The future of the CentOS Project is CentOS Stream, and over the next year we’ll be shifting focus from CentOS Linux, the rebuild of Red Hat Enterprise Linux (RHEL), to CentOS Stream, which tracks just ahead of a current RHEL release. CentOS Linux 8, as a rebuild of RHEL 8, will end at the end of 2021. CentOS Stream continues after that date, serving as the upstream (development) branch of Red Hat Enterprise Linux.
Meanwhile, we understand many of you are deeply invested in CentOS Linux 7, and we’ll continue to produce that version through the remainder of the RHEL 7 life cycle.
CentOS Stream will also be the centerpiece of a major shift in collaboration among the CentOS Special Interest Groups (SIGs). This ensures SIGs are developing and testing against what becomes the next version of RHEL. This also provides SIGs a clear single goal, rather than having to build and test for two releases. It gives the CentOS contributor community a great deal of influence in the future of RHEL. And it removes confusion around what “CentOS” means in the Linux distribution ecosystem.
When CentOS Linux 8 (the rebuild of RHEL8) ends, your best option will be to migrate to CentOS Stream 8, which is a small delta from CentOS Linux 8, and has regular updates like traditional CentOS Linux releases. If you are using CentOS Linux 8 in a production environment, and are concerned that CentOS Stream will not meet your needs, we encourage you to contact Red Hat about options.
We have an FAQ to help with your information and planning needs, as you figure out how this shift of project focus might affect you.
[See also: Red Hat's perspective on this.]
717 thoughts on "CentOS Project shifts focus to CentOS Stream"
This is dumb. The entire premise and the only reason anyone uses CentOS is because it's rebuilt RHEL. Congratulations on undermining that, nitwits.
Feb 2022: RedHat starts sponsoring Rocky Linux! 😐
Yea and after RH owns rocky linux, it is possible for this disaster to happen to rocky too...
IMHO It will be enough that the developers of Rocky do not accept any interference from RH. Do I miss something? :-/
RedHat will destroy Rocky Linux just as it does with the other projects it acquires 😀
I only feel anger. For every RHEL server I have I have ten CentOS boxes. I had just started migrating to CentOS 8. Stream is not an option, I need stability and dependability, not another Fedora.
It is perhaps not unexpected that RedHat will try bugger us this way (remember the transition of Redhat Linux to Fedora), but not less disappointing. I will not be looking at migrating to RHEL.
true that, Redhat just does not acknowledge the fact that CentOS was a good test system for RHEL. We use RHEL in Production and CentOS in test, it was a good solution because:
1) you don't really want to pay licenses for tests systems
2) you want a test system which is 1:1 to the productive
our decision will most probably be to switch Linux distributions and choose some distribution that can at least offer me a possibility to operate systems with decent costs not having to pay for testing.
That's unfortunately just greed from Redhat (or maybe IBM)
I think CentOS must have been eating a large chunk of RHEL market share.
If all a user wants is quality and assurance, CentOS is the perfect clone of RHEL.
The only missing bit is a Partner Account Manager, who could expedite bug fixes for a particular customer.
I see what RedHat/IBM is eyeing is to squeeze more money from such free users.
if you need stability try slackware
I have no doubt to choose Ubuntu as a server xD
use arch linux for a server i guess now
There goes the last reason to stay in a combined RHEL/Centos environment.
Looks like we'll be migrating the enterprise to ubuntu.
For prod workloads with incredible support I would go SLES honestly. Still RPM based but like night and day when dealing with them. I wouldn't go back!
Do you work at SuSe ?
RHEL has, far away, better support and ecosystem than SLES.
We use SLES on a part of the infra and wouldn't go there again.
We have been forced to Canonical because of available expertise amongst support personel, same reason we used CentOS: there are a ton of RH admins to draw from.
Good luck with Ubuntu forced "cloudification" in production environments!
That's ridiculous. Ubuntu is 100% production-ready, no problem. There's also a cleaner version without cloud-init, if you'y.re referring to that - but when you've got hundreds of servers it might not be a bad idea using it anyway.
To summarize: at this moment I have to pay for my Redhat production platforms, everything is normal until then. And I must therefore forget the stability (we cannot trust Redhat for more than 6 years a priori) but I have to test and participate in correcting bugs in my future production version or pay for my development, training or labs platforms. So I have to do on my no-prod environments what I already pay for my production platforms? What a motivating change and deal....
Debian is a better choice, Ubuntu is a Debian compilation, Debian is a distribution from scratch
debian here i come! been using it as a desktop driver. now it will be server too! effffff you rh
I have already approved my org to move our 100 rhel servers and 40 centos server to debian.
I migrated to Ubuntu in 2012, don't regret for a minute.
This is a bad idea. I use CentOS for stability.
Redhat and IBM's ball...their rules...sorry buddy but if you wanted a vote you needed to have been a contributor for the last half decade or two when CentOS NEEDED YOU.
Now that IBM is daddy...they DON'T need you anymore...
Wow, that was rude but it is what it is. I hope Centos Stream take stability to another level.
To a much lower level you mean. CentOS Stream the test version, it's not stable by definition.
I agree that is pretty rude and uncalled for.
CentOS 8 Stream is extremely stable as it is Awesome. Get on with it and move forward. We got a Lot of Good Things Ahead!!
Nothing would have changed by participating other than you would have had a goose egg to show for your effort. Everytime I’ve been pulled in with that line, I just ended up disappointed. By the same logic, you can enjoy a bug-free computing experience if only you let everyone collect your telemetry. (hint: whether you do or don’t, it’ll still be riddled with bugs … but they’ll be glad to blame your lack of participation anyway)
This all sucks, but I appreciate the decade plus that CentOS gave me. The gave me an enterprise-grade Linux bistro for free.
It would be insane to get mad when they finally decide to stop. All I can really say is, thank you so much for letting me have it as long as I did. It certainly increased the quality of my professional life.
I hope this turns out to be positive for Cent.
Have you sold the project to Webpros ? :))
Wahahahaha. I am falling of my chair! Everywhere that crap Webpros and everywhere where they are involved prices increase by 1000%
E.g. Plesk, cPanel, SolusIO, WHMCS and what else. #pathetic #webpros
My employer has a paid RHEL subscription, but we developers often use Centos in VMs and for ad-hoc tasks because of its simplicity and availability. I personally often spin up a Centos VM while working from home, and build and package our proprietary Ruby RPMs (which work differently to 8 modules - multiple versions installed at the same time), and we later deploy the same RPMs to RHEL knowing it will work the same. I also authored my recent Ansible on Centos at home to later use on RHEL.
This will create compatibility problems. I need the assurance of binary compatibility, and I personally work on installed RHEL servers, but do not have access to install or update my own.
Outside of work I am about to update a single C6 VPS to C8 for non-commercial internet hosting, but all of a sudden Ubuntu LTS, like I am running successfully on my Raspbery Pi is looking like a better (safer) option.
Get yourself a Red Hat Developer subscription which is free for non-production work.
You will get a self-supported full RHEL version! For all your virtual machines!
So no issues with compatibility at all.
Go to https://developers.redhat.com/ to register.
You can do that but a dirty secret is a lot of those CentOS systems get used in production for minor stuff. Most RedHat shops I've seen use up to 50% CentOS.
Some even use 99% CentOS..
In all my time I've only ever had to contact RHEL support once, and that was not a good/helpful experience, so why even bother?
I thought tracking RHEL upstream is what Fedora was for?
CentOS Stream will be a much more reliable testing ground before a patch/fix finally makes into RHEL
This was predictable from the day Red Hat acquired CentOS almost exactly seven years ago.
As usual, just think about who pays whom and for what.
and then IBM buys Red Hat. you're right, how predictable. bunch of
IBM.....They did fabulous with OS2 Warp back around 1996. I have faith they will have the same success again. Not worried about IBM and you should not either.
Ibm has been on a 5 year slide - DOWN. After several peaks and drops, they're now $20 less a share then they were in 2016. So tell us again how IBM is doing "ok".
lol! steve I. has a great sense of humor. I was there during a most painful OS/2 Warp adoption. Full fail on IBM. I had many SMB that wouldn't go to IBM for help because of OS/2 Warp's debacle.
You should be, IBM basically buy stuff to cash cow it and run it into the ground by not spending money on development. Notes and Domino is a perfect example of that and their failure with SmartCloud, as well as leaving those customers who were silly enough to purchase SmartCloud in the lurch after selling the lot to HCL.
I Agree with you steve I.
Didn't know that CentOS is a commercial entity, acquired by Red Hat.
NO, I ABSOLUTELY THINK THIS IS THE RIGHT CALL. YOU'RE FREE LINUX IS NOT OUR PROBLEM. SO, YOU LIKE TO RUN 1 INSTANCE OF RHEL AND 9 OF CENTOS. YOU'RE STEALING MONEY FROM US. GO BUY SLES OR UBUNTU THEN!
Steal? This is GPL every thing in RHEL is based on gpl, are they stealing when they take that and charge money?
Stealing isn't the right work.
At the same time, given that RHEL is GPL dominantly, they can't override the license.
But the bigger crux is the public image. RHEL doesn't want to be projected as someone objecting to a RHEL derivative available for free. While, at the same time, they see such clones as a serious problem to their revenue stream.
They could have done it the other way, by obfuscating their changes, like the rest of the world does. But that'd have drawn them bad press.
So this move is an overall better alternate.
* Welcomes IBM employees to the discussion *
Creator first, IBMer later. Speaking for self, I feel being let down - a technocrat was supposed to be at the helm this time.
Simply said, Bye! Bye! CentOS, RHEL stil unwelcome under my server bonnet and welcome solid rock DEBIAN!
Lol this guy feels like all that money he's wasted is finally justified.
The fact that Linux IS FREE is YOUR problem!!!
IBM should stick to the business of giving support to a FREE OS, and not trying to own it.
you are scaring the hell out of me right now
In response to that may all the community then should some making packages for RHEL distro ?
i mean if you really want YOUR money...
In response to that may all the community then should stop making packages for RHEL distro ?
i mean if you really want YOUR money...
Who is the "us" in you're statement? Are you apart of Red Hat?
Keith, that's not the way it works.
Centos gets the "Red Hat Way" in front of new developers all over the world. Those developers, not the companies that hire them, will chose their platforms. Centos' binary compatibility with RHEL has been a critical factor in the success of Red Hat.
Every company I've ever worked in has paid for a license on every single prod server. Large enterprises tend to be hard-nosed about it, but even smaller companies see the benefit of support.
Development servers are a separate matter. While most companies I've worked at did pay for licenses on development servers, some chose to use Centos in development.
Red Hat's decision is short-sighted. It creates unnecessary ill will in the community, which means that the next generation of developers will look elsewhere.
If you can't figure out, this is what will happen: you will chase off developers on Centos first, then companies with production servers on RH as well (and vice versa). You shoot your own leg.
It will be OpenSolaris all over again.
It's more like 80 instances of RHEL and 120 of CentOS. But those 80 RHEL licenses are not windows licenses because the company could save money on the dev and test environments. Once those become expensive, companies are going to save on paying these Linux engineers and go all Microsoft.
Yea right... A few companies may find it worthwhile to shift to Windows, but the entire web is not going to shift to Windows just because of IBM/Red Hat. Other production-quality distributions exist. There will be more jumping distros than jumping RHEL-> Windows.
Even Microsoft has Linux on Azure! And saw value in offering WSL in Windows. And companies won't even save money on engineers as you will just be replacing with M$ Server Admins. Get real!
Your statement is absolute bollocks!!! CentOS has given more value to RH itself. Steal isn't a appropriate 'word' mate , you sound like a millinnieal from IBM . Its a opensource and GPL covers usage aspects.
First, you are writing in CAPS. It is no good.
Second, before IBM, Red Hat used CentOS to become a standard "de facto" in enterprise grade environment, increasing their revenues
Third, if you are thinking to money, you really don't understand the evolution of software, from free software to slave software.
https://www.change.org/p/centos-governing-board-do-not-destroy-centos-by-using-it-as-a-rhel-upstream Sign petition !!
I don't think it is a good way, even though win this time. We should do some thing beat back. Such as do some thing valuable for supporting Rocky Linux.
I wont sign that petition, I can't anymore trust on Red Hat, I dont care about the doomed CentOS, I will all of my will to Rocky Linux and we will learn the lesson: NEVER TRUST A PRIVATE CORPORATION BECAUSE THEY CAN BETRAY AN AGREEMENT!
Well that makes Debian the default choice of ubuntu chased me away with the snap stuff.
Cannot say it any better.
That move from CentOS people is good...
...For Canonical and SuSE! Yep, RHEL/IBM can make more money too.
The Red Hat Universal Base Image - A powerful tool for containerized application developers that provides a safer, more secure and free-of-charge redistributable container base image for creating containerized, cloud-native enterprise applications. With the Red Hat Universal Base Image, developers can more easily create certified applications for production deployment on RHEL and across Red Hat’s open hybrid cloud portfolio, including Red Hat OpenShift. It also enables compatible container images with other operating systems.
The RHEL Developer subscription - A free-of-charge, self-supported subscription for individuals (and soon, teams) who want to develop and test on Red Hat’s commercial, enterprise operating system product. It provides a development/test environment for applications that are meant to be deployed in production on the stable, secure, and high-performing foundation of RHEL. We also recommend that you join the Red Hat Developer Program. For partners, we recommend that you join the Red Hat Partner Connect program.
You think people will trust Red Hat for anything after it screwed over the community like this?
I think this is too harsh on Red Hat.
For the amount of work that the company does, they do need a steady revenue stream, to pay all those engineers.
By nurturing RHEL clones, it hurts their business model.
There are better companies on the list that screw the communities. I won't put Red Hat on that list.
If they were really worried about production systems using CentOS then they could have just kept developing it but made it dev only with RHEL required for any production system. Then again if there is a free development version of RHEL then fair enough.
Their business model has always been to sell support, not code. The problem was their level of support wasn't enough value added for a lot of people.
time to abandon the ship, nipah~
This is absolutely outrageous, this is not how things are done in communities. Major designs aren't just fall from "above".
I fully agree. It seems that IBM has torpedoed CentOS and this project is now sinking. It is sad. The future CentOS Stream feels like beta release of RHEL. Not for production and stable environments. We are looking now to switch to Ubuntu LTS or Debian. Good bye CentOS.
Keep in mind that Debian is a do-o-cracy distribution.
Bugs just don't get fixed automagically.
Debian is a very solid base but that base is solely based on the quality of "upstream first policy", which is something not all vendors like to follow all the times.
This is incredibly disappointing to see. This way I can have a constantly unstable system. Do they think this will mean MORE RHEL sales? Cuz I think it's a good way to make your customers migrate away. Amazon Linux 2 is free.
Whatever moron thought this was a good idea needs to be fired immediately. CentOS clearly doesn't understand their customers and users.
+1 This. I work at a Red Hat partner in Finland and even we get a lot of customers asking for CentOS (higher education, public organizations). We might end up upselling them to RHEL eventually but if they start moving straight to Ubuntu, we cannot even make an offer. RedHat is shooting its foot with a shotgun on this and in a small country might just shift things so much that entire ecosystems gets flushed out with it once enough time passes and adoptions rate plummets.
Some years ago when went through the hassle of switching all our workstations from OpenSuse to CentOS to have binary compatibility with our servers and eliminate some of the hassle, but it seems that all goes into the bin now. Bad move, RedHat, bad move. With stream being in front of the main RHEL version, binary compatibility just went out the window!!
This decision has forced tons of devs to start panicking overnight; just about everything has been built on CentOS, whole business models are based upon the cost of $0. I'm sure the thought was that RedHat and thus business-daddy IBM would reap the rewards of people moving to paid RHEL licenses. The reality is precisely the opposite. SUSE and Debian will be big winners here, not RedHat.
yes, was debating with my new cloud instances whether to go with centos 8 and I saw this... I'll not be deploying any instances with centos and do suse for anything new.
This is a Server OS not a beta testing OS looks like I will be looking for one of the new distros that will be made by how many people you have offended and businesses that can not accept this as a solution including my company.
Similar situation like Office Office acquired by Oracle.
Here IBM is playing the Game!!..
I have changed OS to my notebook when I know this message.
I agree. This is dumb. If I wanted that I would use Fedora.
+1 We have to move to Rocky
+1 this is dumb, stupid, and very on brand for 2020 but I thought CentOS was better than 2020
I don't even Linux. Still, I deeply feel for you guys. This is the dumbest decision.
Basically Red Hat (under IBM) will become another Amazon (AWS) in the sense that both companies love to take gpl software and milk the community without giving back to it - AWS is notorious for this, Red Hat not so much. But clearly things are shifting in that direction.
Hello to all.
This arbitrary measures from IBM/RH affect all Centos users, I have seen initiatives from the Centos founder and other to create more RH clones. In my humble opinion, Linux does not need more distros, what it needs is a very trusted and reliable Linux OS for web servers, truly open source that will not succumb to the temptation of being purchased by greedy corporations, that only cares for money not for users and a good OS. We need only ONE Outstanding, Reliable and Stable Web Server Operating System, just ONE. Please no more confusion and dispersion in Linux, divide and conquer is what we do not want nor need in the Open Source field, we are already plenty divided. Please let us unite efforts in just one platform. My 2 cents.
Thanks and Regards
Nothing to be added to this other than I agree to this statement in its ENTIRETY!!!!
why not freebsd ?
Did the community vote for this change? Other wise it means the CentOS project is not community driven and the entire premise of the CentOS project is a marketing lie.
I wonder how many others have picked up on this and perhaps if the community were involved maybe they really did vote overall for the change...
Well said, good point.
relax dill bag. You are still able to reap the benefits off of someone else's hard work.
The CentOS logo and board seems to have wider audience than spion logo and idea of the RedHat Enterprise Linux. The CentOS was believed to be supported by CERN. It is a pity that the solid confidence and 10 year support were shutdown. Though, It is still good distro for making your webserver or to use it for small servers.
it's time to move to AlmaLinux, an enterprise-grade server OS, an open-source RHEL fork.
So, just a thought... no sense in crying over split milk... bag it and move on.
Create "sentos" in the same way that centos was created. Unroll/repackage/deploy. Shake IBM/RedHat out.
A ceux qui ne sont pas contents : il existe Ubuntu Server et Debian, qui étaient déjà plus répandus que CentOS avant son abandon.
There isn't any sense. Why would anyone use the 'downgraded' CENTOS (CENTOS Stream), which is similar to Fedora that has been there for decades. They will certainly switch to Rocky Linux, ORACLE Linux, HP Linux, AlmaLinux...
I guess my argument "Use CentOS, not Ubuntu if you want most stable production" is out of the window now. Ubuntu it is then from now on.
Well since Ubuntu is just Debian with added fluff, why not stick to Debian proper?
Ubuntu LTS versions are supported for 5 years. Debian is usually supported about 3 years.
Debian has LTS and extended LTS nowadays, and is generally more stable and more reliable than *buntu.
Canonical also supports only a very small subset of packages for the very long time. (Debian ELTS is also limited, but it’s possible to sponsor those packages you actually need, so you have input. And they even sometimes take over backports, like Java 8 for jessie.)
This is more than enough time to even skip one major release.
Debian is also supported for 5 years and under Ubuntu only the "main" repo is supported, which are way less packages than supported in Debian
This is the dumbest move. If I want a stable environment that doesn't cost the earth, I look to CentOS. How much is it going to impact hobbyists and enthusiasts as well?
I disagree that there was any confusion around what “CentOS” means in the Linux distribution ecosystem.
Guess I need to start migrating to Ubuntu or something...
So Why not try FreeBSD?
What about GENTOO 😀
somebody will do another project just like centos only a matter of time
Hopefully an idea slightly less dumb than this announcement... could someone far better at scripting than me write a DNF plugin that tracks the RHEL source release repo and blocks any Stream binary packages that don't have a corresponding RHEL source RPM?
I place myself in the hobbyist / enthusiast category as I have my own domain for which I host at home 2 physical Centos 8 Servers - one is a mail & web-server, the other an FTP & VPN Server - granted it doesn't have the same level of importance as a commercial production server BUT I have shed blood, sweat & tears setting them up, tuning them and writing many, many scripts to get it to the level where I am satisfied. If I do go the Centos Stream route I guess I will have to take more regular offline "dd if=/dev/sd? .." images - especially prior to Kernel updates etc.
dude for LTS - SLES. You actually get more than you pay for.
Today Debian LTS is 5y (https://wiki.debian.org/LTS) and Ubuntu LTS is 5y and since 18.04 is extendable to 10y (but you have to pay for extension).
I don't think Ubuntu would be just Debian with added fluff.
A lot of work goes into making a distribution, integrating, testing, finding and fixing bugs.
And then supporting users. You don't have very informed knowledge about an Enterprise Linux Distribution.
what an unprecedented betrayal of a FOSS community. looks like my decades of using RHEL/CentOS are coming to an end because they simply can’t be trusted
Agreed. Seriously not cool, Red Hat. I've been a Red Hat user since Red Hat 4.1 (and that's before RHEL), and have been a dedicated RH/CentOS/Fedora user since then, and even steered my employer from using a random pile of distros to standardising on RHEL and CentOS.
I have just slammed the brakes on the planned upgrade to CentOS 8 as we have some CentOS 6 systems in production that need to be migrated. Looks like, for now, 7 it is, and see what Rocky Linux looks like.
*cough* OpenSUSE *cough*
+1! SUSE all the way.
+1 openSUSE Leap
Oracle has a converter script for CentOS 7, and they will sell you OS support after you run it:
It would be nice if Oracle would update that for CentOS 8.
OK, this sounds really stupid. If you want to pay for OS (and there are many good reasons why you could), then why pay somebody who just steals code, and not pay Red Hat itself?
The link says that you don't have to pay for Oracle Linux . So switching to it from CentOS 8 could be a very easy option.
It's Oracle. It's a trap.
You haven't heard of OEL
Oracle Linux is free. The only thing that costs money is support for it. I quote "Yes, we know that this is Oracle, but it's actually free. Seriously."
As Oracle Linux is based on CentOS, I'm sure it will be facing issues in 12 months or so...
Oracle already screwed me once with OpenSolaris, and then again with Java licensing. I'm not going there again.
I wouldn't trust oracle as far as you can throw them. Requiring a license for a Java Runtime in a corporate environment, where the corporation isn't asking for support from the vendor, is rediculous.
there is no steal in OpenSource
this quick n'dirty hack worked fine to convert centos 8 to oracle linux 8, ymmv:
rpm -e centos-linux-release --nodeps
dnf --disablerepo='*' localinstall ./*rpm
dnf remove centos-linux-repos
dnf --refresh distro-sync
# since I wanted to try out the unbreakable enterprise kernel:
dnf install kernel-uek
dnf remove kernel
great, I wil give it a try
I've installed it on a VM to have a look around. Yeah, it works, and despite the warnings you need to register to update, that isn't actually the case and 'yum update' works just fine.
However, the minus points are that there is no equivalent to the CentOS-Powertools repo, and their build of the kernel packages are such that the VirtualBox Guest Additions only builds on their UEK kernel, not the RHCK (Red Hat Compatible Kernel), as it's missing a header file in the build. Only Oracle would mess it up such that their own product fails.
The OL-equivalent is ol8_codeready_builder
I use Ubuntu LTS in my clients and I not have any problem. I like CentOS but after this is not a option anymore.
Same here. Ubuntu is the way to go!!!
No, you still can go openSUSE Leap. It will get even closer to the the SLE stream. Next to this, Tumbleweed as tested rolling release is for those who cant wait to get the latest software.
Against my better judgement... *looks at Oracle Enterprise Linux*
I suspect OEL will change to streams as well since they just repackage CentOS and toss in their own kernel. I don't think oracle has any incentive to build a distro on their own. Look at their history with MySQL or OpenSolaris.
IBM is messing up RedHat after the take over last year. This is the most unfortunate news to the Free Open-Source community. Companies have been using CentOS as a testing bed before committing to purchase RHEL subscription licenses. We need to rethink before rolling out RedHat/CentOS 8 training in our Centre.
You can use Oracle Linux in exactly the same way as you did CentOS except that you have the option of buying support without resinstalling a "commercial" variant. Everything's in the public repos except a few addons like ksplice. You don't even have to go through the e-delivery to download the ISOs any more, they're all linked from yum.oracle.com
They rebuild RHEL, not CentOS. RH won't stop publishing RHEL's source codes. CentOS just stops to be the RHEL's rebuild. OL will continue to be that.
Not likely. Oracle Linux has extensive use by paying Oracle customers as a host OS for their database software and in general purposes for Oracle Cloud Infrastructure. Oracle customers would be even less thrilled about Streams than CentOS users. I hate to admit it, but Oracle has the opportunity to take a significant chunk of the CentOS user base if they don't do anything Oracle-ish, myself included.
I'll be pretty surprised if they don't completely destroy their own windfall opportunity, though.
all you are doing is shifting the licensing and support fees from IBM to Oracle. OEL is literally a rebranded RH.
"OEL is literally a rebranded RH."
So, what's not to like? I also was under the impression that OEL was a paid offering, but apparently this is wrong - https://www.oracle.com/ar/a/ocom/docs/linux/oracle-linux-ds-1985973.pdf - "Oracle Linux is easy to download and completely free to use, distribute, and update."
That's entirely correct. There is no charge for OEL. Oracle only charges for support (if you want it).
For now. Oracle is like Needful Things: You *will* regret it at some point down the line.
"For now" == Since 2012 (when public repos were made freely available). Is that not long enough?
Um, yeah, and how long was CentOS free? How about Oracle Java? Putting your eggs in the Oracle basket is a recipe for disaster.
Oracle only publish the last version of the packages in their repository.
If you want to install a previous package revision or not so new package it isn't possible.
If you pay support to Oracle you obtain access to the full repository.
This is a restriction that I found. May be there are other restrictions.
I can live with that because for work purposes maintaining a local mirror of everything you install is something we do anyway as best practice. Mostly because EPEL only carries the lastest version of any RPM already.
Not really true. First, if you take a look at the repo for OL8 Latest, you'll see a bunch of different versions of packages. Second, there are repos for BaseOS GA 8.1, 8.2 etc with the packages that shipped on the install media.
No, all the RPM versions available are provided on the OL yum server, so they're freely accessible.
Of course, you should still run latest RPMs as much as possible for security purposes, whether you're running on your own or do buy support.
[Full disclosure: I work on the OL product.]
Todd says that all rpm are available in the OL yum server.
He must be in the true.
8-10 years ago wit a CentOS6 I couldn't find a old rpm, for update an older Oracle Linux. When looking in the repository with contract I couldn't find without problems.
May be I didn't look in the archvied channel, and different structure of public an privte repositories.
Any way, Oracle Linux is a easy alternative for migrate from CentOS.
I won't choose Oracle if it is critical for your Business if other solutions as Debian/Ubuntu /Suse are available. Oracle isn't a reliable partner: changes licensing models: OpenSolaris, doesn't support RAC in new versions of Oracle Database Standard Edition, and others.
If a change of license by Oracle side doesn't break your business Oracle Linux is a good alternative to CentOS. If this is not the case try to choose other distribution.
I never thought we'd see the day Oracle is more trustworthy than RedHat/IBM. But I guess such things do happen with time...
So, what's the problem?
IBM has discontinued CentOS. Oracle is producing a working replacement for CentOS.
If, at some point, Oracle attacks their product's users in the way IBM has here, then one can move to Debian, but for now, it's a working solution, as CentOS no longer is.
Because it's a trust issue. RedHat has lost trust. Oracle never had it in the first place.
Friends don't let friends use Oracle software. 🙂 (you'll regret it.)
If you need free enterprise Ubuntu or maybe SuSE are good choices now.
The issue is many already have CentOS 8 environment and converting to something like OL is much easier than switching to Ubuntu/SuSE which would require a lot of work.
And when Oracle follows suit, is that when you cough up the cash?
Remember Oracle is the company that literally -stole- OpenSolaris and closed it's source. That is shameless and if you believe in Open Source you can't forgive that.
The goal is not to stay on OL. The goal is to tell management, hey OL 8 is an option so we can continue our product cycle and buy time.
Then by then we might have other alternatives like CloudLinux Project, Rocky Linux and etc.
But you gotta understand, telling management that you wish to switch to a new distribution that will come out in a few months/year isn't gonna fly.
Neither is telling them we are dumping all our development for the past year to switch to ubuntu/suse.
Dev on SuSE should just run, it's not that different from RedHat.
It's the automation that is harder but it isn't that hard. SLES is pretty nice all an all, I worked in a SuSE shop back in the Novell days.
It's up to you man if you want to go to bat with management but vendor lockin is painful and Oracle will lock you in.
Remember what they did with Java...
It's just OL/Oracle Linux these days (since 2010). That's okay though, it's a common misconception.
[Full disclosure: I work on the OL product. Not doing marketing here as that would be inappropriate, just fixing a naming nitpick.]
I have been using CentOS for over 10 years and one of the things I loved about it was how stable it has been. Now, instead of being a stable release, it is changing to the beta testing ground for RHEL 8. And instead of 10 years of a support you need to update to the latest dot release. This has me, very concerned.
well, 10 years - have you ever contributed with anything for the CentOS community, or paid them a wage or at least donated some decent hardware for development or maybe just being parasite all the time and now are you surprised that someone has to buy it's your own lunches for a change?
- if you think you might have done it even better why not take RH sources and make your own FreeRHos whatever distro, then support, maintain and patch all the subsequent versions for free?
That's ridiculous. RHEL has benefitted from the free testing and corner case usage of CentOS users and made money hand-over-fist on RHEL. Shed no tears for using CentOS for free. That is the benefit of opening the core of your product.
Years ago when CentOS was struggling, I sent them $200 since I was using it extensively in our lab at that time.
You are missing a very important point. Goal of CentOS project was to rebuild RHEL, nothing else. If money was the problem, they could have asked for donations and it would be clear is there can be financial support for rebuild or not. Putting entire community in front of done deal is disheartening and no one will trust Red Hat that they are pro-community, not to mention Red Hat employees that sit in CentOS board, who can trust their integrity after this fiasco?
Sieciowski, You're in the wrong place ...
The Microsoft forum is https://techcommunity.microsoft.com/.
It's the other way round.
RedHat is selling something which is made by others for free.
Read the GPL license as a start
Many of us have tried to be part of the CentOS community and do donate servers, mirror resources, and employee time toward making CentOS better. We get screwed just as much as the rest of the community.
terrible move by IBM/RH - the community is what has driven the success of RHEL Enterprise ... and this is damaging off the upstream feed.
generally speaking, centos 8 will
time to think about another LTS distro
This is disappointing and frustrating. If I wanted something that tracked ahead of RHEL, I would use Fedora. I use CentOS because it tracks after RHEL, with all of the benefits thereof.
That said, how is CentOS Stream different from Fedora since they both track ahead of RHEL and therefore seem to meet the same need, and how does it provide value for us, the end users?
Unlike CentOS, Fedora doesn't have to make decisions about what to ship or not ship based on Red Hat's desire to support things.
btrfs, *-devel, and samba-dc immediately come to mind.
Time to move over to Ubuntu LTS then.
How does anyone know that Canonical would not be acquired by someone and change their models of Ubuntu. It is pretty well known they have been trying to sell the company for years. I feel like depending on a free, and unsupported operating system software does not buy companies very much. the coast of a paid and supported RHEL or Suse is cheaper than the potential ris, I think this change makes people assess if it is smart to keep trying to do this approach.
This is a breach of trust from the already published timeline of CentOS 8 where the EOL was May 2029. One year's notice for such a massive change is unacceptable.
Move this approach to CentOS 9
This! People already started deploying CentOS 8 with the expectation of 10 years of updates. - Even a migration to RHEL 8 would imply completely reprovisioning the systems which is a big ask for systems deployed in the field.
Well, you get a rolling beta-testing distro for another 9 years. Exactly what RH/IBM wants to bump profits. It's no longer identical to their product, it's not stable enough to have same level of the enterprise support. Therefore many CentOS users have to become RHEL customers or move to another emerging clone (which will struggle with lack of support initially).
This is unbelievable to me - and is going to be damaging to the Open Source Community at large. Published LTS dates should never be changed so dramatically with such little notice.
I'm embarrassed for those that made this decision. You turned your back on the community you currently claim to help.
This. Up to today, we were at the "we're sunsetting CentOS7, new machines will be CentOS8" stage, still ironing out issues with Application Streams etc. With CentOS8's support now much shorter than CentOS7's, we will probably halt and, who knows, probably consolidate on Ubuntu LTS.
Very much this. We already have many CentOS 8 hosts in production and now we have to migrate them to CentOS 7 as it has the longest support.
This is an absolute betrayal of trust. Shameful.
Yup, if they did this with RHEL9, no problem. But cutting the EOL of an existing product by 8 years?
Burning this trust is going to hurt RH much more than they might guess.
Definitely, cutting a live and on going production is litter ally a backstab to the community, they should have wait for RHEL 9 and announce it years before doing so,
Now this is just gonna pull people as far as possible from RH
+1000 to this...they are entitled to change direction for a future release (still, I don't agree with it), but changing EOL of an existing product by 8 years is simply unheard of
This is a PR nightmare for RedHat
So in other words, you're making it into Fedora.
This does raise the question of what's going to happen to Fedora. "Fedora has a valuable place as the proving ground for RHEL" doesn't sit well alongside "we're introducing CentOS streams as the proving ground for RHEL".
When CentOS Stream was first talked about, they published this - https://fedoramagazine.org/fedora-and-centos-stream/
Basically, Fedora => CentOS Stream => RHEL
My trust in Oracle is not enough for me to invest time and effort in Oracle Linux. Very sad about this news...!
The announcement means that you'll have to invest time and effort in something. CentOS is effectively being cancelled; the brand will remain, but there's no product comparable to what people thought of as CentOS before.
My office switched the bulk of our RHEL to OL years ago, and find it a great product, and great support, and only needing to get support for systems we actually want support on. Oracle provided scripts to convert EL5, EL6, and EL7 systems, and was able to convert some EL4 systems I still have running. (Its a matter of going through the list of installed packages, use 'rpm -e --justdb' to remove the package from the rpmdb, and re-installing the package (without dependencies) from the OL ISO.)
I am considering creating another rebuild of RHEL and may even be able to hire some people for this effort. If you are interested in helping, please join the HPCng slack (link on the website hpcng.org).
(original founder of CentOS)
Not a programmer, but I'd certainly use it. I hope you get it off the ground.
This sounds like a great idea and getting control away from corporate entities like IBM would be helpful. Have you considered reviving the Scientific Linux project?
Feel free to contact me. I'm a long time RH user (since pre-RHEL when it was RHL) in both server and desktop environments. I've built and maintained some RPMs for some private projects that used CentOS as foundation. I can contribute compute and storage resources. I can program in a few different languages.
Thank you for considering starting another RHEL rebuild. If and when you do, please consider making your new website a Brave Verified Content Creator. I earn a little bit of money every month using the Brave browser, and I end up donating it to Wikipedia every month because there are so few Brave Verified websites. The verification process is free, and takes about 15 to 30 minutes. I believe that the Brave browser now has more than 8 million users.
Wikipedia. The so called organization that get tons of money from tech oligarchs and yet the whine about we need money and support? (If you don't believe me just check their biggest donors) also they keen to be insanely biased and allow to write on their web whoever pays the most... Seriously, find other organisation to donate your money
Please keep us updated. I can't donate much, but I'm sure many would love to donate to this cause.
sounds great, thanks
Not sure what I could do but I will keep an eye out things I could help with. This change to CentOS really pisses me off as I have stood up 2 CentOS servers for my works production environment in the last year.
I would certainly appreciate and use that! Frankly I saw this day coming since RH bought CentOS but I think there is a need for a bug-for-bug compatible EL version.
Please, work together on that, if possible:
1 how long tell you get the 1st version released
2 do you need funding if so how to help.
This is the project he is working on.
This announcement is very difficult to distinguish from just saying: "CentOS is being abolished. In future, you can beta-test RHEL, and we'll be using the 'CentOS' branding to describe that experience."
If there's a difference between what I've written and what the post above actually says, I'd be very happy to understand it.
That's pretty much how I read it, too. I didn't realize the CentOS name had cachet among the bleeding-edge bunch. Seems a strange application to me.
Why make CentOS Stream the primary innovation hub for the RHEL ecosystem when you already have Fedora for that. CentOS has always been a fully open sourced rebuild of the stable enterprise-ready RHEL and this is what drives such high demand for it. If it's not as production ready as RHEL then what's the reason for using it going forward? People will just switch to Debian.
Absolute terrible decision, abandoning what made CentOS a good alternative to RHEL.
LOL... CentOS is RH from 2014 to date. What you expected? As long as CentOS is so good and stable, that cuts some of RHEL sales... RH and now IBM just think of profit. It was expected, search the net for comments back in 2014.
IT's not like we weren't expecting it, but at the very least we thought they had the decency of maintaining their commitments.
So if CentOS 9 didn't exist, okay fine. We even thought CentOS 8 might not exist but it was promised and it had a promised commitment date.
What no one expected was for IBM/RH not to at least honor their commitment to LTS. That isn't professional at all. Who wants to deal with a company who can pull out their commitments whenever they feel like it?
Why? What's the point of a stable enterprise distro when you are going to make it a rolling release? Thank you for ruining CentOS. You're awesome!
Thanks for your hard work over the years.
For what reasons you re-invent fedora?
My entire infrastructure has to change now. GOOD LORD.
*HALTS MIGRATION OF EVERYTHING FROM C7 to C8*
Here the same. And no Rhel is not an option.
Same here. product was code complete in November and we were letting customers beta our centos8 version.
I have absolutely no idea what I am going to do now. It would take me 6-8 months full time to move to something else, and they sure as heck arent going to pay for RHEL licenses.
The fact that I am even considering OEL is making me nauseous.
Yup. This exactly. Now we get to hit the brakes and scramble for a bit.
Yeah, I just saw this and we're halting as well.
Canonical probably loves this decision... mostly because you're going to push people over to using Debian and by extension Ubuntu with this decision.
Increasing the market share of the .Deb branch of the Linux world and future software compatibility for those operating systems.
Meanwhile you've horrible hurt the stability of future software releases on RedHat by cutting off developers from a free alternative to RedHat Linux that was up to this point 100% binary compatible.
Now if I want to build my app it might work on CentOS, but as soon as it's installed onto RedHat. It crashes thanks to needing an updated package that's not on RedHat yet.
Also, perhaps it's time to update your business model and streamline it to match the rest of the Linux world.
After using a rolling release distro for years on my personal systems, I am very happy hear CentOS is going full rolling release. I will have to say organizationally, some of the old school people I work with (mostly in management) are having a hard time how to handle our image build and release cycles. I would appreciate a blog post in the future for what to expect for image release cycles (datacenter and cloud).
This isn't just rolling, this is basically rolling beta. Not thrilled.
Keyword "personal distro". If a doctor was using a computer to do surgery on you, would you be okay with a rolling release on that machine?
CentOS/RHEL is for enterprise usage and requires higher reliability. If you want consumer grade, that is what Fedora is for.
But you are confused about something, there is no "rolling release". You still have up upgrade from CentOS Stream 8 to CentOS Stream 9! All that is changing is now you'll be getting less stable code.
I also want good shit for free and am mad that the party is over.
Yes, I too love when IBM/Red Hat makes a community distro into just another beta testing distro that will barely work on servers. We all should buy RHEL and shut up because the community ain't shit.
Yes, lets release a roadmap for Centos 8 support, get a ton of the community to begin migrating over to it, only to have the rug pulled from under them. Beautiful move.
This here is the real issue.
While I vehemently disagree with the direction CentOS is taking here, they are within their rights to do so. It's free stuff, who are we to demand things?
However, bringing out a new distribution (CentOS 8), marketed with support for as long as RHEL (2029?), then mid-release changing that to end in 2022 is an absolute shambles.
We have already started deploying CentOS 8 in the field expecting it to be supported for many years to come, now we're going to have to put a halt on that and figure out what we're doing next (and no, paying for a RHEL license is not an option).
Genuine thanks for the all the work over the years, Team CentOS, but this move is awful. It should have happened with CentOS 9.
>Who are we to demand things?
The community. It's FOSS not a corporation. You have every right to shit on them for bad decisions and to have a hand in its governance.
Let's not forget that Red Hat is getting a ton of "good shit for free" also.
Like kernel for the start...
You do know that Red Hat re-invests half of it's revenue into open source projects, such as a the kernel. There are few free rides when it comes to Red Hat and open source - before they contribute and help the projects.
This is a pretty clear indication that you people are completely out of touch with your users.
Bad idea. The whole point of using CentOS is it's an exact binary-compatiple rebuild of RHEL. With this decision RH is killing CentOS and inviting to create a new *fork* or use another distribution. Do you realize how much market share you will be losing and how much chaos you will be creating with this?
"If you are using CentOS Linux 8 in a production environment, and are concerned that CentOS Stream will not meet your needs, we encourage you to contact Red Hat about options". So this is the way RH is telling us they don't want anyone to use CentOS anymore and switch to RHEL?
That's exactly what they're saying. We all knew from the moment IBM bought Redhat that we were on borrowed time. IBM will do everything they can to push people to RHEL even if that includes destroying a great community project like CentOS.
"it removes confusion around what “CentOS” means in the Linux distribution ecosystem"
My level of confusion regarding CentOS just went from quite low to fairly high. As a long-timer, this seems like a sad development.
Yeah, excatly. CentOS was never confusing to anyone in the Linux ecosystem, even new users understood it. Now though, what the heck.
Not unprecedented - look at Red Hat 9 and the start of Red Hat Enterprise. Fedora was spun off at that point as a way to retain developers interested in FLOSS.
See also discussion on LWN (lwn.net)
Agree completely. To go back on that LTS date is a terrible idea. Ton's of shop's have spent significant time uplifting their 6/7 environments to 8 and then this gets dropped. Unbelievable...
Debian / Ubuntu it is then !
Well now I regret deploying CentOS 8 last week. Can't wait to re-do all that in a year...
Anyway thanks for all the great work in the past. Sad to see CentOS die like this though. Was a great ride while it lasted.
What's the point of Fedora now? Why do this? That's really bad for everyone that invested time on Enterprise Linux (EL) derivatives.
If RHEL would be like RHL was in the late 90's and early 00's I'm fine with it, but I don't see RHEL being free without support anymore.
This is just killing CentOS. CentOS have the true winner in clones war, Scientific Linux, PUIAS, and others that I don't even remember the names converged to CentOS.
So our hope now it that the CERN folks continue the game with Fermilab and reconsider Scientific Linux 8 to be back.
What a disappointment with this move from Red Hat.
It's not something that's *going to* kill CentOS. It is precisely that. That's what they've announced. The fact that they're re-using the CentOS brand for something different doesn't really mean anything.
What? Why? Who came up with this idea? Fire him/her. You just shot yourself in the leg.
I assume 2021s mor relevans search will be: How to migrate from CentOS to Ubuntu...
First CoreOS, now CentOS.
It's about time to switch to one of the *BSDs.
FreeBSD makes a dammed good server if you can swing it past IT.
Unlike RedHat they do backports and are always 100% API backwards compatible so you actually *can* upgrade it without breaking your applications.
That is false. It seems you do not understand what RHEL is. At day 0 of a RHEL major release, software is selected and then, for the next 10-12 years, Red Hat does backporting for the complete ABI. Significant work. No other Linux distro comes close.
Read more here: https://access.redhat.com/articles/rhel8-abi-compatibility
No, everyone knows you can't go from a major version of RHEL to the next without a re-deploy.
It almost NEVER updates cleanly without breaking your custom apps on top. Only FreeBSD (and Debian sometimes) seem to get this right.
You're right that they do it on the major versions.. but not between them.
FreeBSD 4 is API compatible with 12.
I'm really glad that I haven't upgraded to CentOS 8, I was considering it for 2021. I will be looking for a new Enterprise OS to replace CentOS 7 with now. Feels like a lucky escape.
Wow. Well, I guess that means the tens of thousands of cores of research compute I manage at a large University will be migrating to Debian. I’ve just started preparing to shift from Scientific Linux 7 to CentOS due to SL being discontinued by 2024. Glad I’ve only just started - not much work to throw away.
Not very often does one see such a respectable brand like CentOS/RH turn its back to its user-base in such a way.
We migrate to Debian with thousands of servers.
Only in 2020...
I'm going to contact Red Hat and ask for options, for sure. Who can I flog, and what can i flog them with?
Well thanks for a whole lot of nothing. I spent the last week setting several servers up with centOS8 expecting 10 years. Now I'm on a rolling release. Well if I'm going to be on a rolling release, I guess I'm using archlinux for everything like on my personal computer. A big thank you for not following through on our listed commitments that people have made decisions on. No faith in cent os.
Not a rolling release. A rolling pre-release.
This is ridiculous. Completely destroying the entire point of CentOS for the majority of us. Sign this petition please: https://sprsrv.cc/VlGNQ
I've signed, but the bond of trust relating to EOL dates has been broken, so I'm really not sure I would want CentOS 8 now even if RedHat back down. For me the same would now apply to purchasing RHEL if I were inclined to do so. Who is to say they won't change that as well?
Red Hat did change. Years ago, when your subscriptions expired, you could continue to use RHEL just without the updates. Now if your subscription expires, you have to decommission the server. You can no longer continue to use it when the subscription expires.
This is just a money-grab to move people onto RHEL paying subscriptions.
My hot take is much different than the rest of what I'm seeing here. Fedora has always been the fast-moving active development distro, with RHEL/CentOS as the stable LTS. This mostly just gives the community at large more access to the process by which the latter is created.
It sounds like the community will have more insight into, and potentially more influence over, that process. We'll see how RHEL is maintained, and how new features that were *developed* in Fedora or other upstream projects are *integrated* into RHEL. Previously that has been mostly invisible magic unless you worked at Red Hat or a few big customers or partners.
I miss this kind of access supporting some of my current customers, to that could be nice. It might actually give me a reason to use CentOS instead of just my developer access to RHEL. For people who were not after that kind of visibility and participation, I guess it's less exciting.
If you just want a stable, RHEL-compatible OS to use as a development target, platform for non-production lab systems, and such, you can still get $0 access to genuine, branded RHEL through their developer program. Log in to developer.redhat.com, check out the terms, and you'll probably be fine. I use this to work on stuff that will eventually be deployed to production on paid RHEL all the time. CentOS never quite exactly perfectly matched the names of repositories and such anyway.
If you want to run production servers with the stability of RHEL and the price tag of $0, this is probably a bigger issue. But hey, Red Hat has always given us *more and better* access to source than the various licenses require. They don't owe us pre-built binaries. Heck, they don't owe source that exactly matches the commercial release to anyone who didn't get the commercial release. So, if they change the approach to CentOS in a way that favors people who want to participate at the expense of people who want to free ride, it's hard to hold that against them.
And you know if it gets too hard to free-ride, someone will start another under-resourced rebuild project that unpredictably fails to release any updates if certain key people are vacationing. Just like CentOS used to be before Red Hat started pouring resources into it.
You're correct to say that RH don't owe anyone pre-built binaries. That's why CentOS began - other people put in the work to turn source into binaries and creating a community around it. They succeeded. RH then took over CentOS with the promise that they would keep it going....
"Heck, they don't owe source that exactly matches the commercial release to anyone who didn't get the commercial release."
Correct me if I am wrong, but by the very nature of open source, they actually do owe us, the community, access to the binaries. We may have to compile it and removed trademarked items like logos, but the source code is "free."
Read the terms. The developer program is not even signable for a non company. And the usage of RHEL is allowed purely for development. Have fun with the audit rights of Red Hat.
What about all the personal and non profit users of CentOS out there? If you take the terms halfway serious, this is no option at all.
Here's my hot take, I don't want to have to give Red Hat my email address and my name to download this distro. Fuck that.
IBM is declining, hence they need more profit from "useless" product line. So disgusting
Looks like I need to move my machines to another distribution. If that will come true centos will be not useable any more
- Was I good server OS?
- Yes... You were...
This is such a terrible idea I don't know where to start. The stability is going to go away immediately. Rolling releases are cute for your laptop but managing thousands of servers it becomes a nightmare. You thought your systems has package drift between minor version releases just wait until you get into the rolling releases. Years ago when RH released their intention to charge I had to pivot from RH 9 into Slackware to save my small ISP. CentOS has copied RH so much that now it has come full circle.
Slackware -> free all the way
An entire team worked for months on a centos8 transition at the uni I work at. I assume a small portion can be salvaged but reading this it seems most of it will simply go out the window. Does anyone know if this decision of dumping centos8 is final?
"This is a Community mantained site. Red Hat, Inc is not responsible for its content."
Well, is it?
I jumped to centos back with the redhat / fedora split around 03. This is insane. If I wanted fedora, I would use fedora. Guess I need to move to a better LTS distro.
Unless the community can center on a new single proper fork of RHEL, it makes the most sense (to me) to seek refuge in Debian as it is quite close to CentOS in stability terms. Already existing functioning distribution echosystem, can probably do good with influx of resources to enhance the missing bits, such as further improving SELinux support and expanding Debian security team.
I say this without any official or unofficial involvement with the Debian project, other than being a user.
And we have just launched hundred of Centos 8 servers.
What's wrong with Ubuntu LTS. I never had any instability issues
You're dependent on Canonical rather than a FOSS project explicitly not dominated by a single corporation.
This is a clear business strategy, I will return to Debian.
Another one bites the dust due to corporate greed, which IBM exemplifies. This is why I shuddered when they bought RH. There is nothing that IBM touches that gets better, other than the bottom line of their suits!
This blog casually ignores the #1 reason for running CentOS: It's a stable clone of RHEL. I can't tell if this is an intentional omission or what, but it's tone-deaf and out of touch.
Few sites use CentOS because it's a development fork of RHEL. The vast majority, I'd guess over 75%, use it because it's a solid, enterprise-quality OS with great community support.
Some detractors will argue that CentOS users have been leeching off of RHEL this entire time, but this ignores the history of CentOS. The entire mission of CentOS for a good 15 years was to be a solid clone of RHEL. There was an uneasy relationship between CentOS & RHEL for a while, but eventually, RHEL embraced CentOS as a valid sibling.
Dude, is it a joke?
Are we in April fool?
This is the perfect ending of this terrible 2020.
My compliments, I can throw my last 10 years of work in the toilet.
Absolutely insane. To not have an exact binary compatible community rebuild means also getting much more lesser fixes and hints for improvement for RHEL. Sad news for the community, sad news for RHEL!
Eh. CentOS is not exact binary compatible with RHEL. AFAIK, CentOS Stream is closer to RHEL than CentOS Linux is.
This is a big mistake. RedHat did this with RedHat Linux 9 the market leading Linux and created Fedora, now an also-ran to Ubuntu. I spent a lot of time during Covid to convert from earlier versions to 8, and now will have to review that work with my customer.
I just finished building a CentOS 8 web server, worked out all the nooks and crannies and was very satisfied with the result. Now I have to do everything from scratch? The reason why I chose this release was that every website and its brother were giving a 2029 EOL. Changing that is the worst betrayal of trust possible for the CentOS community. It's unbelievable.
"And it removes confusion around what “CentOS” means in the Linux distribution ecosystem.*
Pretty sure no one's confused about CentOS till this announcement.
RIP Community Enterprise OS.. just when you think 2020 can't get weirder.
RIP CentOS. Congratulations RH/IBM, a nice betrayal from you to end this 2020
What a colossal blunder: a pivot from the long-standing mission of an OS providing stability, to an unstable development platform, in a manner that betrays its current users. They should remove the "C" from CentOS because it no longer has any connection to a community effort. I wonder if this is a move calculated to drive people from a free near clone of RHEL to a paid RHEL subscription? More likely to drive people entirely out of the RHEL ecosystem.
Before this absurd turning off stable Centos (and betrayal of trust of IBM/RH) RHEL/Centos community was big in corporate area.
Short sighted decision of IBM managers ..(
> "When CentOS Linux 8 (the rebuild of RHEL8) ends, your best option will be to migrate to CentOS Stream 8, which is a small delta from CentOS Linux 8, and has regular updates like traditional CentOS Linux releases."
But if you read the FAQ, you also learn that once they start work on RHEL 9, CentOS Stream 8 ceases to exist...
Not sure where it says that?
That one says CentOS Stream 8 matches RHEL 8 support dates.
I'm not a fan of this move at all, but I guess only time will tell how well CentOS Stream works or doesn't.
Not that set of FAQs, this one:
Q6 - "Will there be separate/parallel/simultaneous streams for 8, 9, 10, etc?"
Re-reading it, I overstated; the prior stream won't cease as soon as work starts on the next: they anticipate "an overlap of approximately 2 years between one stream and the next".
i.e. If you want to go down the Stream route, you'll be the upgrading servers running it every 2 years.
To be fair, it's a rolling release-- you'll be "upgrading servers" on a regular basis.
For my personal system, this is fine.
For researcher desktops who need latest software without a major rebuild, this could be a benefit-- but when SuSE switched to a rolling release, they tended to forget things like VMWare drivers and NVidia drivers, rendering Leap useless for desktops and VM's.
I don't know if that changed-- I left SuSE as a result.
Seems early for April Fool's Day
Remember, you're still in 2020!
This makes CentOS completely irrelevant for running production workloads now so after just completing the migration of some rather large platforms to CentOS 8 from 7 this comes as an unpleasant surprise, time to start planning the move to Debian or Ubuntu.
Amen! I was just starting to wonder when RH will kill CentOS... I was starting to think I was wrong back in 2014. Remember that "CentOS announced the official joining with Red Hat while staying independent from RHEL"... LMAO, many were stupid enough falling for that. The same happened when IBM came into focus... "IBM would make RH stronger", maybe stonger in greed.
The money move the world... during this pandemic increace centos system and down rhel... lost money a now "hey guys destroy centOS and i will paid $$$$$"
No they won't. People will just switch to something else such as Debian or Ubuntu...
If the community feels this strongly, they should let RedHat and IBM management know. Write letters and make calls (emails don't work these days). Write articles. When some high-profile organizations migrate away from RHEL/CentOS/Fedora they might wake up. Of course IBM has a long history of such miss-steps....
From a RHEL perspective I understand why they'd want it this way. CentOS was probably cutting deep into potential RedHat license sales. Though why or how RedHat would have a say in how CentOS is being run in the first place is.. troubling.
From a CentOS perspective you may as well just take the project out back and close it now. If people wanted to run beta-test tier RHEL they'd run Fedora. "LATER SECURITY FIXES AND UNTESTED 'FEATURES'?! SIGN ME UP!" -nobody
I'll probably run CentOS 7 until the end and then swap over to Debian when support starts hurting me. What a pain.
Sorry, but saying like it's all abount money is weird.
I've reported *RHEL* bug because I found it on CentOS! Right now, I should get paid from Red Hat for that since this is not more community, and it's only about money. And all other people too.
Red Hat may earn more or less money on RHEL, but let's be honest - people using CentOS reported stuff to Red Hat on regular basis so Enterprise Customers had better software they never asked for (in my case I found quite stupid bug in NTP). Why would they can benefit from money, and other contributors *to RHEL* can't?
This is more than unfair right now.
Personally I won't switch to Debian because it hurts me when I need to use it (too much wierdness on it), but I use it at current company. But Personally i'll stick to CentOS Stream because I have no better choice so far (I don't really trust Oracle to migrate to Oracle Linux).
Don't trust Red Hat. 1 year ago Red Hat's CTO Chris Wright agreed in an interview: 'Old school CentOS isn't going anywhere. Stream is available in parallel with the existing CentOS builds. In other words, "nothing changes for current users of CentOS."' https://www.zdnet.com/article/red-hat-introduces-rolling-release-centos-stream/
I'm a current user of old school CentOS, so keep your promise, Mr CTO.
What is this? Another 2020 joke?
Goodbye CentOS. Was nice knowing you. Time to look for another distribution, since you guys obviously decided to abandon the users you had for... whatever reason.
Oh, right. The corporate overlords or whatever.
This a very, very dk move.
There is nothing wrong with Centos-Streams, in fact it is a fine project that brings more community involvement into the RHEL release process.
The issue is that support was promised for Centos 8 through 2029, and now that will not happen. This is a breach of trust between the community and Redhat that can never be restored.
Even if they reverse this decision, will anyone trust it? I think not.
It's done. The End. Just walk away. Centos is no more, find an alternative.
Yup that's the killer for me.
Changing an EOL to 8 years earlier is an awful move for a company like RedHat. Trust = destroyed.
The same might happen to CentOS 7
Yes please stand by your word. Going back on the 2029 EOL date like this is a horrible decision. It should have been implemented at the beginning of RHEL 9 _OR_ at least given a few years minimum before switching (which would give users ample time to find an alternative and/or some group to spin up a new community EL OS based on RHEL source code).
IBM? Is that you?
That was quick:
"Old school CentOS isn't going anywhere. Stream is available in parallel with the existing CentOS builds. In other words, "nothing changes for current users of CentOS."
From the same article:
'To be exact, CentOS Stream is an upstream development platform for
ecosystem developers. It will be updated several times a day. This is
not a production operating system. It's purely a developer's distro.'
Read again: CentOS Stream is not a production operating system. 'Nuff said.
We just got all of our centos6 users migrated to Centos7 and Centos 8 just to NOW read this. And I saw none of this covered in your emails.
Can's say as this has been a good call for me at all.
Same! We just finished migrating 4 large applications from CentOS7 & RHEL7 to 8, I was literally days away from ordering the servers and getting licensing in place for the production servers.
Just one word: FAIL ....
Too bad about the 3rd party software support for software on RHEL from companies that dont want and dont have a subscription and only advertise RHEL support because they can test on Centos. We'll see a bunch of software with multiple OS targets drop RHEL now, I think.
This is a bad idea, If this continues I'll have to find another distro to use. I used CentOS because it tracks just behind RHEL. I feel like you're going to lose a lot of the community with this shift.
This makes my decision to go with Ansible and CentOS 8 in our enterprise simple. Nope, time to got with Puppet or Chef. IBM did what I thought they would screw up Red Hat. My company is dumping IBM software everywhere - this means we need to dump CentOS now too.
Feels like a stab in the back.
This is exactly what it is.
You should rename project then, because It's no longer "Community" enterprise os.
You can just use "RHEL Stream". Simple, right?
Yes, good point. The name has become a misnomer.
What I find ironical is also the fact that Centos 7 now officially has a longer EOL than Centos 8!
Ironic, and it puts those of us who have recently migrated many of our development serves to CentOS8 in a really bad spot. Luckily we haven't licensed RHEL8 production servers yet -- and now that's never going to happen.
With this announcement, I would no longer put much trust in that CentOS7 EOL date either...
I Sarted some days ago a migration process from servers runing 6 and 7 to 8, and I feel like and idiot now.
I can't believe what IBM is actually doing. This is a direct move against all that open source means. They want to do exactly the same thing they're doing with awx (vs. ansible tower). You're going against everything that stands for open source. And on top of that you choose to stop offering support for Centos 8, all of a sudden! What a horrid move on your part. This only reliable choice that remains is probably going to be Debian/Ubuntu. What a waste...
What IBM fails to understand is that many of us who use CentOS for personal projects also work for corporations that spend millions of dollars annually on products from companies like IBM and have great influence over what vendors are chosen.
This is a pure betrayal of the community. Expect nothing less from IBM.
This is exactly it.
IBM is cashing in on its Red Hat acquisition by attempting to squeeze extra licenses from its customers.. while not taking into account the fact that Red Hat's strong adoption into the enterprise is a direct consequence of engineers using the nonproprietary version to develop things at home in their spare time.
Having an open source, non support contract version of your OS is exactly what drives adoption towards the supported version once the business decides to put something into production.
They are choosing to kill the golden goose in order to get the next few eggs faster. IBM doesn't care about anything but its large enterprise customers. Very stereotypically IBM.
Like some, I have already invested a bit of work getting my systems ready for CentOS8, and that work is now probably going to waste.
But even more concerning for me is that CentOS has broken trust by reverting CentOS8 EOL from 2029 to 2021. I don't think it's a coincidence that this happens mere months after IBM bought Redhat; and this clearly confirms we can now no longer trust CentOS owners Redhat and IBM to stick to their word with CentOS, RHEL, or anything else...
Not only breaking the support promise but so quickly (2021!)
Business wise, a lot of business software is providing CentOS packages and support. Like hosting panels, backup software, virtualization, Management. I mean A LOT of money worldwide is in dark waters now with this announcement. It took years for CentOS to appear in their supported and tested distros. It will disappear now much faster.
Community wise, this is plain bad news for Open Source and all Open Source communities. This is sad. I wonder, are open source developers nowadays happy to spend so many hours for something that will in the end benefit IBM "subscribers" only in the end? I don't think they are.
What a sad way to end 2020.
I don't want to give up on CentOS but this is a strong life changing decision. My background is linux engineering with over 15+ years of hardcore experience. CentOS has always been my go to when an organization didn't have the appetite for RHEL and the $75 a year license fee per instance. I fought off Ubuntu take overs at 2 of the last 3 organizations I've been with successfully. I can't, won't fight off any more and start advocating for Ubuntu or pure Debian moving forward.
Red Hat killed a great project. I wonder if Anisble will be next?
This is a nuts move! Once upon a time RedHat was a trusted name -- but after this, who could possibly recommend RHEL for a long term stable platform?
I am in the process of deploying CentOS8 development servers, and RHEL8 prduction servers for multiple customers... That is now not feasible as licensing dev servers & VM's is not going to happen in this company.
Even I could convince other folks to stick with RHEL, why would I after being backstabbed like this?
And as others have mentioned, many of use CentOS for our personal projects which gives us the experience to use RHEL in our work. Given what is happening with RedHat now, what sysadmin in their right mind would pick RHEL?
I would guess that RHEL8 won't even make it to it's so- called EOL after this.
Absolutely disgusting. You vile, terrible people.
Goodbye to you and thanks for shanking your user base.
I hate IBM more now.
We expected this and as such are ripping RHEL and CentOS out of all 15,000 boxes.
Nice way to shoot yourself in the foot.
Have you decided on what the replacement OS would be? Mind sharing what your thoughts are?
Looks like they're moderating comments out of existence, too.
What a horrid mistake. Incredibly stupid. Now I have to run that Ubuntu garbage.
Hoping that stabbing Open Source community in the back, will make it switch to commercial licenses is absolutely preposterous. This shows how disconnected they're from reality and consumed by greed and it will simply backfire on them, when we switch to Debian or any other LTS alternative. I can't think moving everything I so caressed and loved to a mess like Ubuntu.
How about trying Mageia? ^_^
IBM is destorying Red Hat. Sad!
news and event in 2020 it's more like war
Assinine. This is completely ridiculous. I have migrated several servers from CentOS 7 to 8 recently with more to go. We also have a RHEL subscription for outward facing servers, CentOS internal. This type of change should absolutely have been announced for CentOS 9. This is garbage saying 1 year from now when it was supposed to be till 2029. A complete betrayal. One year to move everything??? Stupid.
Now I'm going to be looking at a couple of other options but it won't be RHEL after this type of move. This has destroyed my trust in RHEL as I'm sure IBM pushed for this. You will be losing my RHEL money once I chose and migrate. I get companies exist to make money and that's fine. This though is purely a naked money grab that betrays an established timeline and is about to force massive work on lots of people in a tiny timeframe saying "f you customers.". You will no longer get my money for doing that to me
In hind sight it’s clear to see that the only reason RHEL took over CentOS was to kill the competition.
This is also highly frustrating as I just completed new CentOS8 and RHEL8 builds for Non-production and Production Servers and had already begun deployments. Now I’m left in situation of finding a new Linux distribution for our enterprise while I sweat out the last few years of RHEL7/CentOS7. Ubuntu is probably a no go there enterprise tooling is somewhat lacking, and I am of the opinion that they will likely be gobbled up buy Microsoft in the next few years.
Unfortunately, the short-sighted RH/IBMer that made this decision failed to realize that a lot of Admins that used Centos at home and in the enterprise also advocated and drove sales towards RedHat as well. Now with this announcement I’m afraid the damage is done and even if you were to take back your announcement, trust has been broken and the blowback will ultimately mean the death of CentOS and reduced sales of RHEL. There is however an opportunity for another Corporations such as SUSE which is own buy Microfocus to capitalize on this epic blunder simply by announcing an LTS version of OpenSues Leap. This would in turn move people/corporations to the Suse platform which in turn would drive sale for SLES.
OpenSUSE Leap releases are supported for 36 months. OpenSUSE Leap and their LTS (SLES) share the same base.
SUSE is now an independent company - Micro Focus sold it in March 2019.
In the words of Valve developers "This is utterly fucking retarded"
In hind sight it’s clear to see that the only reason RHEL took over CentOS was to kill the competition.
This is also highly frustrating as I just completed new CentOS8 and RHEL8 builds for Non-production and Production Servers and had already begun deployments. Now I’m left in situation of finding a new Linux distribution for our enterprise while I sweat out the last few years of RHEL7/CentOS7. Ubuntu is probably a no go there enterprise tooling is somewhat lacking, and I am of the opinion that they will likely be gobbled up buy Microsoft in the next few years.
Unfortunately, the short-sighted RH/IBMer that made this decision failed to realize that a lot of Admins that used Centos at home and in the enterprise also advocated and drove sales towards RedHat as well. Now with this announcement I’m afraid the damage is done and even if you were to take back your announcement, trust has been broken and the blowback will ultimately mean the death of CentOS and reduced sales of RHEL. There is however an opportunity for another Corporations such as SUSE which is own buy Microfocus to capitalize on this epic blunder simply by announcing an LTS version of OpenSues Leap. This would in turn move people/corporations to the Suse platform which in turn would drive sales for SLES.
So the inevitable has come to pass, what was once a useful Distro will disappear like others have. Centos was handy for education and training purposes and production when you couldn't afford the fees for "support", now it will just be a shadow of Fedora.
This is a bit sad. There was alway a conflict of interest associated with Redhat managing the Centos project and this is the end result of this conflict of interest. There is a genuine benefit associated with the existence of Centos for Redhat however it would appear that that benefit isn't great enough and some arse clown thought that by forcing users to migrate it will increase Redhat's revenue. The reality is that someone will repackage Redhat and make it just like Centos. The only difference is that Redhat now live in the same camp as Oracle.
What is with all the FUD? Christ sakes, don't any of you people freaking out mirror repos? This is really not a big deal. Certainly not a big enough deal to jump ship and rebuild your entire infra on debian or some such.
Well, how would you feel if you used a stable enterprise distro for your server in a production environment but was then forced to have to use non production quality software because the developers decided to change the end of life date of the stable distro from the 9 years that was set to a year without asking the community. I'd be pissed because I'd have to either deal with a shittier beta version of RHEL or I'd have to switch my servers to Debian so I'd have a stable production ready server.
Uh yeah, this is exactly that big a deal.
Exactly this, RobbyC. Mirror your repo. Test your image. Update your repo in stage. Test your image. Rinse. Repeat. It's like, one more step.
Anyway, "Switches over to Ubuntu".
I am very sorry for this betrayal.
Look, we've lost the last 10 years.
Maybe we should spend another time in the ubuntu trash from now on.
Use Oracle Linux!
Happy to report that we've invested exactly one day in CentOS 7 to CentOS 8 migration. Thanks, IBM. Now we can turn our full attention to Debian and never look back. Here's a hot tip for the IBM geniuses that came up with this. Rebrand CentOS as New Coke, and you've got yourself a real winner.
OS2 Warp R.I.P. 1996
"don't any of you people freaking out mirror repos?"
not sure exactly what you mean by this. can you explain? i mirror the centos 8 linux repos but i'm not sure what that has to do with the change to stream. yes stream has repos setup exactly like regular centos 8 linux repos (afaik) but the contents wont be the same.
as far as whether or not it's a big deal or not... that depends on the individual. for many switching from existing centos 8 to a continuous release beta version of something between fedora and rhel is a really big deal.
Wow, Cent-StreamOS, good luck finding a positive comment in here to quote to whoever the hell paid you to do this or litigated you into a corner. Day one and we are already discussing alternatives.
I haven't looked at the program because I do the same thing you do, but I believe you can develop on RHEL for free now:
Everyone predicted this when redhat bought centos. And when IBM bought RedHat it cemented everyone's notion.
I agree with you.
And on the next comment!
I had always advocate for CentOS as a stable distro....
This change, makes me change my mind about FOSS and future business stability
Thankfully we just started our migration from CentOS 7 to 8 and this surely puts a stop to that. Even if CentOS backtracks on this decision because of community backlash, the reality is the trust is lost. You've just given a huge leg for Ubuntu/Debian in the enterprise. Congratulations!
What the..? IBM should stop doing killing CentOS! That is the stupidest decision!!!
This news has upset users all over the world.
RIP CentOS :'(
Thanks for all those years.
It's a nice betrayal from IBM/RH to end this 2020.
This has nothing to do with IBM. It has everything to do with the fact that governing board of CentOS has majority of Red Hatters. Red Hat wanted this years ago WAY before IBM.
Shocked and appauled by this.
We use RHEL exclusively at our company and have a large contract with RH. We recommend our users use Centos at home for personal learning of RHEL. I use Centos for home server use. We make heavy use of third party repos which will likely suffer/disappear now. Centos also provided a fast environment to test things we planned to deploy on RHEL with extra licensing (e.g Clustering)
Additionally, we benefitted from a large Centos community providing extensive bug analysis, blogs and how-to information, applicable to RHEL.
I will be contacting my RH account manager tomorrow to register my disgust. Not that I think it will help, some MBA likely thinks this will help the bottom line, when it won't they will have already moved to the next gig.
I am senior system admin in my organization which spends millisons dollar a year on RH&IBM products. From tomorrow, I will do my best to convince management to minimize our spending on RH & IBM, and start looking for alternatives to replace existing RH & IBM products under my watch.
This is disgusting. Bah.
As a CTO I will now -today- assemble my teams and develop a plan to migrate all DataCenters back to Debian for good. I will also instantly instruct the termination of all mirroring of your software.
For the software (CentOS) I hope for a quick death that will not drag on for years.
This was ought to happen after IBM came on board
C'mon guys.. How many of you are actually using centos just to avoid paying to rhel? Just switch to Arch, Debian or whatever and get over with it.
Well one day when you become a CIO or CTO, then you'll understand what is a fuss about.. 😉
I'm not so sure if our comments are deemed as "valuable" as the headline says; in fact, especially since Red Hat was bought by IBM, I even doubt anyone who matters in this subject is even paying attention, but here goes anyway, just to relieve some of the frustration..
Yes, as mentioned multiple times: cutting down support by 8 or so years when people started using it is a betrayal of the trust that your community had in you, and so is making CentOS (which started as Community Enterprise Operating System) essentially a beta product.
If you really pull this through (which I think you will, because: why listen to us mere mortals when you think you can make more money, which remains to be seen), you don't deserve the trust we gave you anymore.
There was NO confusion whatsoever about what CentOS is/was to the people using it. Maybe there was confusion in your corporate minds; probably because it didn't make you a dime? Therefore you didn't create less confusion: what you did is create a lot of confusion where it wasn't before.
The conclusion can only be that you didn't want to clear up any confusion, you wanted to get rid of the competition and get more RHEL revenue. The people who where governing CentOS should NEVER have joined Red Hat.
I guess we all should have known something like this was coming to us. First when CentOS joined Red Hat, and later when IBM bought Red Hat. I mean: who can trust corporations and what they're telling us?
Companies need to make a profit to survive, I get that. So far I always understood that Red Hat did just fine on its own, that's what they made me believe, doing things the way they were doing them.
But alas, this is what you get when corporations step in and greedy people want more, more, more, and even more money in their pockets, even though they already have so much of it.
I now feel kind of foolish for hoping that CentOS would remain the way it has been until now; maybe with a couple of changes, but still remaining sort of the same. I must say I certainly didn't expect this dumb move amid a running version that was presented to be supported to 2029.
We are at the brink of migrating our old data-center environment to the cloud, and in the process we were going to migrate as much as we can from CentOS 7 to CentOS 8. Instead of just progressing, we now have to decide whether we'll even keep using CentOS, because it's doubtful if we can still trust you in your motives and if CentOS will be as stable as it was before.
If we stop using CentOS and standardize on something else, it also means that remains to be seen whether I should renew RHCE in the future, because: why bother?
I expect decisions like these will also have to be made by others, so, Red Hat/IBM, you might end up not only NOT have the revenue of us buying RHEL instead of using CentOS, but also LOSE certification revenue. The future will tell..
Finally, you already have Fedora and there has never been any confusion about how Fedora related to RHEL. There really is no need for another beta product.
All of that said and without any sarcasm: please, please surprise us positively and think it over - again.
My (very large, aerospace) employer is dropping RHEL for Oracle Linux, via in-place upgrade.
That seems to be more and more a good idea...
Is Scientific still out there as a RHEL clone?
Scientific Linux is only available for EL7. They decided against doing anymore builds for major releases for whatever reason. If you are interested in a potential successor, keep your eyes out RockyLinux.
CentOS is dead, long live Rocky Linux (Rock solid Alternative)
No, Scientific was discontinued in April 2019. I believe they decided to use CentOS 8.0 instead of releasing a new version of Scientific.
well I guess they will start the project again... 😉
Oh how I hope this to be true !
We have several HPC clusters and have been using SL5, SL6 and SL7 for the last 10 years and are about to complete the migration to CentOS8 in januari.
We will complete that migration , too much already invested and leaving it at SL7 is a bigger problem , but this has definately made us rethink our strategy for the next years.
I have some C8 servers to migrate to Debian now. Not a big deal in itself, I know some people have to do hundreds. Ubuntu is not usable due to their slowly enforced drift towards snap and snap store from deb packages and repositories.
The big deal is that this will basically kundermine RPM based ecosystem. Developing and testing on CentOS just became worthles, focus shifted from RPM to DEB. Developing and testing on CentOS and releasing for RHEL was convenient and didn't involve any licensing mess to slow things down.
I very much understand the need for revenue and how CentOS was seen as cutting into it, but it might actually have been a thing that fed RHEL. There is a need for quality commercial support in production, RHEL is a prime choice for that. Now they just don't have a training and proving field anymore.
Fedora just became a big question mark as well.
Great day for this announcement! We were about to install CentOS in one of our machines. Our option is now crystal clear: would you guess it? It starts with "D" and ends with "ebian".
IBM are seeing every CentOS install as a missed RHEL subscription. I see a release in-step version of CentOS as a boost for RedHat.
Why should they not be able to do this? The point of Linux is that people can do what they want? And it isn't like there aren't enough distributions? I'm not sure I see what the major problem is. I understand that it's inconvenient but you don't pay for CentOS and this is why? How can you complain about this it's free...
By the way if anyone at CentOS reads this do what you want to do ignore these other idiots.
For one it was promised this wouldn't happen when CentOS was taken over by RH and subsequently by IBM and for two, what happened to the 2029 EOL date?
I'm almost positive people would prefer still to pay a small fee to continue having stepped releases of CentOS vs streaming potential garbage at their servers, but you can't also expect people who have hundreds of servers running CentOS to up and buy RHEL licenses. I'm glad you don't see this as a problem, clearly you don't run many if any servers but to call others idiots, well that just shows what an idiot you are doesn't it?
These statements do not hold ground.
Not all people can do what they want, because not all people are able to; perhaps because of lack of knowledge (no, not everyone can learn this), time restraints or other reasons.
No, you can just switch to another distribution. Apparently you don't have to administer lots of servers and/or are not automatically deploying servers. The process of everything will change and doing so means a massive amount of work. If you would, then you'd know that it's just not viable for everyone to do this.
CentOS used to be a project separate from Red Hat, and it should have stayed that way. What people have been afraid of since 2014 just happened.
Probably should have been more clear. I work for a large company and am one of two people that support the Linux environment the vast majority of which is CentOS 7 and 8. I say this as someone who's work will be greatly effected by this. But you can't blame the CentOS project for that, at the end of the day the company choose use an OS that is free and run by a community. The CentOS guys aren't at fault.
You're correct, IBM is at fault, not CentOS.
Sorry but I have to catch my flight from RPM to DEB. See you all!
This seems such a bad move that undermines the spirit of OSS and breaches the trust of tens of thousands in the intellectual community. This will have repercussions for sure.
As a lot of us here, I've been in the CentOS / RHEL community for more than 10 years.
Reasons of that choice were stability, long term support and good hardware vendor support.
Like many others, I've built much of my skills upon this linux flavor for years, and have been implicated into the community for numerous bug reports, bug fixes, and howto writeups.
Using CentOS was the good alternative to RHEL on a lot of non critical systems, and for smaller companies like the one I work for.
The moral contract has always been a rock solid "Community Enterprise OS" in exchange of community support, bug reports & fixes, and growing interest from developers.
Redhat endorsed that moral contract when you brought official support to CentOS back in 2014.
Now that you decided to turn your back on the community, even if another RHEL fork comes out, there will be an exodus of the community.
Also, a lot of smaller developers won't support RHEL anymore because their target weren't big companies, making less and less products available without the need of self supporting RPM builds.
This will make RHEL less and less widely used by startups, enthousiasts and others.
CentOS Stream being the upstream of RHEL, I highly doubt system architects and developers are willing to be beta testers for RHEL.
Providing a free RHEL subscription for Open Source projets just sounds like your next step to keep a bit of the exodus from happening, but I'd bet that "free" subscription will get more and more restrictions later on, pushing to a full RHEL support contract.
As a lot of people here, I won't go the Oracle way, they already did a very good job destroying other company's legacy.
Gregory Kurtzer's fork will take time to grow, but in the meantime, people will need a clear vision of the future.
This means that we'll now have to turn to other linux flavors, like Debian, or OpenSUSE, of which at least some have hardware vendor support too, but with a lesser lifecycle.
I think you destroyed a large part of the RHEL / CentOS community with this move today.
Maybe you'll get more RHEL subscriptions in the next months yielding instant profits, but the long run growth is now far more uncertain.
I truly "press F to pay respect" to the former CentOS team, and would like to thank you guys for the excellent work among all those years.
I also wish you real success with the fork Gregory, but I doubt it will get the attention CentOS had in the past quickly enough to be the de-facto solution we would need.
Also, why bother supporting a company which takes efforts to kill it's community ?
As to you IBM, I give you another type of "F".
I'm usually not flaming in comments, but for you IBM, I'll happily make that exception.
"I also wish you real success with the fork Gregory, but I doubt it will get the attention CentOS had in the past quickly enough to be the de-facto solution we would need."
I'm not entirely sure about that. I remember when MariaDB forked from MySQL almost immediately when Oracle bought it. Now MariaDB is the default DB for most Linux distro's. I'm hoping that something similar happens for "Rocky Linux", as it's apparently going to be named.
Rob, I truly hope so, but in the meantime, a lot of people will leave the ship and generally speaking will get more frightened of using a community enterprise OS.
The fear may also come from IBMs future plans for RHEL.
Anyway, it can't get worse right now, and having a fork is the best that can happen now.
Hopefully the guys from SL linux will also contribute to Rocky Linux.
Your comments are spot on. ISVs often only bother with RHEL because of the large Centos base. Same with third party repos. As a desktop Linux user (commercially) this is only viable with 3rd part repo,s VLC for us is an obvious item. Will even Google bother with Chrome on RHEL without Centos? MS Teams is a corporate product we need in the current circumstances, will MS bother without Centos.
I have to admit -- I've not seen such a short-sighted, destructive decision in a long,long time. Par for 2020, for sure.
For all the disappointment, outright anger and vitriol, and stated destruction of trust that has been so thoroughly laid out in the earlier replies here, I believe that it is still possible for IBM/RedHat/CentOS to regain a large portion of what they've just lost -- If they publicly admit a WTF moment and *quickly* recant this decision.
In our university setting, we entitle whole hypervisors to enable our RHEL VMs, but we still run mostly CentOS-based systems (7/8) for two chief reasons -- 1) same production consistency between RHEL and CentOS, and 2) Don't have to waste time screwing around with subscriptions-management-overhead, etc. (CentOS x following RHEL x by a few days is not that big a deal for most, but we still pay the RedHat piper...)
Today, that blade-in-the-back stings... But there is a limited window to gain back some of that trust. If not, I don't think I can hold off the tide pushing to transition away from RHEL/CentOS after this.
FWIW, over the years we've seen several mostly-parallel efforts pack up and close down because they didn't see the longer term value in competing with CentOS. (White Box Linux, Scientific Linux, etc.) If no recant is forthcoming on this decision, I have to wonder if there is enough groundswell to restart one or more of these other efforts.
(Hmmmn... I'm coming up on an early retirement in a couple years... perhaps that could be a great way to keep me occupied...)
We will migrate our 600+ CentOS servers to Ubuntu LTS
good bye Centos
Too bad, centos is not stable anymore
🙁 I just started to migrate to CentOS8 from CentOS7, now I have to think what to do next.
Don't like this idea.
Thank god I did not start to migrate my sites to Centos 8, will stick to Centos 7 and will use Debian on future servers.
Not a big deal to be honest. I love CentOS but we all can live without it. Sad, but well..
jejeje... at the same situation...
Great, you have just done a complete mess of what once was a perfect test&prod environment. Debian now is not an option, but one of the requirements.
Exactly my thoughts. And I don't think there are any other Distros which offer a longer EOL, if I'm not mistaken. (rolling updates obviously don't count)
Some years ago IBM bought Informix. We switched to PostgreSQL, when Informix was IBMized. One year ago IBM bought Red Hat and CentOS. CentOS is now IBMized. Guess what will happen with our CentOS installations. What's wrong with IBM?
"Community-driven", ok... (sarcastic laughter)
It sounds like you're suggesting that the community wasn't actually begging to saw off the branch it was sitting on!
Our annual subscription bill for using well over a thousand RHEL servers is eye watering. Our distributed and diverse development base makes use of CentOS servers to both keep costs down and ensure no license breaches occur - they are free to develop new ideas without constraint. Should I now be thinking of moving to another platform to continue doing what already works very well for us?
I am so glad that I left my 40 servers on C7 so I do not need to worry till y.2024, only 2 of them are on C8 which I will rectify beginning of the next year... But after that, I'm moving to Ubuntu for at least 5 years of support. I have no man power to simply accept IBM's "let's fool around" decisions.
what if the similar decision is made for C7?
Maybe "openEuler LTS" is going to be our new CentOS
This remembers me what happened when Oracle bought Openoffice.org
So they forced people to create Libreoffice which nowadays dominate the market.
And what today about a "LibreCentos" ?
Branch IT and start again!!!
Fuck!! the same happen with Zentyal what is happening to You? CentOS community base will have to branch the project and start again. Or rebase to debian.
Well that's annoying.
I work for a 25K+ organisation that has been a Red Hat customer from the start.
Those of us that *develop* for the RHEL installations use CentOS because we can easily spin up a new dev or test VM when we need to without having to make a business case for it. If something even looks like it costs money then there will be hoops to jump through which costs the most precious resource: time.
I guess IBM/RH are too focused on those few customers that would otherwise pay but this will place roadblocks in front of developers for paying customers too.
And for cutting out eight years from a published EOL date? That will burn a lot of goodwill.
The logical next step is to cut off CentOS 7 support from 2021, "just to focus on current stream version".
Things could not get worse, it's 2020, right?
Now please remember CentOS was promised to be supported till May, 2029.
All these open-source ventures are based on trust. Now I wonder how would RH/IBM/whoever try to regain the trust, after suddenly assigning CentOS 8 extinction event to 2021.
A very poor decision, but I've been unhappy with the direction CentOS has been heading in lately anyway.
I've used CentOS for 10 years, but it's time to bid it farewell. This is the kick up the ass I needed to start migrating all my servers to Ubuntu server.
What's frustrating is these decisions seem to be made arbitrarily, and without any consultation.
As far back as I can recall, this is a completely unprecedented betrayal of the release support roadmap for a major Linux distribution. Stunning.
I am a system administrator for a large ISP/MSP and we will be ditching CentOS, scaling down RHEL and moving to Ubuntu and others.
The whole premise is utterly ridiculous, but completely killing 8.x support (which was supposed to last until what, 2029?) at the end of 2021 just as we've finished migrating a lot of 6/7 to 8 completely seals the deal.
All of us complaining to each other and making suggestions to a blog is unlikely to accomplish our goals. I suggest we all combine our efforts and let @ArvindKrishna the CEO of IBM who claims to be an "avid learner and listener" and CEO of @RedHat @PaulJCormier know how we feel. We should all start mentioning them on Twitter and get some shareholder visibility on how this is a lose-lose for both the shareholder and the CentOS user. They could:
Put the original 2029 EOL in place for CentOS 8.
Or at least offer some kind of route for us to easily transition to RHEL at low cost, or purchase security patches on a subscription basis.
Instead they are going to force most of us to abandon CentOS and Red Hat entirely. We don't want to have to rebuild our entire infrastructure simply to go from CentOS to RedHat.
Start the social media campaign immediately, or we are destined to lose.
You've broken trust in published EOL dates not just for CentOS but for RHEL by extension. If you are willing to alter published EOLs for one product, why should anyone trust that the fallout from changing it on RHEL in the future wouldn't be considered just a cost of business if someone thought it would make more money? Published EOLs from you can no longer be relied on, and there are other distributions that haven't made such a colossal error in breaking the trust of their users.
Thank you for making my December more complicated as we are now pivoting to redesign projects which were already preparing for deployment away from using either CentOS or RHEL, really what I wanted to cap off 2020 /s.
You people at IBM/RedHat are betraying and crucifying your own users and community, right after we have started numerous CentOS 8 systems in production, after months or even years of planning, investing in know-how and testing.
This is really irritating.
If you do not recall this policy, not only you will cause huge problems to thousands of administrators out there; you will suffer their wrath.
Unless you recall this policy and provide CentOS 8 as promised until the end of its life-cycle, you are proving imposters by luring thousands to using a product which you planned to effectively abolish.
I still hope that you will not disappoint CentOS admins and users so badly and that you will continue to support CentOS 8 (and CentOS 7) in its current/expected form.
This move deserves a rebranding!
How about swapping the 'e' for a 'u' ?
Community UNTrustworthy OS
Funny timing on this, just switched over 50 severs from Centos to Oracle. We also switch a dozen RHEL servers over the same night and cut our support fees in half! Oracle is free, including the patching and from what we tested, a better solution for us. Now Centos is dying with IBM Red Hat guys screaming how good this is for them.
Some time ago I switched all the kit I'm responsible for over to ALT Linux and it's been a wise and stable move ever since.
It's russian crap.
This has been clearly a way to force all the Centos Enterprises users to pay for licenses. It is clear that RedHat was losing a lot of money due to the competition with Centos and they have solved it hard. They also know very well how difficult it is for a company to migrate all its servers to another distribution. A game of the dirtiest.
Thanks for the news I ALMOST began updating to Centos 8 and you clearly saved me the time. From this news I am testing a Debian 10 install since it appears Centos is no longer a viable option. It's been a fun several years but now that a big old corporation has hold of both RedHat and Centos neither remain a linux option for me. Good luck but I think you just put a nail in your own coffin.
IBM is about a useless company. Thanks for destroying perfectly good software.
Apparently that answer is yes. I knew IBM grabbing up red hat was gonna be a bad thing.. didn't expect it to be this bad. I haven't played with Debian in years but it's time to get back into the learning curve. At least my Centos 7 boxes will be fine till 2024.
That has became uncertain now. With this move, one can't trust CentOS 7 support to remain the same. They very well may terminate it along with 8.
Can we know if this is an unanimous decision from the Governance board ?
Forget CentOS. Let's do Rocky Linux like RHEL great again. Story like MySQL to MariaDB. OpenOffice to LibreOffice. Only community transaction from one project to better project.
This is a sad day. I've used CentOS for years for many different projects and it's always been rock-solid stable.
I guess I'll need to move to an alternative to get this reliability from now on.
Congratulations on your own coffin nailed. I am using CentOS7 and I think the time has come to switch my system over to Debian or Ubuntu with your decision. At least I'm fine until 2024
For now. They might change their mind again. Who knows when commitments mean nothing.
Step 1: Get lots of people to to use product binary compatible with RHEL
Step 2: Remove that product, force them to pay you money
The fact that we were promised EOL of 2029 and got backstabbed to an EOL of not even end of CentOS 7 but before CentOS 7 is a slap in the face!
We just spent a ton of resource jumping from 6 to 8 thinking it we would be fine until 2029, well so much for that...
I guess I have no choice but to move away from CentOS and RHEL because we simply can't trust the LTS commitments.
Who wants to do business with a company who you can't trust?
It does not sound like you were doing business with Red Hat if you were just using Centos. Or are you saying you were using some Centos and some RHEL? It is not clear from your post.
Time to leave centOS/redhat.. Goodbye..welcome to Ubuntu..
Looks like openSUSE will get lots of new subscribers now: While Leap gets closer to SLE - currently they use the same sources for the base packages, in the future the same binaries, Tumbleweed as the tested rolling release is for everyone who cant wait to get the latest software.
Remember, openSUSE is an independent community that builds great tools beside the Distributions....
Remember when RedHat, around RH-7.x, wanted to charge for the distro, the community revolted so much that RedHat saw their mistake and released Fedora.
You can fool all the people some of the time, and some of the people all the time, but you cannot fool all the people all the time.
Even though RedHat/CentOS has a very large share of the Linux server market, it will suffer the same fate as Novell (had 85% of the matket), disappearing into darkness !
As I predicted, RHEL is destroying CentOS, and IBM is running Red Hat into the ground in the name of profit$. Why is anyone surprised? I give Red Hat 12-18 months of life, before they become another ordinary dept of IBM, producing IBM Linux. CentOS is dead. Time to either go back to Debian and its derivatives, or just pay for RHEL, or IBMEL, and suck it up.
> Given this, we’ve informed the CentOS Project Governing Board that we are shifting our investment fully from CentOS Linux to CentOS Stream.
So basically CentOS 8 has been defunded. I doubt the governing board is happy either. It'll be interesting to see what Karanbir Singh does next.
Looking at all these comment (and I agree wholeheartedly), a very fitting scene from "Good Morning Vietnam" springs to mind:
1 guy called and said it was visionary, the other 1100 said it was not. - "Hey IBM, you svck. Eat a bag of sh^t. That's pretty much to the point."
Ironic how a comedy is the most fitting one...
Welllllll then. 317 servers to move to Ubuntu LTS and I WON'T bother moving any back if you reverse the decision.
What a surprise at the end of 2020, hope there is a win-win solution.
Finally, centos is rolling release. Every year I ask why oh why couldn't this distro be rolling release?? Well, ask and you shall receive! This distro clearly has its finger on the pulse of those who use it!
It isn't going rolling release.
You still have to upgrade from CentOS stream 8 to CentOS stream 9.
The only thing that is changing is you are getting less stable release.
I genuinely cannot tell if you are sarcastic or stupid.
is it good? I wanted to switch to Centos in a short time.
CentOS as it existed is now dead. CentOS Stream is a "development platform".
I am mid-migration from Rhel/Cent6 to 8. I now have to stop a major project for several hundred systems. My group will have to go back to rebuild every CentOS 8 system we've spent the last 6 months deploying.
Congrats fellas, you did it. You perfected the transition to Debian from CentOS.
Well “Those who do; decide” and that is going to happen.
- RedHat stops CentOS stable releases
- Somebody creates a fork and fills the gap.
Some guys thought they were better than others and they were capable of handling any consequences followed by their foolish decisions. Yeah, before everyone left them.
I find it kind of funny, I find it kind of sad.
The dreams in which I moving 1.5K+ machines to whatever distro I yet have to find fitting for replacement to are the.. Wait. How could one with all the seriousness consider cutting down already published EOL a good idea? I literally had to convince people to move from Ubuntu and Debian installations to CentOS for sake of stability and longer support, just for become looking like a clown now, because with single move distro deprived from both of this.
This is a monumental shoot in the foot.
For years i have used CentOS for home projects and enterprise projects. Those enterprise projects always turned into RHEL projects because of CentOS.. This is really a bad move for IBM.. Well, i knew something bad would happend when IBM purchased RedHat...
"Shift of project focus" == Dropping a product we promised support for through 2029.
Just a shame, waste of well provided good software
Literally in the process of migrating from Debian to CentOS for production servers as this got published. Stopped, went back and redeployed with Ubuntu LTS. Thanks a lot guys
Happy to donate and be part of the revolution away the Corporate vampire Squid that is IBM
This is terrible.” Let’s take the reason everyone uses our operating system and shift our focus to the opposite”. Brilliant, love you IBM. I hope a project steps up to replace the “10 year stable release enterprise OS” need. Please let me know if anyone has an alternative other than Ubuntu’s (5 year support).
Ok thnx bye!
Red Hat's word now means nothing to me. Disagreements over future plans and technical direction are one thing, but you *lied* to us about CentOS 8's support cycle, to the detriment of *everybody*. You cost us real money relying on a promise you made, we thought, in good faith. It is now clear Red Hat no longer knows what "good faith" means, and acts only as a Trumpian vacuum of wealth.
First I want to say I have been using RedHat/Centos for 23 years, probably still have some of those old CD's somewhere. RedHat/Centos in my opinion the standard, and has been a great solid platform.
I would ask first, how bad is this? Lets not get carried away, lets give some time to the "Centos Stream" team to provide more data, and philosophies!
It's very possible they will maintain the same solid, reliable and predictable OS as they have done in the past.
2nd Jumping ship is insane at least at the moment talking about all these other flavors of linux, and converting everything tomorrow... slow your roll.
There are excellent solutions out there, yes Ubuntu, Debian, FreeBSD, Gentoo ( I have a friend that swears by it), Arch and others which have their own direction to pick since they are also based on RHE.
I have a server or 2 running Debian and Ubuntu but everything else is Centos, my brain speaks Centos, these others have subtle learning curves.
My main point is, we need to see now if the direction by RH has been written in stone, and what those other distros plans are before jumping in direction.
I'm hoping Centos Stream can be the viable alternative.
Other things to consider, core internet servers like cPanel/WHM and FreePBX run very well on Centos, what direction will they take, will their research help you forge what may be a better decision then an article we read last week?
Either way, we have a window of time and need to not overreact on our frustration of what on the surface seems will be a break from RHE...
Trust me, Centos is hands down my favorite.
1 year is hardly enough time. Just planning alone can take a year. And good luck explaining to management why all your plans fell through.
At the very least they could have supported CentOS 8 until 2024 like CentOs 7.
Sorry, buts it's hard to believe that Centos Stream would be a stable solution just based in their testing abilities.
I good example I like to give is about the maintanance many EPEL packages that use rolling lauch like EXIM. From times to times something broken there, is not about becouse the maintainers are incompetent, but becouse the individual and irregular lauches are too hard to keep, specialy when we have diferents systems integrated. So to solve this I switched to Postfix that was maintained by core system. With with rolling lauches all the work I only need to do once per year can happen any time in suprise.
Yes, there are subtle learning curves in distribution changes (except for FreeBSD where there are many more things different, it's a different S.O) but is no more harder than learning curve os learning a new CentOS changes (ex.: I wasted a good time learning systemd when I changed from CentOS 6 to 7).
The cPanel/WHM already jump the boat!
90% of CentOS systems out there are probably production systems. Why ON EARTH would anybody be pleased to hear that CentOS is going to be the upstream of RHEL?
Big companies being big companies I understand. I am a capitalist at heart too, but please stop insulting the intelligence of your userbase (which is primarily engineers, FFS) by feeding us lines like "And it removes confusion around what 'CentOS' means in the Linux distribution ecosystem."
There was no confusion. We wanted to receive maximum value/benefit from the 99% FOSS components of RHEL by supporting our own systems instead of paying annual licenses on all our CPU cores. Don't talk down to us.
Just admit that you wanted to close a "loophole" that you shortsightedly viewed as giving away an enterprise-grade product for free. This will not cause a mass migration to RHEL, though, so the joke's on you guys.
I fully agree with the general feeling expressed by the preceding posts. It seems that IBM has torpedoed CentOS and this project is now sinking. It is sad. The future CentOS Stream feels like beta release of RHEL. Definitely not for production and stable environments. We are looking now to switch to Ubuntu LTS or Debian. Good bye CentOS.
I've been a Linux sysadmin for 23 years or so now. Debian at home, Red Hat / CentOS at work. Red Hat Developer subscription for me at home - self support. Debug RHEL 7 and 8 at the same time - no, only one subscription. There's a CentOS mirror and an EPEL mirror on my desk next door - and Debian and Ubuntu. I've told my colleagues to look carefully as to how we support systems. CentOS Streams - a rolling beta to support the next RHEL point release - no security support, changing regularly. NO. I can't say I blame anybody - but good luck with whoever's keeping CentOS Streams going - you couldn't even build a stable CentOS 8.2 for a while because you hadn't sorted out dependencies and were dependent on missing packages
Trust is built up over months and years, and can be shattered by one wrong move.
Can we trust RH to support C7 until 2024, or will they e.g. ditch it in year's time?
You've just lost my six figure RHEL production licenses
It seems that the fork is on. https://github.com/rocky-linux/organization/wiki/Contributing
Terrible decision to kill most popular distro for political reason, why not use Fedora for upstream and leave CentOS alone.
RedHat used to be my favorite type of Linux distribution and I was using it extensively since mid 90's but unlike other people here I am not mad at RedHat.
Might be that I used up all my hate when they killed RedHat Linux distribution in favor of RHEL.
For a short time I considered to join mr. Gregory Kurtzer in his initiative to create a Centos fork from ground up but I simply don't see the point in creating another clone of RHEL as RedHat can seriously undermine the effort simply by not releasing src.rpm files in the future.
Remember that not all licenses require source to be provided to a customer and they certainly don't define in great detail the form in which source code should be made available.
In my opinion we should just let it die and focus our efforts on contributions to other functional Linux distributions already available.
On the long run it would greatly benefit us all as well as those Linux distributions and their userbase.
Currently, there are only three options one should consider:
a) convert to RHEL and purchase the required licenses
b) migrate to an alternative stable Linux distribution
c) delay the decision by moving to Centos 7 for the time being
Unfortunately, there are no other independent RHEL-like Linux distributions that are proven to be stable and up2date but there are other types of Linux distributions that were around almost as long as RedHat and one shouldn't be afraid to consider them.
As for the Ubuntu, I am unable to dismiss the RedHat->RHEL and Canonical->Ubuntu analogy. What makes you guys think that the same thing couldn't happen to Ubuntu?
Some time ago I read someone defending the Debian project agains Ubuntu arguing exactly this. That at same time that big linux corporations can levarage a linux distribution, they can also can destroy every thing with bad decisions.
That time I was thinking thats will never gona happen with a big and at that time innovative Canonical and was much more unimaginable something like this with an old and famous Redhat. But now everything changes.
Probably I will start to defend full contributed like opensource maintained projects like Debian instead of a bussiness corporated like Redhat and Canonical, becouse at any time, without caring with anyone they can change everything spreading demage in all directions.
CentOS has already lost tremendous market share to Ubuntu. This seems likely to be the final nail in the coffin. Really unfortunate, as I'd rather not use/recommend Ubuntu, but the primary reason I can convince people to use CentOS on servers, even people that use Ubuntu on their development laptop/desktop, is that it is more stable and predictable for a long period of time.
How are you going to take an OS that should have 10 years of support and, without warning,tell people sorry it's ending next year? Why would people trust you after pulling something like this? The great 2020 dumpster fire continues...
TL;DR: "After 16 years, CentOS (Community 'Enterprise' Operating System) abruptly abandons the its role and primary reason for existing. Decides to become a beta release for RHEL instead."
To the people comparing CentOS to Ubuntu. Sadly there is no comparison. Debian is great but still not a CentOS to most enterprise environments. The only real option at the moment as much as i hate what they stand for is the http://yum.oracle.com/oracle-linux-isos.html, The plus side is they release ISO way faster as well.
Those taking about Re-basing to Debian clearly do not understand the whole benefit of rebuilding Redhat packages and all the development / back-porting having already taken place, we also got RHEL certified server hardware to work without playing games in CentOS.
So, to summarize history here:
1. IBM and Oracle start using an open source OS that's superior to their proprietary OSes because it's popular and stable. They *only* do this because that OS is open source and they can compile it and provide their own support for it; safe in the knowledge they won't have the rug pulled out from under them.
2. Company makes enough profits from enterprise licenses of the open source OS that they can stop providing a free build, and they close the free version and release "Fedora".
3. Another organization takes the time to release the core OS as free again. Over 10 years of this happening it results in the Company growing and posting a 2019 profit of $300 million on revenues of $900 million.
4. Company gets purchased by IBM as a strategic competitive advantage of having a stable OS.
5. Company again decides to close the open source build, and slashes the existing End Of Life date by 90%, causing all users to abandon ship as this breach of existing End of Life guarantees means they can no longer trust the company - the entire reason IBM and Oracle started using the OS in the first place.
6. IBM and Oracle eventually start looking for another OS to use...
Addendum: Users of IBM's other open-source tools are also frightened away from using their other cloud services such as API Connect/Loopback, as the trust in *already released software* support timelines is now broken.
Not sure if this was on Red Hat cards the day they brought CentOS to their ecosystem, or is it the usual IBM short-sightedness to give more importance to short-term gains vs long-term. In any case, a trust built over years and lost in a matter of months, is going to be impossible to restore - it simply cannot be trusted anymore even if this was rolled back.
In future, I am going to think 5 times before I choose an OSS backed by a for-profit org.
Thankfully, I just recently started on Linux only at a personal level, and though I had bet heavily on CentOS after having considered all my options in detail, its Good-Bye CentOS, and Hello Debian...
Devuan is pretty good also.
I'll just add another carrot to this angry stew. It's my regret for upgrading my CentOS 6 prod servers to 8 instead of 7 with assurances that we'd be cooking until 2029. It's time to find another chef. 🙁
The option we have is to use Oracle Linux 8 which is based on Red Hat
The future of CentOS Linux is CentOS Stream. Awesome News.
Yes awesome news. For their competitors.
Like many others I have upgraded multiple home servers from 7 to 8 specifically for long term support dates, plus a couple upgraded from 7 to 8 as rdoproject.org has packaged openstack ussuri for 8 (7 stopped at the train release).
I will not be downgrading them to 7 for a longer support lifetime, like others I will look at alternatives, anything network/tftp-boot installable to make life easy but not ubuntu as its slow migration to putting everything into snap packages creates the most insecure environment I have ever seen..
This dumping of C8 will have a flow on effect to not just small developers but large groups like rdo package for fedora/centos/rhel and may just drop centos as fedora is obviously the test bed for rhel changes.
Personally I switched to centos for some servers due to constant changes in fedora breaking things, as the same thing is now intended for C8 centos also is no longer in my future.
And then there are desktops. I use fedora desktops to develop and package (rpm) applications to test on centos before releaseing. I manage all servers and desktops with puppet. Do I want to have to have to create puppet configs for multiple OS's ?, no, the desktops will also be migrated to whichever OS I choose which does limit options. It does mean I won't be producing rpms anymore; but then most users are happy to compile from source these days I just won't test or fix issues on rhel based systems.
For large enterpises managing a mix of centos/rhel with tools like puppet or chef probably will also not want to start creating multiple OS configs and select a 'standard' system for the enterprise which if they are not willing to convert C8 systems to paid for rhel systems will not be a redhat enterprise any more.
(grin) However this may just be a strategic move by ibm to migrate all users off rhel alltogether in the hope users who suddenly have a desperate need for a new OS start purchasing the thousands of os2/warp install dvds that have been sitting in warehouses for decades.
This is a centos blog, so this is a good time to say thanks to the centos team for all your years of hard work until now. While sometimes it may seem like it is not appreciated judging from the outrage in comments on this post your work has indeed been appreciated and used by both individuals and organisations, lots of them. Its place in the ecosystem will be missed.
It seems cloudlinux has also announced they will release a free open source version of RHEL/CentOS 8 that is binary compatible with all tools and support till 2029. Released in 2021 Q1
So now our options are Oracle Linux, Rocky Linux and CloudLinux version. At least some good news I guess that others won't leave us hanging...
Yes, that appears to be so.
Their summary of the blog post above is a lot more succinct, because it avoids all the circumlocutions that the author deployed: "Red Hat just killed CentOS as we know it". That's a lot more to-the-point and relevant to CentOS users than talk about "shifting focus", removing confusion, etc.
Is deciding not to build point releases a CentOS decision? I ask since the source rpms will be released either way unless that changes. How is it Oracle and company will continue to build the point releases and not CentOS or am i missing something here? Sadly i think this will have some major impact on RedHat and less for the hundreds of appliances vendors using a redhat rebuild of sorts.
Is there anything IBM doesnt get their hands on and screw it up?
I have an honest question to those who might have used both CentOS Stream *and* Ubuntu:
A concrete reason for me to be using Ubuntu over RHEL or CentOS in the datacenter has been the widespread community support, the vast amount of packages and comparably fast adoption of newer features, in combination with long term support and actual working upgrade paths between major versions in most cases. Granted, I only run Linux servers on top of VMware hypervisors rather than on bare metal, but I've never really had any nasty surprises in production environments over the years, so obviously Ubuntus LTS server versions aren't *that* unstable.
Is the goal of CentOS Stream to compete directly with Debian Testing/Ubuntu? Then I'd say it's probably stable enough for many actual production use cases.
I'm very sorry to see the original intention of the CentOS distribution get shredded in this way, though, and the lack of honor and loyalty to the community that gave the distribution its name is appalling to say the least.
Keep your promise and keep CentOS stable. The stream will not fullfill production needs.
A better approach would be to have a stram release as extra build.
So devs can choose between Centos and the stream version.
In this way Stream can prove to be stable enough for production servers.
Please post an survey about this change so user and devs who use CentOS can contribute to the decicion making. Since this is a key reason for stream. So let the community decide.
Basically, if you want a stable version, pay for RHEL. Sad!
This is dumb as it gets. Well, back to Debian then.
A message to IBM "I must break you..."
Clearly somebody with "shallo thinking" at IBM/RHEL thinks that CentOS users are going to suddenly subscribe to RHEL.
This will not be the case. This will leave a bad taste in the mouth.
CentOS users will instead go elsewhere after feeling that CentOS/RHEL/IBM cannot be trusted.
This is a very shortsighted and naïve move.
Ubuntu 2020.04 is good enough to switch
How can we be sure that Ubuntu sponsor Canonical will not get acquired and change the rules of how freely Ubuntu stable versions are available. This whole thing is making me rethink if it makes sense to rely on free unsupported software at the operating system level. I don't trust oracle and cloud linux also seems like a commercial company wanting to profit from Linux like Red Hat. Hmmm...
CloudLinux really seems to "get" it. Well done with this confidence-inspiring post:
> Why We Are Doing It
> 1. We have all the infrastructure, software and experience to do that already. We have a large staff of developers and maintainers that have a decade of experience in building an RHEL fork, starting from RHEL5 to RHEL8.
> 2. We expect that this project will put us on the map, and allow people to discover our rebootless update software and Extended Lifecycle Support offering.
So they candidly admit they won't have to allocate lots of resources to this new gig, because they've already basically been doing the work behind the scenes. Seems like evidence of a smart and sustainable business decision to me. This absolutely will put them "on the map" with this opportune announcement. And then they also openly recognize that they're not being 100% altruistic, because users of their community distro might sign up for their proprietary offerings. Awesome, that's how to make money in the open source ecosystem, and it doesn't insult their users' intelligence with a bunch of corporate whitewashing. Then the best part is here:
> What Will We Do To Make Sure That It Doesn't Go Wrong
> We plan to make all the build and test software free, open-sourced, easy to set up, so if we ever go in the wrong direction - the community can just pick up where we left off.
Wow. CloudLinux does indeed "get" it. Keep your eyes on this one folks. Between CloudLinux's new offering and Rocky Linux I'm sure the open source community will pull through. Probably even more cool forks coming down the pipeline. No need to jump ship yet.
Isn't this quote "We plan to make all the build and test software free, open-sourced, easy to set up, so if we ever go in the wrong direction - the community can just pick up where we left off." exactly what Red Hat did here.
Keeping everything open source still so that others can still create variants. And it sounds like this is a for profit business and plans to charge money for support and maybe extra capabilities. Why does this sound any better than Red Hat is my ?
Looking at the link below the packages in RHEL
use many different licenses. Among them BSD
which doesn't require source code to be published.
If RH decides to do that, then it will be
impossible to create fully compatible rebuilds.
Another reason to look for alternatives?
Uhm, those licenses are what the original producers/maintainers of those packages are using. They would be the ones pulling the plug on distributing source, at which point eve RH would probably stop including them in the distro.
If Red Hat stopped publishing source for *all* packages, then that'd be a decision that RHEL is no longer an open source distribution. They'd be betting that people don't care about being able to get the source any more. That'd be a huge step to make and they'd have to be betting that they'd suck more ex-CentOS users in in the short-term and keep them, than they'd lose in the long term from others who wanted the source.
With Rocky Linux and the CloudLinux fork, it appears that the short-term result of Red Hat deciding to abolish CentOS (which is what it is, the fact that they're using the same moniker for a RHEL upstream in future isn't relevant) is that they've immediately created two new RHEL competitors, over which they will have less control than they did over CentOS. Was that clever?
So if we upgraded our systems to CentOS 8 we are basically getting updates and maintenance cut off from the date they promised and leaving us in the dust with no stability? That's really underhanded and puts many organizations in quite a predicament. Looks like we will need to find options other than RHEL or CentOS after they went back on their word.
Not good for cloud, but I have a slackware server. That thing is gonna be around longer than the dinosaurs!
We're not interested in accepting the risk that this decision created.
A lot of companies end up using RHEL because third party tools often require that you use RHEL. So it's only natural that a person would use CentOS for other environments to bring complexity and cost down.
We started a DAR today to look at other server grade Linux OSes and the removal of CentOS 7 by June 2024.
It's a very bad decision by RHEL, I truly loved CentOS due to its stability.
We had plans to migrate 400+ production RHEL servers to Centos 8. I guess we have to find an alternative solution. This is very very bad move by Red Hat/Centos!
We can go back and forth Ad Infinium on commercial interest versus duty to support the user base of the community but that misses what are most likely the actual point(s). This move is both a strategic and architectural one at its core if you step back and look at it. When you consider the entire landscape of Red Hat's product portfolio this makes a lot of sense. One of the latest trends in software is the immutable operating system which is great in many ways but rather different from how the RHEL distro (or other RPM based distro) currently works. So if we consider the fact we have to get from where RHEL is currently to where Fedora CoreOS is eventually there will need to be architectural midpoints. This has been occurring in CentOS for quite some time if you take a look at it. AppStream (which is present even the current point version) of CentOS bears a lot of resemblance to a midpoint between a traditional RPM repository and rpm-ostree. Now, aside from expense and being denied freedom to modify the software, what we are really concerned with are breakages in referential integrity from a code perspective. Will I be calling functions/methods or interfaces whose declarations have changed in type, have they been deprecated altogether, etc. Obviously that is the impetus of having a point release we can target in the first place but it's far from the only solution. Even then that's only relevant if you rely on system libraries. Lots of implementations are more concerned with the version and changes in the Java runtime than anything.
If Red Hat implements something to the effect of certification channels (streams) in their repositories that you could lock to that would ensure that the versions of system packages aren't in excess of what would be present in a given point release then the rolling aspect would be moot. From a repository standpoint it would deduplicate a lot since the streams would be nothing more than graphs that organized the packages present and imposed limitations based on constraints you specify.
They could even go further and create or modify a build system that flags changes in system libraries that have a high likelihood of breaking referential integrity with existing code bases on a given point release. Pair that with an implementation of something to the effect of a scanner that heuristically identifies those things in a given code repository while potentially offering suggestions to fix the findings (much in the same way current IDE recommender systems do now) then you've further mitigated the headaches associated with a change like this.
Lastly, if you look at the fact they've just implemented Btrfs in Fedora, ultimately you could end up with build systems that could boot strap multiple concurrent versions of certified streams in read only images on a single machine and bi-directionally test point release equivalents without having to spin up multiple VMs. Personally I like the idea of having that sort of capability but it takes change to get there and historically that isn't received well when it breaches a certain threshold of progressiveness.
Whether or not they choose to implement something like what I've mentioned in the repositories remains to be seen. Objectively though, it would probably be wise for them to continue to support the community so as not to alienate potential customers and stifle innovation that may come from sources they can't hire. Just my thoughts on the matter though...
What you say Brian Terry is at the heart of the foreboding on stability. And it's a universal problem with implicit solutions in the way you have stated it. I hope all those with an interest in avoiding the problem take a note in what you've described. PS, what you describe for system libs applies equally well for VM's (including JVM). And that means upstream compliance first.
It is a very sad developement, especialy for us working in universities and using centos as a rock solid solution for our infrastructure. Other than that I really think that this is an unfortunate decision for redhat too. What this does is that it shrinks the RHEL userbase in the long term. I do not know if you understand this but at some point even capable linux administrators seek for commercial support on critical infrastructure. Redhat with centos had a big pool of potential customers to fish on.
And that did work. Redhat as a company did not fail, and I do not think that economics where going south before IBM took it over. I hear many people talk about openSuse, or SLES, well sorry but at least here in Greece Suse is non existent. If you take a look for what distros 3rd party packages (free or otherwise) are being maintained SuSe is usualy not among those distros. I mean that this will not go the way redhat imagines unfortunately. It will shrink their userbase in the long run. and frankly I do not think much people will use stream except from situations where there cannot be an alternative. (ovirt hypervisor maybe). I mean we where only stuck on v3 kernel throughout centos7 because we trusted it was stable. CentOS users are the least possible candidates for beta testing. Call it by it's name allready: we killed centos
They burried the lead...
"If you’re using CentOS Linux in a commercial deployment, we suggest you look at moving to RHEL for the added management technologies, security, and support that are an integral part of the RHEL subscription. Our sales teams can help you identify the appropriate offerings that match your use case."
Just move to the best Linux distribution on the planet: Debian.
Three years EOL only, unfortunately. In this light Ubuntu with 5 years looks better. Having reinstalled hundreds servers every 3 years might be a nightmare.
Jessie will have been supported 7 years when it goes out of ELTS in 2022.
This was ought to happen after IBM came on board
IBM has now officially taken over the mantle as the company where all good software goes to die.
Sounds like Oracle Linux is about to get a heck of a lot more popular. Redhat, you are shooting yourself in the foot.
We don't WANT CentOS to be upstream of RHEL. We want 100% compatibility.
Here we come Amazon Linux2, our security department was already strongly pushing us away from Ubuntu.
I'd trust AWS a lot more than Redhat to keep supporting a product
IBM - the new Oracle.
Good job guys, RedHat - that was a backwards acquisition, and this is a backwards choice.
"The future of the CentOS Project is CentOS Stream" - there is no future for CentOS. CentOS means Community enterprise OS, not RHEL preview/beta OS.
I feel sorry for everyone in the CentOS team. I can't imagine that they like the new direction of the project.
I try to get more understanding of what really is the impact? and is CentOS Stream is as bad as being portrayed here.
About alternates, I would rather slash my wrists than using *buntu, *suse, or oracle.
Not interested in Amazon Linux either. I can survive with debian though.
I am now mainly concerned about the future of Fedora project.
To make it simple.
Fedora -> CentOS -> RedHat
So before a point release update goes to RedHat, CentOS Stream is the beta test platform for that point release.
Currently, it should have no impact on Fedora. But who knows what the future will bring with IBM on board.
Your options as far as forks of RHEL 8 go are:
Current = Oracle Linux, Springdale Linux
Upcomming = CloudLinux Project(2021 Q1), Rocky Linux (The current timetable is 2 months for test builds that will leverage on CentOS and ready for deployment in 6 months)
Good thing it's early in my company's lifecycle and we only have a dozen or so VMs and our product is not so committed to the RHEL family that we can't change.
* No RHEL, either in production or lab. They can no longer be trusted to abide by contracts or other promises any longer. So much for their public statements about 2029/CentOS 8.
* No CentOS, either in production or labs. Red Hat is going to mess with it to try to force conversion and squeeze free QA out of people. If we find a bug we no longer will have certainty whether it's us or CentOS breaking a package. We don't have that kind of time to waste.
* No Fedora anywhere (it was never an option, and CentOS becoming Fedora in all but name #imports that). Production and the DevOps/build/test environments are no place to roll dice or be someone else's QA department or vulnerability/exploit lab.
All of our developers are upset about this too, and there is a LOT of talk about shifting our product off of the RHEL family onto Ubuntu; if we did so, there would be no resistance internally. Even our CEO is asking if Red Hat can be trusted any more. My answer will be "of course not".
* sigh * I guess it's Ubuntu. Canonical doesn't grub for money quite like IBM does. They haven't pulled the Vader “I have altered the deal. Pray I don't alter it any further” move yet.
Goodbye CentOS. Hello rockylinux.org
Smelling an state pressure on tech companies to drop centos/Linux as other states are curious about GNU/Free OS!! I ain't spy it 🙂
All tech companies have started praising RedHat/IBM ... !!
Hail Rocky Linux... Tech mass protest...
Why is it that i keep readin RedHat cannot be trusted? As far as i Know RHEL8 will be supported till 2029. That said CentOS can carry the burden of not being trusted for the long term as it is they who have adopted that they will focus on Stream instead. As i have mentioned RedHat is still releasing the packages in src format and nothing stops CentOS from building them. I believe this is a CentOS stearing commity choice more than it is a redhat choice. Someone please correct me if i am wrong.
Because who knows when RedHat might say tomorrow, RHEL 8 is now dead in 2 years, please pay for premium subscription for upgrades.
They simply can't be trusted to keep commitments they made.
As they saying goes, fool me once, shame on you, fool me twice, shame on me. People aren't up for being fooled twice.
CentOS is owned by Red Hat.
"As i have mentioned RedHat is still releasing the packages in src format and nothing stops CentOS from building them." ...except for the fact that RedHat owns CentOS, and RedHat told CentOS to stop building point releases.
I can't believe RedHat would do this. Countless thousands of people have spent countless hours creating CentOS 8 machines and now they will go without any updates, making them USELESS as long-term machines. It is selfish and a huge betrayal of the CentOS and Linux communities. This will backfire horribly as people run far away from anything RedHat, including RHEL. How can ANYONE trust a company who would do this?
+1 Couldn't agree with you more. We are switching completely away from RedHat because of this.
All the SAs and Devs of the world, unite! Do not recommend any IBM/RedHat products to your company or friends.
The reason we all loved Red Hat were the investments they made into ecosystems that would help them fight Microsofts monopoly in the long run.
Most likely all these former Red Hat projects will go like all the MySQL and Java stuff went under Oracle.
However, the need for RHEL is disappearing. Docker / Kubernetes and associated tools are managing what OpenStack never did, and once you've containerized your workloads the underlying OS is not nearly as interesting as it used to be.
So IBM will now be milking those who are unable to migrate.
+1 this change should never happen.
+1 this should never happen in the first place, I wonder what will RedHat do with so many backslashes from community
We (the company I work for) are already planning to transition to another Linux distro. Sad to see CentOS ditching their original philosophy of being an Enterprise grade free OS based on Red Hat.
This was our driver to chose RHEL and CentOS as our preferred distros. Now we have started to move away from this.
The way this change came out, breaking EOL, clearly gives us the impression that the constellation IBM/Red Hat/CentOS is not to be relied on in the future.
So, what exactly stopped RH/IBM from creating another distribution (or simply keeping both CentOS Linux and CentOS Stream)?
Instead they had to hijack an existing (very popular) project and completely change the central goal of said project.
There can only be one explanation. And that is NOT a technical one. Any technical reasoning could have been resolved by keeping both variants, or creating another distribution. They simply want to get rid of a free alternative to RHEL.
Dropping CentOS 8 just adds some salt to the wound.
Any talk about the potential use cases for CentOS Stream is meaningless. Those use cases are not the same use cases as those of the original project.
It's like hijacking Amnesty International, forcing a drop of the human rights activism, picking up rights for arctic fauna instead, motivating the move with "Those poor seals really need someone to speak for them".
Well yes, but you don't need sacrifice Amnesty for that.
RH/IBM, please just give CentOS back to the community! You can still drop any financial support. That's a much much better alternative than pissing off your customers and the community. I'm sure financing can be crowdfunded considering the situation for many current CentOS deployments.
And YES - most of your RHEL customers also use CentOS in its original form, believe it or not.
I've always professionally used CentOS/RHEL for servers and CentOS/RHEL/Fedora for workstations, and been a strong advocate when I have not been in a deciding position.
I guess that's going to change now. A shame, especially since I really like Fedora.
Yet another IBM dickhead executive with a game plan for a bonus after contributing zero, nada, nil to the development of CentOS.
That's the beauty of FOSS: Centos 8 Will be not available? Fine, we fork it. Cant't wait? OK, we are options available, Debian, Ubuntu Server, SuSE... I Think until December 2021 (when support for Centos 8 will end) we'll have a fork (Rocky, Cloud Linux...) up and running, so, no worry!
PS: Oracle had a bad history with Java, Open Solaris, Open-Office, Java, MySQL and counting... I don't trust them to adopt OL.
I've not tried Oracle Linux yet, but to be fair, looking into it shows that they've supported it as they said they would for 6 years now. If they are going to pull some bait-and-switch with it, then there's no sign of that yet.
So are you going to honour your initial EOL date for CentOS 8 -- which is why I installed it -- or am I gonna have scrap this work I've been doing to deploy this new server and use Debian instead?
Do what you like with CentOS, but you have one chance to redeem yourself in my eyes and that is to commit to your original EOL timeline for CentOS 8. The one that you showed me when I downloaded it. Do you understand how damaging it is to your reputation to give one EOL -- which people use to make business decisions -- and then change it drastically?
Do this, or join the ranks of Apple, MS, Oracle and other companies I won't go near with a 10 foot pole. At this point, you are no longer a trustworthy steward of technology and there is zero chance I or my company will ever purchase a support contract from you.
Well said Sean. I am thinking the same.
Most of my machines run CentOS 8 already, some left on 7 but what was done here is a joke, totally not on the enterprise level.
Some months ago i was thinking...
- Redhat (the company) does business with its free and free distribution called "RedHat", it's around 1995
- to make more service and profit contracts the RedHat distribution is transformed into Red Hat Enterprise Linux "RHEL" (circa 2003) no longer free (only the sources, as required by the GPL are available), to satisfy the community and to experiment, Fedora is launched. Too bad that in the production environments we don't want experimentation but STABILITY.
- CentOS was born as a recompilation of RHEL sources (about 2004),
- CentOS 6 and 7 spread a lot in server rooms,
- 2018 all CentOS employees hired by RedHat!
- 2018, RHEL 8 is released, recompiling RHEL 8 in CentOS 8 delays almost 1 year
- 2018 CentOS Stream is born: a version not recompiled from RHEL sources but for development and testing of RHEL (not suitable for production ... here we go again ...)
- 2019 RedHat is acquired by IBM ...
It seems clear to me that the IT giants (Google, Amazon, IBM, Oracle, ...) have now figured out the ways to get around or overcome the obstacles imposed by free software licenses and are making huge profits from it.
What will become of CentOS ??? Difficult to say but it is present in a large chunk of server environments and hardly does not stimulate the money appetite of the IT world.
Now I know it.
You're (unfortunatelly) so right... :-/
Boooooo! Way to derail all the projects I had planned to work on in 2021. Now my team and I have to work on cutting systems over to a new distro. Thanks a lot!
The CloudLinux distro will be called Project Lenix:
Rocky Linux also released timelines:
You killed CentOS. Thanks for that. Ubuntu Server, here we go.
This is bed!
We like CentOS.
New to Linux (very old in Windows AD environement), I choose CentOS 8 for my lightweight, stable "non Windows" apps and servers after "abandonned" Fedora...
What a disapoitment a month later to see that I choose the wrong Linux distro!
Maybe you should have continued support with FreeBSD instead of putting all of your eggs in one basket with Centos...
Centos and RHEL is now dead. That is sad. There was no confusion before you turned around in a moment. With this move, you make sure we and our customers will not be involved with any IBM/RHEL/Centos for the future to all our data centers.
RIP the last greatest Linux distro! cPanel have already announced they're moving over to Ubuntu.
They didn't say "moving". They're planning to support both Ubuntu and the CloudLinux RHEL rebuild, so there's continuity for existing CentOS users.
But for me the more important is that they said "Note that cPanel will not support CentOS Stream."
The abandon of currently RHEL infrastructure will be not now, but probably for any new version they will launch.
For me, the RHEL once more again broken the trust in it. This will probably change the ways that everyone will relate to most linux based companies including cPanel. The cPanel choose to abandon FreeBSD long ago and concentrate all their efforts only in RHEL based dists, changing everything to RPM and adapting all their systems to theirs standards (like systemd).
With the Redhat destroying the CentOS in order to gain some more money they rapidly changed the way not only the cPanel will have to focus and depend in just one dist and 3rdparty solution.
I've just rebuilt the virtual machine I used to provide to my mech. eng. students based - now - on Devuan Beowulf.
It has been complicated to solve all the library compatibility issues raised by a couple of binary-only solvers which were officially supported on RHEL alone, but in the end it seems I succeeded.
We do not depend on the RH ecosystem anymore. Bye, and, btw, it has never been real love ;).
The geniuses of marketing and business have arrived and have imposed their bullshit on technicians and fans. It is a story already seen, be careful not to end up like Commodore! Clowns!
No Centos, not you...
Dont touch centos pls!!
I use CentOS for stability. 🙁
...not anymore going by this news
I'm very deluded by this news and joined to the petition http://chng.it/TrM5f2ZK8S
Great move for ... SLES ... SuSE must be rejoicing about this move.
I hereby declare that if such move to stream happens we (as Boğaziçi University) shall no longer provide mirror services to CentOS repositories.
Slashing CentOS 8 lifecycle is a low blow.
Time to start looking for a new production Linux distro.
I really can understand someone want's more money, yet i can't understand they make decision like that, out of nowhere (even more after last year huge change).
So, now, even if i had money (and a veeery good reason) i would avoid it from now one. I'll go Debian or SUSE ...
I have used CentOS for 15 years. The reason? STABILITY and confidence in the team that maintain it.
I recently upgraded my main website to CentOS8 thinking there would be many more years of stability, but now I have to get out by 2021!
Yeah this is a hard one. For me, its about being able to provide a solid server at my small biz clients, using Linux distro that's not changing too fast. That way I can do an install and do very little maintenance over the life of that hardware. Rolling distro's always seem to be way too bleeding edge for me. Of course, now I do all virtual, but still, I want something not changing too much. CentOS 7 has been perfect for that. 8 has been good till today.
And its not that I don't want to support the company financially, its that my small biz clients can't afford enterprise pricing. Its already the biggest reason I try not to use MS server's for them just to do simple file sharing and/or a web server. (also the Linux ecosystem and ability to use different or customize your software) It just costs way too much for such simple things. And I NEVER need support! I've never once had to get support for Linux systems, so I'm not paying a monthly/yearly fee for something I don't use. I would pay a one-off fee to license a server, as a way to support their package repo maintenance, if it were reasonble, like $25/server. This is a Linux *distro* that's repackaging the *services* and open source software I use.
I don't know, I have very mixed feelings about this. Mainly I just feel like "Oh, there's yet another company pulling the rug out from under me and moving to charge more money from customers that *can* pay them, while turning their backs on those that can't afford enterprise pricing." Honsestly, I don't blame them for that, but I don't like it.
Never even heard of Rocky Linux, but now I'll be checking that out because I just don't like Ubuntu. Its fine, and 20.04 I like more, but again, they keep changing things like network configs on every version and its really annoying!
I could complain more and more, but I think its all been said by others.
For the last few years, IBM destroys everything it touches (there was a time IBM helped put humans on the moon).
If Gregory Kurtzer does achieve another distribution for the community, might want to look in to not selling it again, regardless of what the vendor promises.
Sadly, there's not much of a reason to stick with CentOS without "CentOS Linux". People come here for stability, knowing a version will be supported for a decade rather than a few years. At this point, I'd rather use Ubuntu as at least I'll have 5 years of support.
I am not in principle against the transition to a CentOS Stream only world (I say this as one using CentOS as a workstation, not as a manager of a huge array of production servers). The latest GCC toolsets being available months in advance of CentOS Linux is enough reason for me to want to use Stream. But I foresee some practical problems that need to be resolved for this to actually work, because these issues plague me now.
My CentOS and RHEL systems are all virtual, and unfortunately mostly running on Windows, and this is something I cannot change. I also need to be able to move VMs between platforms, and for a great many reasons beyond the scope of this discussion I use Oracle's VirtualBox, because anything else is just too painful.
The problem is that Red Hat are not shipping the VirtualBox drivers with their kernel, and are unlikely to decide to do so. The driver build process is complicated by the fact you cannot determine kernel code content from the initial version number 4.18.0 because the Enterprise kernels are hybrid and contain code from later versions, as far ahead as 5.x. Therefore it is almost guaranteed that every Enterprise kernel update at a point release is going to break the build of VirtualBox drivers (this happened with the release of RHEL 8.3, and is currently the case with Stream's latest 4.18.0-257.el8.x86_64). Every point release kernel (and most Stream updates) must be treated as an exception in the build, and VirtualBox code changes are required to get the build to work, and this takes weeks (based on past experience).
Coping with RHEL and CentOS Linux breaking VirtualBox every six months is one thing, but coping with Stream breaking it every few weeks is going to be unworkable for me (and others). I can only think of three possible solutions.
Red Hat choose to support VirtualBox users and bundle working drivers with CentOS Stream kernels. This will make the kernels different from RHEL, so I can't see this happening.
Migrate CentOS-Plus to Stream and put working VirtualBox drivers into the CentOS Stream Plus kernel, so people who need VirtualBox support can get it via the Plus kernel. This of course requires that Plus gets migrated to Stream, and I don't know if this will be the case, and I suspect not.
Add a VirtualBox package to the repos that people can install to get the drivers. Put it in EPEL, Playground, whatever, but put it somewhere. This is probably the most viable option, but requires that Red Hat perceive a need to do it.
I really do want to keep running with CentOS, but running Stream on VirtualBox is going to be a very bumpy experience unless the driver issues are somehow addressed.
I thought I couldn't be anymore surprised after all we had in 2020 - seems I was too optimistic...
Are you serious about such move? Please, reconsider!
While this is very unfortunate, especially the part about cutting short the supported timeframe. I was expecting this a long time ago - once RedHat acquired CentOS - I thought it was to shut it down. When that didn't happen immediately, I figured RH was doing a quiet death by not updating CentOS. When CentOS 8 was released, I thought it might be safe to use it again.
If CentOS Stream is the "development" branch of RHEL - then should we expect the Fedora project to fold soon too? They seem redundant at this point since I thought THEY were the development branch of RHEL.
Long term, I think this will backfire. A lot of bottom-up/grass roots shops would use CentOS to try things out. Executives would pay for RHEL so they could sleep at night. Without that option, I think devs will choose another distro that offers official support contracts.
Well, so long and thanks for all the fish.
...from the other side, from FAQ they do promise:
Q5: Does this mean that CentOS Stream is the RHEL BETA test platform now?
A: No. CentOS Stream will be getting fixes and features ahead of RHEL. Generally speaking we expect CentOS Stream to have fewer bugs and more runtime features as it moves forward in time but always giving direct indication of what is going into a RHEL release
What sounds good, hope it can be true...
I think i'll dig for my slackware's floppies...
gizas... what a wreck..
Best news ever. I hope this will end the dominance of ancient CentOS in VFX studios.
2020 continues to deliver the hits 🙁
This is a horrible idea and really puts enterprises using centos in a hard place. 6 goes EOL and now after all that hard work getting on 7 we get to do it all over again migrating to a new distro.
Thanks a lot.
Goodbye Centos. My organization and I need a stable Linux OS and not a Fedora clone.
Apparently the word "Enterprise" means nothing to the CentOS project.
You do you.
New name proposition: BeFoDuOS (Beta for Dummies OS)
Wow, this is bad news.
What a complete mess this will turn out to be.
you do realize that you've just thrown the Scientific Linux community (that was converging to CentOS 8) under a bus?
The right moment to announce this would have been *before* releasing CentOS 8.
You cannot change the EOL over night from 2029 to 2021 and not expect to loose your user's trust.
This is the very bad bad thing: announcing it afeter more than a year from the release of CentOS8.
We started from MMarch/April to install new server with CentOS8, and we migrated alla old CentOS6 Server to 8.
So in one year we must find a replacement, redfactor all our orchestration to the new distro, understand how to do thing wiht the new distro as close as possibile as we do with Centos, and migrate every CentOS8 Server......in One year?
The decision is made based on IBM greed probably and I don't like it but I am hoping there will be some other project that will rebuild RH 😉
Didn’t take long: https://projectlenix.org
First it was Spacewalk (no longer supported as of May 2020) since CentOS 8/RHEL8 isn't using standards anymore, and now CentOS is slated for an early EOL/EOS. Don't want to use RHEL due to its expensive support model/contract support, so it seems Ubuntu/Debian seems to bew the way to go now. What a shame.
It is interesting that I am hearing from other universities that are saying they will no longer purchase IBM servers because of this more. Time will tell if it is talk or they are serious.
In the words of Mr. Chow from Hangover "so long g** boys".
Terrible move you guys did there...
What a shame that I have to pull out oracle linux.
I did the same and surprisingly i also found that the uek kernel supports btrfs.
I will continue down this road till i see a reason to change direction. That said i would highly recommend you guys mentioning Ubuntu try Debian stable. It has most of the stable feature most of us want / need. And they have a CVE database which list backported fixes. So can be used in the enterprise just sadly no 10 years of support last i checked.
I wonder what organizations with thousands upon thousands of servers like CERN will do. They seem to have a worldwide influence on what a lot of labs and universities end up using through their computing grid (WLCG).
Well this news just tops off a "Great 2020"....
What is next on the chopping block ? Ansible AWX?
in fact centOS will die, less and less RPM will be developed, we will see an explosion of people using ubuntu going forward
This should have been expected since IBM acquired Red Hat ... lets hope that HP, now that ClearOS which is basically CentOS clone will loose the future path, will not acquire Canonical and do something similar with Ubuntu ...
I hope all the developers of all the tools you rely on start waking up to the fact that you are bathing in the cash provided by their hard work. I suggest the licence they use is amended.
What a pitty! 🙁
This move will move a lot of users away from Red Hat. Moneywise for Red Hat a stupid decision.
Centos = Stable,
Not stable = NOT CENTOS and:
Welcome Debian on board...
don't go to Ubuntu, better go to clean Debian
100% agree with this comment. Debian over Ubuntu any day of the week, I would actually say it is closer to CentOS in relation to packages.
Debian is rock solid, but you got stuck with stone-aged packages. If its not a problem, go ahead. If you want newer packages or had a brand new hardware, Ubuntu Server is the way to go.
Wow! Guess I will need to find a new server OS.
What a joke. It sounds to me like the only reason is greed. There's no rationality behind this decision. Many many people depend on the stability of centos. This only causes more confusion.
oh my god
When IBM bought RedHat ...
I tried telling fellow Fedora users that they may as well move now, 'cause IBM will f--- things up. Everybody kept telling me how stupid I was. LOL, I didn't even fathom they'd f-up CentOS.
They moved all the Fedora servers and in process of that they were downgraded; because f--- Fedora. Now, CentOS *is* Fedora.
RH and anything living off of it is now irrellevant. Thanks IBM! From a Fedora user since 1. RH Linux linux user since 4.x. Responsible for at least 20 commercial users across several companies.
Now for personal/community, I've already moved to Manjaro. For future enterprise who knows, SuSE probably.
Time to move ~600 servers (of which are 96% CentOS distro's) over to Debian I guess. RIP CENTOS EOL!
Red hat, this is a bad move for several reasons.
1. You are eroding the trust you built with oss. Your taking over the CentOS project was to prevent what happened with prior when the leader disappeared. To ensure it could continue in a stable environment.
2. CentOS is the distro people like me used to learn. When we become professionals, we help continue the legacy of using RHEL products in business. Ubuntu will now take that share when people start using ubuntu more. You will lose business for short sighted gains.
Feel Not Good!
Debian 10.7 can't compile boost 1.75, 1.74, 1.67, ...
Ubuntu 20.10 builds boost 1.75.
1.67 is already in Debian stable. 1.74 is in Debian testing. Both an apt install away - apt-cache show will help you search package descriptions for packages.
Essentially, you shouldn't need to build much from source / go searching for third party repositories like EPEL but it is an adjustment. [Debian and CentOS user for years.]
Don't forget to sign the pedition!
I learn how to be a Linux SRE thanks in large part to CentOS, so it's sad that it is now going away. Not surprising, but still sad.
You could have avoided much of the current bad feeling in the community by continuing to support 8 until it's planned EOL. Killing it off next year is just throwing a hand grenade into the room you've just left.
Does this mean, to use centos stream we need to use/install RHEL?! If yes, then where is the "free" and "opensource" part of it?
Wow, need to update my recent recommendation then. This will likely alter our client's course from RHEL to Debian 🙁
I echo everyone else - this is a really dumb decision which will lose Red Hat market share. Mind you that's what IBM does to the companies it buys ....
Honestly, it seems the current CentOS/RedHat boards do not quite understand the symbiosis CentOS/RHEL.
The community wants CentOS "classic". People are not interested in "low-cost" RHEL (or "no-cost for some special cases"). People are interested in using as much as they want, and pay as much as they need. Many companies need the support and are happy to pay for that. Some other companies do not need support at all, but are even happier to pay a bunch of RHEL licenses to a company that is doing good for the community. Some others cannot live without some of the extra Enterprise repos.
RHEL is big thanks to CentOS. Without CentOS, some people will move to other 1:1 binary compatible distros, but most of them will migrate to openSUSE/SELS or Debian/Ubuntu. It is going to be a big failure for RedHat. But there is still time for a rectification before too many people make up their minds. Soon it will be too late for RedHat. I am hoping for such rectification or reconsideration.
Honestly, from a many years RedHat fan.
Since they are backpedaling anyway with the whole Rocky Linux idea, why don't they just leave CentOS alone and then make Rocky the unstable community edition. The name sure fits the project.
We use a mix of RHEL subscriptions and CentOS. I know we'll be dropping our RHEL subscriptions because of this, we're not supporting a company that yanks the rug out from under us.
We just dropped 14 RHEL licenses (and 5 centos boxes). It took me a month to convince the management but i finally succeed. We had more RHEL machines than CentOS because we used RHEL for the desktop and CentOS for servers. Till the end of my IT career I will advice any future employer to drop any RHEL dependency. This is a red line that RHEL should never be crossed.
i also love cent os and Linux and also included in my top 10 list of best Linux Distros for Beginners
In thinking about this now for two+ weeks and considering the impact near term and longer term, I give Red Hat credit for being bold and making a move that they likely knew could make a wide range of people upset.
My guess is that they may not have known how widely CentOS was being used by large companies, individuals, small businesses, governments, and by hardware and software companies that build appliances or bundled solutions. They would only know if someone told them I assume, or maybe used it to negotiate a better price on RHEL.
While I see uproar, it feels like they are trying to create a fresher model to keep their core product alive and vital, which they need to do to survive in the cloud world in my view. And in parallel, reduce competition from a competitor that they helped create and make safe. I think they need to do both of these things to say vital and strong and continue to advance commercial open source business and technology development models.
I feel if they don't do these sorts of things they will end up being the next Novell - create a billion dollar business with a server operating system, and then flail downhill for 10 years when a competitor, Windows NT in their case, comes about. In Red Hat's case it feels like the public clouds are the biggest competitors, and if they don't act fleet-of-foot they will slowly be marginalized.
CentOS free usage seems to have gotten incredibly wide, with many people thinking it is close-enough to RHEL, and backed by Red Hat, so safe and secure. I am glad to see Red Hat not sitting back. As my dad told me there are three kinds of people in the world. 1) Those that make things happen, 2) those that watch while things happen, and 3) those that wonder, what happened. I think Red Hat is acting like category 1), after being in category 2) on this topic for the last 3-4 years.
> And in parallel, reduce competition from a competitor that they helped create and make safe.
They did not help create CentOS. They took control of CentOS and killed it off. Its place will be taken by another distro that will slide in to fill the gap, provided there is community support. Just so we are clear, RedHat uses the work of others for "free" because that is the open source licensing model. Debian is a perfectly fine "enterprise"-worthy distro that is completely community supported. So, regardless of what your dad told you, this is just a case of a company trying to flex its muscles. Losing goodwill in the open source community isn't a great plan. You only have to look at Oracle to see that. They have the best alternative to CentOS that is completely free and is a breeze to shift to, but people are hesitant because of their reputation.
It's always bothered me that everywhere I work (for-profit corporations) they use CentOS instead of Red Hat. When I'm job hunting the recruiters want me to have Red Hat + Satellite server experience and they don't even know what CentOS is. Perhaps this will benefit Linux in the long run.
Based on reading this, some articles and some other posts, I am starting to feel like the main reason Red Hat changed the CentOS situation was that too many large for-profit users were using CentOS when they likely knew they should be purchasing RHEL. I see names like Disney, GoDaddy, Rackspace, Toyota, and Verizon in articles, and in another one comment I saw Wal-Mart and Salesforce.com mentioned as being massive CentOS users. Feeling like people should maybe encourage them to be more fair in the commercial open source world.
This article mentions some of the for-profit users and makes some good fair points I think.
I work for a company that uses RedHat and CentOS. The reality is the support in general is not needed. When it is needed it is for an edge case, I think Redhat need to revamp the whole licensing model. Pretty sure they would get more support contracts that way. Oracle has an interesting model at the moment. You pay per call but they are giving their OS (clone) away. RedHat can go to a similar model and stop loosing customer base to the clones (Oracle/Rockey/Lenix...). Also what i tend to see is we use centos for Dev 90% of the time and RedHat when in PROD due to the support option. My logic would say just offer the OS and let the corporate entities select their support options.
I have used CentOS, Fedora and RHEL for over a decade. All this is going to do is send a larger number of people over to Oracle Linux.
Its an option but!!? Oracle?
Ubuntu Time!!! Because this movement i will go to move all my workloads in AWS from CentOS an RHEL to Ubuntu. Fail Red Hat for this movement.
As many already outlined the short-sighted intention of this move, i see another issue arising from this:
Crippling down a valid enterprise-ready distribution can be very irritating as far it goes for reliability and trustworthiness of "Linux" in general.
Many it-decisioners and stakeholders just wan't to see progress in their projects and finish up. A move like this, will cause many disruptions and workarounds which will negatively affect the outcome of the product. And this perception, these experiences, will also be associated with opensource in general (at least partly). So, in the end IBM/RedHat will lose double, because the it-world will think twice, with whom you lay into bed and to what conditions.
For me personally i'm negatively surprised. I already had a taste of IBM's new policy, when they stumped down CloudForms in favor of their product (IBM cloudmanager) and i also had a bad feeling when the old RHCE has been transformed to another Ansible marketing show. I love Ansible, don't get me wrong, but why sacrificing a well respected certification for another ansible cert, when they already had something related to this?.
So for me, this is unfortunately a decision point about how the *nix-journey will continue, but it's quite disappointing that IBM/RedHat is putting money above trustworthiness.
I don't think this comment makes a whole lot of sense.
Red Hat did not change anything in its version of Linux that they market as being the one to use for production workloads, with security fixes and hardware and software certifications, and a long life cycle. From Red Hat's view, which was documented on their website and the CentOS web site, CentOS is not an "enterprise-ready distribution". That is what they position RHEL to be.
It seems to me that people who wanted to get all the value of RHEL without paying Red Hat anything for it are the ones that appear to be upset in many of the posts. I think they need to ask themselves if they were being fair to Red Hat by doing this. I am personally shocked Red Hat did not do this many years ago.
I also am bothered by posts where people don't want to pay Red Hat for RHEL, and want to use CentOS for embedding in their products or for consulting work, where they are expecting to get paid for their product, or their work. Again this seems unfair to Red Hat and hypocritical.
Just sayin !
Maybe it doesn't make sense to you as there are different environments and different needs.
We are paying red hat for commercial rhel incl. satellite and so on. But the thing is, we simply can't use a rhel subscription everytime a testing instance gets spinned up. centos was a ideal starting place to explore, to test and simply start working. work which didn't had to be remodeled or modified just to match production. centos stream is simply not needed since there's already a valid candidate for it: fedora.
it wasn't the centos users only which profited from red hat like scavengers, which your are trying to portrait. red hat profited also very much of the community with inputs/bugfixes and a starting point / an entrance to "their world" including rh certification, representing them in a kind of way, recommending the whole rpm/rh/centos way of doing things.
to me its the wrong way of treating a community / a project which also accelerated their business and success.
ps: about your statement of "CentOS is not an enterprise-ready distribution". Literally lookup what CentOS actually stands for.
I see that the acronym uses the word "Enterprise" for the E in CentOS.
But in looking at the CentOS project website I don't see anywhere it states that it is an Enterprise Ready Distribution.
I looked at the description and the FAQ.
The main description starts off with "CentOS Linux is a community-supported distribution derived from sources freely provided to the public on..."
A "community-supported distribution" would not qualify in my view for an Enterprise distribution. Since the community has no accountability for response time or security patches as far as I can tell.
Saya sangat setuju dengan Anda Pak Bill.
Tidak adil untuk Red Hat/IBM jika kita hanya menggunakan CentOS dengan label RHEL versi gratis. Jujur, saya sendiri termasuk orang seperti yang Anda katakan, menggunakan CentOS untuk produksi, tidak hanya sebagai tempat pengujian. Saya termasuk bagian dari orang-orang yang tidak adil terhadap Red Hat.
Saya harus belajar mengapresiasi hasil karya orang lain atau perusahaan. Lagipula Red Hat toh masih menyediakan sumber terbuka yang mereka gunakan untuk membangun Red Hat, jadi mereka tidak mengkhianati apapun dalam hal open source software.
Ini hanya pandangan saya.
*ps: sorry saya menggunakan Bahasa Indonesia, bukan Bahasa Inggris. Silakan translate jika memerlukan artinya, atau skip jika dirasa tidak penting. Terimakasih
[Google translation added by admin]
I totally agree with you Mr. Bill.
It wouldn't be fair to Red Hat / IBM to only use CentOS labeled the free version of RHEL. Honestly, I myself am one of the people as you say, using CentOS for production, not only as a testing ground. I am one of those people who do not do justice to Red Hat.
I have to learn to appreciate the work of other people or companies. After all, Red Hat still provides the open source they used to build Red Hat, so they don't betray anything when it comes to open source software.
This is just my view.
* ps: sorry I use Indonesian, not English. Please translate if you need the meaning, or skip if it is not important. Thank you
[end of translation]
This is all about power !!!! And it's been done by all major companies. If the independent linux developers community would not have been as dispersed as it is over all kinds of distro's, but instead would unity around one core distro with a well thought out license, those companies wouldn't even dare or be able to pull this stuff!! I say one core distro to rule them all (all being some specialized forks).
There is still Oracle Linux. Oracle provide even faster updates then CentOS.
Oracle Linux 8 Update 3 which based on RHEL 8.3 was released last month:
It should be fairly easy to move from CentOS to Oracle Linux. However, I think the cPanel installer is currently not supporting Oracle Linux even though it is compatible to RHEL.
Red Hat kills CentOS as a viable production operating system, even for use as a test environment for apps targeting production RHEL environments. I will be migrating my CentOS 8 servers to Rocky Linux. And I will be scrapping plans plans to adopt OpenShift as RedHat might kill the free version of it too. I will be looking for truly free alternatives.
I am curious if you are working in a for-profit business, or consulting to a for-profit business, or for personal use or a non-profit or school type environment. If it is the non-profit/education I can see how you want to use as many no cost things that you can. But if it is the for-profit world, I don't understand why you would think it is not fair for Red Hat to charge for the products that they create, to help pay for the thousands of engineers they hire to create RHEL and OpenShift. Also, if you are in the for-profit world, do you pay others for their software. I am trying to understand the situation and your strong position. Thanks
I "used to" recommend RHEL, but not anymore. They becomes the bad guy. When CentOS proyect agreed to be attached to RH,
it was under the agreement that RHEL would give sponsorship and updates for 10 years to each CentOS version.
But they have violated the agreement, no matter if they are new owners, the CentOS community has been betrayed.
Never again will I recommend to my clients to buy RHEL, we must give our full support to Rocky Linux, and never again trust a Corporation that can change the rules of the game at will as we have seen.
Hi @ all!
Why not move to Oracle Linux now!?
Oracle does a very good job and they should seize the opportunity.
Yes, Oracle Linux is a very good operating system, very stable, fully RHLE compliant. With the decision on Centos Stream, I removed centos from 3 of my servers and installed Oracle Linux.
But there is a problem. Oracle is also a corporation. If Oracle Linux takes all the acceptance and becomes as used as Centos was while it lived, Oracle could do the same thing it did with mysql or the same thing that RedHat did with Centos.
The lesson is clear: community projects should not depend on corporations.
Rocky linux won't make that mistake again.
There is no point in going into the redhat ecosystem, time for a change, Debian or Ubuntu, although for those who know each other better Debian because it is a system created from scratch, and Ubuntu is based on Debian but it is easier to use. I have been using Centos for 15 years and decided to migrate to Debian
I have to uninstall centos 8-stream quickly. It is so unstable I have never seen anything like it. You type a letter it gives you about 15 of the same character. It is frustrating to work to work with it.
I don't know what to do. I have just come to the server world.
"The CentOS Project
The CentOS Project is a community-driven free software effort focused on delivering a robust open source ecosystem around a Linux platform."
Have you ever read that??? It is at footer of centos project site.
I could sugest an update, because CentOS Project IS NOT ANY MORE a community-driven free software. Now, Centos stream is an effort from redhat to kill Centos as we used to know it.
The decision about centos strem, for sure, was not taken from the community 😉 It was taken by new Red Hat owners.
yes, information needs to be updated, centos are no longer centos
CentOS stream is not an option for me.
Up to now I use CentOS 7 on my private web and mail server and I like the stability and the long support. I am very disappointed that RedHat is choosing this path. 🙁
You have just lost a dedicated tester. There isn't a single person I know that will stick with this change.
don't waist your time on CentOS Stream crap..!!
We can switch to Debian Linux
Let's say goodbye to Redhat and Centos and can switch to Debian:
-- "Debian takes security very seriously. We handle all security problems brought to our attention and ensure that they are corrected within a reasonable timeframe. Many advisories are coordinated with other free software vendors and are published the same day a vulnerability is made public and we also have a Security Audit team that reviews the archive looking for new or unfixed security bugs."--
You are going to loose soo much customer base after this you dimwits!
So, we are all turning in to beta testers for RHEL now tracking ahead rather then level pegging in some areas and behind on others (given stable hand-me-downs from RHEL).
And folks are moaning? for some, yes!, I agree and do use CentOS in production as just as equally good as RHEL without the support packages. But now tracking ahead I can see issues.
I was kinda happy with the notion of using nightly builds as being bleeding stable edge and even then just mainly stick with a good production stable.
CentOS is becoming the RHEL testing playground which will in the long run breed a larger, faster and more efficient or adaptable development force rather then attracting production users.
I guess for them it equates more within the monitory system and seems like CentOS is paying its dues, where as they, RHEL, just see CentOS presently as a rebranded RHEL with no use to them.
This isn't the first of changes towards a more monetary-driven direction.
In 2014, Red hat payed off a non-profit board and re-badged some of them.
Now, they've spoilt them enough to drop the stable product altogether, replaced by one with a contastantly moving target, the very oppositive of what CentOS users expect.
As for the lousy license free proposition: have Red hat realised their money lies in 16+ server production installations?
It's preposterous to presume an unstable codebase will be used in production installations.
You just don't change the rules of the game mid-game, that's what this is.
No thank you, I'll happily pass on your fantastic support.
Well done Red Hat a.k.a. IBM!
Nobody will buy this "shifts focus to" bullshit in this community, save it to your marketing guys who can't change a lightbulb."CentOS Stream" what you can test for free if that is your hobby, CentOS the stable version you can use for systems what matters. Two completely different things. IBM simply don't understand how open source works, and what it's benefits.
I am thinking it will be technically possible to keep patch level same as RHEL. May special plugin for DNF can be developed. It should get versions of packages from RHEL sources repo and then do the update in CentOS accordingly.
rhel -> centos/free (3-year support for server/desktop) -> centos stream -> Fedora
get more value.
I've been running Red Hat software in various forms since shortly after the Halloween release. It's been a good, long run but it looks like it's time to move on. Disappointing.
Here you can read the truth: https://www.theregister.com/2021/01/26/killing_centos/
"Brian Exelbierd, responsible for Red Hat liaison with the CentOS project and a board member of that project, has told The Register that CentOS Linux is ending because Red Hat simply refused to invest in it."
Instead of finding a way out for everyone, Red Hat tries to dictate to users how to use CentOS. Simply use CentOS Stream.
These issues are the reason I depend upon FreeBSD for all my Server needs.
You've (CentOS Community) suffered a sucessfull Embrace, extend, extinguish event .... Seems IBM Learned something from the Micro$oft debacle, after all. One thing is vitally clear, IBM/Red Hat have lost the moral right to call it a "Community ENTerprise Operating System" .. call it Blue Pill and be done with it.
Wow so disappointed Debian here I come.
What a stupid suicide! Stop it before it's too late.
I mean, this is kinda bad I guess. Is there like a clear comparison between CentOS 7 and CentOS 8 vs CentOS Stream? What features are we missing from CentOS 8 in Stream? Are they going to be ported? My biggest concerns are is there going to be the same support or stability wise? Lots of questions still to be answered regarding this move by Red Hat.
I mean hosting companies will face a blow as many prefer it because of its similarity to RHEL!
centos is dead, as a copy of fodera. so sad !!!!!!
I was delaying it as much as I could as I grew up with Red Hat, then continued with Centos and now I see that this delay was just a terrible mistake. I should have moved to Debian looong ago and wouldn't have many problems with stupid ideas from Red Hat. Also, Red Hat was taken over by IBM 2 years ago - I think this was the reason of the shift in strategy. I don't like Red Hat's collaboration with Microsoft too, it can only get things worse, definitely moving off to Debian now.
It's a pitty, After years of activity with Centos I will return back to Debian !! I made a big change years ago, I'm really frustrated from this decision.
Almalinux or RockyLinux? Which will end up with best support/development/not become RHEL in a few months?
I wouldn´t trust Almalinux it´s RHEL(CloudLinux) all over again. I would go for RockyLinux
What if this is just a strategy by IBM to get the for-profit freeloaders only using CentOS out of their way - seems to be working.
thx for everything and good bye.
it was a great decade with you.
which distro most of you choose now ?
We developers don't typically work for free and more than 80% worldwide have an annual income more than a Redhat yearly subscription.
We gladly spend tons of cash to buy Macs and phones.
The work the servers that run CentOS do isn't free, neither do we charge for just electricity and hardware.
It's only fair that you pay for a RedHat subscription now if you want stability and you desire that platform.
If you want to be adventurous be with CentOS Stream.
Any developer/administrator above 25 years old has profited a lot from free CentOS and few have returned anything, how much more companies.
I have a singular system that run centos and three that run redhat. This is going to make me move away from centos all together.
The easiest path for me is to find out if we can use ubuntu LTS for that box because getting approval for purchasing is a lot more difficult than not requiring it.
Thanks for writing this article, and all the best to the centos community, but I'm dipping out here.
This is a sad but predictable turn of events, I saw it coming and switched when IBM got it's foot in the door. It's also not the worst to happen, that's coming.
SLES seems to be the only option for Enterprise now. Ubuntu is a good desktop solution but it's philosophy is not Enterprise Server worthy.
For all your Non-Enterprise solutions take a look at Devuan, it is currently the best option for the future of FOSS. https://www.devuan.org/
I am a scientist, and frequently do "free lance" research at home using my computers running CentOS. I do not make money from this, it is gratis. To have to BUY RHEL subscriptions and pay for support I don't need/want is forcing me to look elsewhere. CentOS is a stable environment, I left Fedora a decade ago because it is not. I do not have time to keep updating/upgrading every 6 months. This is greed, and no one else is getting my money. I heard Scientific Linux (based on CentOS) already has gone. I'm left with-- Ubuntu. Nice Job, IBM.
Did you hear about Rocky linux?
This is great hopefully Ubuntu will benefit and become the Linux de facto, cent os and red hat are just fragmenting the linux ecosystem more than it need to be for financial gain.
Linux is fortunate to have a plethora of distributions from which to choose. It's just not that hard a choice to move away from CentOS to another stable Linux version. IBM on the other hand appears to have executed a coup - moving CentOS from following RedHat development to leading Redhat development, thus potentially saving themselves considerable time and effort (if it works out this way). I'm no fan of paying for 'stolen' open source software and I never would pay for Linux which I have used since I bought a set of Slackware CD's from TransMeta. When I was still a sysadmin, we used CentOS on IBM hardware and had to pay IBM to support that hardware, mostly because of a dearth of adequate documentation and a relectance by IBM to provide any information without the meter running.
It is my understanding that centos-stream is an online operating system. In light of all the damn hacking online (I am trying to keep it clean here! My choice of words are not appropriate!), anyone who thinks that centos-stream online operating system is the future is making a GRAVE serious judgment call. I have been a computer and electronics geek since the 1970s and I am telling OFFLINE is the ONLY way to maintain absolute protection from hacking. Given that, I think you can see that I AM CORRECT IN THIS DEDUCTION. A masterful FIREWALL is THE ONLY way to keep the HACKING COMMUNITY OUT OF YOUR ONLINE SYSTEMS. NO MATTER WHAT YOU SUPPLY FOR SECURITY IN YOUR centos-stream WILL NOT WORK. It WILL FAIL, and your users will end up getting hacked. ReTHINK the absolute constantly need to be online, and switch back to the OFFLINE system that allows users to protect their systems from hacking, using online connection ONLY to download emails, and a severely restrictive browser through absolute FIREWALLING of ALL SOCIAL MEDIA (including google), which is most likely where people's information gets stolen.
SLES seems to be the only option for Enterprise now. Ubuntu is a good desktop solution but it's philosophy is not Enterprise Server worthy.
Red Hat kills CentOS as a viable production operating system, even for use as a test environment for apps targeting production RHEL environments. I will be migrating my CentOS 8 servers to Rocky Linux. And I will be scrapping plans plans to adopt OpenShift as RedHat might kill the free version of it too. I will be looking for truly free alternatives.
This is great. Hopefully, Ubuntu will benefit and become the Linux de facto, centos, and red hat is just fragmenting the Linux ecosystem more than it needs to be for financial gain.
My great love will die .. no more Centos for me ...
Centos is installed on my laptop and on severel Servers ..
May have 8 HP DL380 G9's and a 6 cage 3par with all switches available to donate to the project soon. Currently in Telehouse,NY datacenter running LIVE traffic.
Somebody let me know how I could donate them to the non-IBM centosproject or rocky projects otherwise we will pay a scrapper to take them as we have with several other datacenters.
Will have many G9/G10 servers available in Germany as well in the coming months.
I completed migrating all my CentOS 8 servers to Debian 11.
A bitter ending and bitter news for your friends
I have used CentOS, Fedora and RHEL for over a decade. All this is going to do is send a larger number of people over to Oracle Linux.