about this kind of stuff) jumped on with both feet, without even giving sysvinit a goodbye kiss. I understand your sentiment about narrowing things down, but I don't know that I'd trust most of the distros to do it "right".
If you look at the whole survey, Ubuntu, CentOS, and RHEL have a total of 87%. Add #4 Debian at 6% and #5 Microsoft Windows Server at 3% and you're up to 96%.
Essentially nobody is using the rest for OpenStack work. If, instead of forking, developers tried merging low use projects, there would be more maintenance and development resources for the merged project and users choices would be simplified.
Sounds like you assume only one person/group builds all these rebellious distros.That would waste time and energy. But it's not so. At most, a few people/groups build/maintain a couple of aspects. eg Desktops, Browsers, Tools, Multimedia, etc. And again, you're free to choose. I stick with Kubuntu - no big surprises and mostly pleasant. You don't have to try every distro.
Actually, there is only one Linux. That would be the kernel.
When you take that kernel and add the packages that you, personally, need/want, you have a distro. Ideally, all of us should have our own individualized distro ... unless you actually know a person who uses a computer EXACTLY the same way that you do, of course.
So no, there is not "too much choice" out there ... if anything, there is not enough.
At this stage of the game, we should be moving away from marching in lock-step, all using the exact same set of tools, and instead customizing systems to suit each and every user. (For example, MeDearOldMum has no need for the kernel source and most dev tools, so I don't install that stuff for her ... the rest of her system is similarly customized).
Sadly, that would be hard ... and people like the easy way out. So they just use what is given to them, without bothering to think about it. And so many/most are stuck with the systemd-cancer, for example. Including most down-stream distributions, which makes it look like there are too many distros, even though there are really only a few.
So yeah. I don't use OpenStack, but I do have a few servers that I needed to upgrade last year, mostly because I want to avoid "the cloud" aka somebody else's computer. I'd been running CentOS, but when I started testing CentOS 8, I discovered that it was absolutely awful. Ok, fine, 7 is still supported for a few years, I'll just use that, it's workable and maybe they'll get their shit together.
Then the stream thing got announced. Big NOPE from me, these things need to run for years with nothing but security updates, I'm trying to keep my pain in the ass level low.
And now those servers are running Xubuntu. Yes, with a GUI, my servers have GUIs, because sometimes it's easier to do things with a few clicks, sometimes it's easier to type stuff, and so I do both things. No, they don't boot into wtf-ever they're calling runlevel 5 these days, but if I type startx I want that GUI there, and XFCE is a LOT nicer than the current Gnome horribleness.
I'm not surprised at all that Red Hat is getting ditched. Now that Rocky Linux is a thing I could go there, but why? It's still got the Red Hat 8 awfulness where things just seem to be broken at random. The only way I'd use it now is if I had some specific project that didn't work on Xubuntu.
I too am holding off on 8, and sticking with 7 for my new development machine. Still got 3 years to see how things develop.
Bailing out of Centos to a RHEL Developer subscription.
... and XFCE is a LOT nicer than the current Gnome horribleness.
Maybe, for the time being.
But the forecast is rather gloomy.
With the jump to 4.16 and CSDs you may be in for some of that Gnome stuff you make an undeservedly polite reference to.
No way I will upgrade my box to Xfce 4.16, I'll pin 4.12 till I find something else that suits me.
Interesting, you bailed out of CentOS before moving on to 8. I tried for three years (or more) 7 and decided to not go further and go back to unsupported 6.10 Final. :-(
Now I currently don't use CentOS as a server OS, but just for my desktops. After so many years (and using it) I still can't stand Gnome 3. I remember being 10 years ago at a RedHat Summit in Boston where the creator announced this refresh. Can't remember his name, but I've hated what he did (not him personally !!) ever since.
Haven't decided where to go next. Mmm, funny enough CentOS 6.5 seems to have lesser problems for me than 6.10, like a properly working YouTube with sound in FireFox, etc. <g> Doesn't work in 6.10, probably some DRM issue.
A most places where I've worked since both distributions have been around, the people who choose what to put on servers have picked something from the Red Hat family, and the developers have mostly chosen from the Debian family, which has occasionally brought slight incompatibilities to light.
May I point out the obvious as someone who has used CentOS, RedHat, SunOS & Windows in a production environment and developed on them. There is one common feature: CentOS was by FAR the worst experience as a developer.
Perhaps CentOS was rubbish and people were forced to use it? (As I was.)
I have also used Suse, Debian, Ubuntu and now use Gentoo/Linux. The latter two I even use at home, when I need most reliability.
CentOS is literally not fit for purpose now, since CentOS users wanted (and were getting) RHEL without paying RH. I can't blame either side for their positions, but I tend to blame CentOS users more.
I've been using Debian forever*, though, and Debian wants my help, not my money. I'm looking into adopting a couple of packages that have lost their maintainers.
*1999 was my first permanent install, so not literally forever
Freeloaders is a little bit of a naive assumption.
I share much of my work for free, as do many others in the open-source community. In turn we do expect others to do similar. These day, the *only* way to get high quality software is through this kind of community process.
And if they can't (like CentOS seems unable to), they should move aside to make way for a project who can. Preferably they should actually close down to reduce fragmentation too but that is up to them.
So no, this isn't like the world of Windows or gaming where people just want stuff for free and demand more and more like little fat children. Likely all of us in the open-source community have spent a considerable sum of money (including our time) investing in the ecosystem. I personally have spent vastly more money on open-source than on proprietary software because access to source code *is* worth more and especially since I do not engage with DRM.
But its fine. The project is forked and we continue with our lives. All it has succeeded in doing is waste a little more of our time recreating the infrastructure and re-organising.
Sorry, three months late posting a followup.
Red Hat doesn't care that you share your code. Red Hat sells a product, a part of which is your code that you shared for free.
Red Hat wants money in exchange for their product. CentOS was RHEL with the parts that said RHEL filed off. Functionally identical. Red Hat bought CentOS and put a stop to that. So there's another fork going to provide a RHEL-clone for free under the terms that Red Hat has accepted for their product.
So you took offense to the word "freeloader". I apologize. I was attempting to get across that CentOS users aren't paying CentOS or Red Hat for the work being put into their respective products. I don't think that was an inaccurate description of end users of the product.
Red Hat agrees with me, though, which is why there's a new project to recreate what CentOS was before Red Hat bought it.