Want to read Slashdot from your mobile device? Point it at m.slashdot.org and keep reading!

 



Forgot your password?
typodupeerror
×
Red Hat Software Linux

CentOS 8 Ending Next Year To Focus Shift On CentOS Stream (cyberciti.biz) 136

Well here is a surprise for those that have long used CentOS as the community-supported rebuild of Red Hat Enterprise Linux... CentOS 8 will end in 2021 and moving forward CentOS 7 will remain supported until the end of its lifecycle but CentOS Stream will be the focus as the future upstream of RHEL. From a report: For those relying on CentOS 8 to enjoy the reliability and features of Red Hat Enterprise Linux 8 but without the licensing costs, etc, that will end in 2021. At the end of 2021, CentOS 8 will no longer be maintained but CentOS 7 will stick around in a supported maintenance state until 2024. The CentOS Project will be focused moving forward just on CentOS Stream as the upstream/development branch of Red Hat Enterprise Linux. CentOS 8 users are encouraged to begin transitioning to CentOS Stream 8. The CentOS Project announced this shift in focus today via the CentOS Blog. Red Hat's announcement meanwhile is promoting the change as beneficial to CentOS Stream.
This discussion has been archived. No new comments can be posted.

CentOS 8 Ending Next Year To Focus Shift On CentOS Stream

Comments Filter:
  • by kriston ( 7886 ) on Tuesday December 08, 2020 @03:20PM (#60808836) Homepage Journal

    So no more CentOS LTS releases? The article's author is right: this is definitely going to move people to Ubuntu Server LTS because RHEL isn't exactly cheap.

    • So no more CentOS LTS releases? The article's author is right: this is definitely going to move people to Ubuntu Server LTS because RHEL isn't exactly cheap.

      This seems like a good way to shrink the user base of RHEL in general. Without the entry level CentOS, there will be fewer customers willing to shell out the cash for RHEL.

      • by Yeroc ( 125826 )

        Exactly. Running CentOS gave you a trivial upgrade path to RHEL when you needed/wanted support. It also exposed people to the RHEL ecosystem which is significantly different from Debian / Ubuntu. This will no doubt decrease their market share over time.

        • It's going to certainly seen a significant migration over to the Debian ecosystem, that's for certain.

          I built a few systems on CentOS over the years, no real complaints other than that I've always preferred Debian over the Redhat distros. I've actually been mucking about a bit with FreeBSD, just have to get my head around it, but it's nice to work with a very Unixy-y operating system where a lot less liberties have been taken (looking at you, systemd), but it's a different beast with some different tools. I

          • by laktech ( 998064 )
            I'm deploying some infrastructure for a personal project and decided to launch it on FreeBSD. I've really enjoyed it and will continue to use this platform on future efforts.
        • That's a good point. And one that has kept me from wanting to take any sort of dive into the RHEL ecosystem.

          Not because I'm lazy.... well maybe because I'm lazy.... but the learning curve felt steep enough in the very few times I've considered it that I've never bothered. I'm sure this probably works in every-which way.

          Back in the late 90s/really early 2000s my first experience with Linux was Slackware. It was Slackware because my friend had a Slackware CD I got to use to try it. It's what I learned it

        • Yup, I wonder what this will do to shift the market in 5-10 years. I started with Red Hat because it was free and available (back in the RH 5 days, pre-Fedora), just without updates, but back then that wasn't a big deal with security as it was already secure enough. Post RH9 I moved to Fedora (Core at the time), and then happily embraced CentOS when it was available for longer life than Fedora. Great for long-term server stuff. I've got some locked down CentOS5 and CentOS6 servers, with multi-year uptim

          • pretty similar path here. Have used/supported CentOS for years because of the path to Redhat(which I picked up in the mid 90s) but had already started moving to Debian and Ubuntu, have been running 20.04 LTS on all my workstations with kubuntu since it rolled out.

            This will mean I finally completely take the plunge debian/ubuntu all the way. Plus my play stuff ;)
        • by kriston ( 7886 )

          How trivial was that upgrade to RHEL? It didn't seem so trivial the last time I did it; it was more of a complete re-installation.

          • by Yeroc ( 125826 )

            Yes but any automation scripts and any of your binaries are pretty much guaranteed to work without re-compiles/re-testing.

    • ...as in Oracle, who would like nothing more than to become "free RHEL" in their place, complete with the ability to add support later. This is a very bad move.
      • I haven't had a whole lot of dealing with Oracle. Is Oracle the bad part of the move (I know they are a big player, are they as evil as or perhaps more evil than RedHat?)? Or just a bad move for RedHat, dropping something that still has a following, and then potentially facing competition?

        I'm just surprised to hear as relatable a name as CentOS loosing support.

        You would think having people living within your ecosystem would be a good thing, paying or not. Because if they are in your ecosystem, they aren'

        • by Wolfrider ( 856 )

          > For the regular guy,;I imagine (perhaps naively) that some entity will end up picking up the slack

          --I KNEW it was a bad idea for Scientific Linux to stop their fork awhile back...

      • by kriston ( 7886 )

        Oracle offers a script that automagically changes your CentOS installation to Oracle Linux quite easily. It's practically the same as CentOS except for the kernel.

        • by PCM2 ( 4486 )

          It's practically the same as CentOS except for the kernel.

          And Oracle isn't messing around with the kernel they ship, either. They know who their customers are and they don't ship sketchy kernels.

      • by PCM2 ( 4486 )

        Well it's Oracle, so it's certainly not going to be free. I once asked Wim Coekaerts who actually paid for Oracle Linux. He said "Oracle customers." And that's probably about right. You wanna get all your support from one phone number? Go for it. I mean, already who do you call when you want to tune your OS for an Oracle database -- Red Hat or Oracle? So you might as well just install their OS image, too, and Oracle actually does a decent job of shipping patches, too.

    • > So no more CentOS LTS releases? The article's author is right: this is definitely going to move people to Ubuntu Server LTS because RHEL isn't exactly cheap.

      I've slowly moved from 100% Fedora + CentOS to all Debian 9, then 10, ever since IBM announced its acquisition. I knew that marked the end of the open community.

      Having sworn off Debian in 2005 as unsalvageable, they've changed their ways and I've changed my opinion. Now, almost everything I loved about Fedora is in Debian and almost everything I

    • CentOS has support? :)
      I mean it's a completely free as in beer and free as in speech product. What support do you expect. I'm being a little tongue-in-cheek here.

      While I totally understand why those with an enterprise bent are upset, let's look at this a little closely. Centos Stream will have branches for each release. Centos Stream 8 is going to be around as long as RHEL8 is. So the "LTS" nature of centos does not go away.

      So what changed? Right now, if I run a centos 8 server, I get critical and sec

      • Having centos stream 8 being the upstream of rhel 8, means taht centos stream 8 will get the package updates before rhel 8.3 will. In a dev/test/stage/prod release cycle, this allows one to use centos 8 in the earlier stages to confirm changes and adjust to them BEFORE rhel 8.3 is released, and taht critical customer upgraded on day one and found an ABI compatibility problem with your product :) So there is a potential silver lining.

        I have no interest in being a beta tester for Red Hat. I see how that worked out for Windows 10 Home users.

        • why do you assuem you're beta testing a package that is released on centos stream 8 before it's on rhel 8.. the rpm has already been tested, verified and ready to go. do you think it gets any more additional testing before it's released on a rhel8.x point release?

          Because in most cases it does not . If you are beta testing the apckage on centos 8 stream, then by that same logic youa re currently beta testing rhel 8.3 packages when they are released.

          • why do you assuem you're beta testing a package that is released on centos stream 8 before it's on rhel 8.. the rpm has already been tested, verified and ready to go. do you think it gets any more additional testing before it's released on a rhel8.x point release?

            I assuem so because I subscribe to Occam's Razor. This is exactly the model tried and adopted by Microsoft: free beta testing for their paid "Enterprise" users.

            Because in most cases it does not . If you are beta testing the apckage on centos 8 stream, then by that same logic youa re currently beta testing rhel 8.3 packages when they are released.

            This fails the logic test. You claim a pre-release stream is not in any way a trial balloon for the RHEL distribution. If you are correct then there is no legitimate purpose for changing CentOS' entire raison d'etre into a new, separate, different upstream product. People use CentOS to get RHEL binaries without the costs.

            Admins adopt CentOS fo

  • You will have to reverse engineer RHEL9 like the PC Bios in the 1980s if you want it for "free".
    • by xack ( 5304745 )
      This will backfire hard. They will get a short term revenue bump from management panic buying rhel licences until the open source community has forked successfully. I expect RHEL 9 not to be released and Red Hat just minimally supporting RHEL 8 for the remaining companies that haven't escaped yet.
    • by dfghjk ( 711126 )

      The "PC Bios in the 1980's" was literally published in complete source code form, including comments, in a reference manual that IBM sold in retail stores.

      So, if you mean like that, then sure. Reverse engineer by literally reading the widely available source code.

  • I thought that IBM promised that they weren't going to weren't going to mess with Red Hat's core business model as part of their merger? This looks like a pretty blatant cash grab to force users to get more signed RHEL Enterprise support agreements to me.

    I'd also imagine that this is going to backfire on them... all of those folks in AWS who are using CentOS images now will just migrate the Amazon Linux instead. It's pretty much RHEL compatible anyway.

    • by oldgraybeard ( 2939809 ) on Tuesday December 08, 2020 @03:26PM (#60808868)
      Wait! you think big corps keep promises. Foolish one ;)

      I was wondering what would happen when IBM bought Redhat. Now we know, IBM will drive the SaaS subscription business and drive a stake in to the community.
      • I'm sure I remember lots of comments on this site when IBM bought Redhat along the lines that IBM would kill the golden goose.
        This looks like the start of that process to me.
        • Very true. I do wonder which IBM(I recall IBM is splitting in to 2 businesses) will get Redhat.
          I am guessing the one with the Technology(designated survivor ;), would not make much sense for Redhat to go with the marketing dead weight and the debt(ready to go belly up).
          I am guessing the surviving part(Tech part) of IBM will keep Redhat.
          • But this whole deal does smell like it is being driven by marketing and mba dweebs.
            Could be that Rednat is the SaaS you can survive slop to the marketers and mba's leaving.
    • I thought that IBM promised that they weren't going to weren't going to mess with* Red Hat's core business model as part of their merger?

      * For sufficiently loose definitions of "not going to mess with".

    • For any use case where CentOS was the preferred solution, there's several other solutions right within grasp from a technical perspective. Whether management will allow the technical folks to make that decision is another thing entirely.

    • by dnaumov ( 453672 ) on Tuesday December 08, 2020 @04:31PM (#60809160)

      You got this backwards. Anyone who's been following the IBM/Redhat developments and actually paying attention knows that Redhat has basically engineered a reverse-takeover of IBM. This is basically confirmed by current employees, both Redhatters and IBM old-timers.

      • by Yeroc ( 125826 ) on Tuesday December 08, 2020 @04:39PM (#60809204) Homepage

        That may well be but it matters little. There's no escaping the relentless demands of shareholders on a blue-chip stock.

      • You got this backwards. Anyone who's been following the IBM/Redhat developments and actually paying attention knows that Redhat has basically engineered a reverse-takeover of IBM. This is basically confirmed by current employees, both Redhatters and IBM old-timers.

        “In place of Dark Lord, you will set up a Queen. And I shall not be dark, but beautiful and terrible as the Morning and the Night. All shall love me and despair!”

        -Lady Galadriel

  • by chmod a+x mojo ( 965286 ) on Tuesday December 08, 2020 @03:32PM (#60808894)

    I mean it took what, a year+ to get even KDE or XFCE into the community repos for version 8? Xfce4 was there like 6ish months late, but in an only half usable state.
    Up until then you had GNOME shoved down your throat as it was even installed by default. You had to actively select not to install Xorg and GNOME.

    They shit the bed hard when doubling down on "GNOME is the best DE EVAR!!!1 and we refuse to help support anything else even though our users demand it...".

    Cent 7 was good, supported a massive amount of desktops both natively and through the community repos, and was all around more polished than 8 has shown to be. Therefore them giving up on "normal" 8 is not a big surprise.

    Plus what enterprise is going to even look at them now that they want to basically become a rolling release? That isn't stability you can trust for mission critical software or operating systems.

  • I don't mind a rolling release for a desktop OS. In fact, I prefer it.

    However, not all server workloads are containerized, and, until the point when they are, it's not appropriate IMO for a server OS to have rolling releases.

    We deploy software against very specific versions of CentOS 7, and we do that for a reason. Haven't even upgraded to 8 yet. A lot of this software is not under our direct control, and we have no way to know what will happen, without VERY extensive regression testing, if an update cau

  • there better be an way to migrate to Stream with DNF

    • I did this on one of my test machines.

      I just added the Stream repos, did a dnf update, and... everything got all fucked up.

      Thankfully, we hadn't moved anything real to 8, everything is still safely on 7.

  • My guess... (Score:5, Insightful)

    by Junta ( 36770 ) on Tuesday December 08, 2020 @03:51PM (#60808974)

    So for one, I think CentOS doesn't make much sense as a RedHat owned and derived project. It used to, but when RedHat took it over, we have RHEL and then RedHat paying a team to go through and effectively de-badge their own software and produce an easier to download, easier to mirror, and easier to update knockoff of themselves? A knock-off that is *far* superior to use unpaid than RHEL is, and without any easy way to sell support to an in-place installation, when support is the crux of your business? A situation made all the more embarrassing by Canonical and Oracle, both of whom have the 'easy to acquire and update' with upsell to support in-place without reinstall or complex migration process.

    The simplest approach? Just make RHEL like Oracle Linux. Free to download and update, with optional registration and support.

    I would not be surprised if this is on the table, but there is a disagreement internally. Someone else is all onboard with canning CentOS, but they imagine keeping RHEL as-is would just force all those CentOS users to convert to paying RHEL customers, rather than scare them off the RHEL ecosystem entirely.

    I think this gap between announcing end of CentOS and announcing their RHEL plans is to take some chance to examine the fallout of the decision to have data to back one call versus the other.

    I personally think they'd scare more people into Ubuntu than they would convert to RHEL, and that would be a very bad situation for RHEL long term. So for their sake they really need to make RHEL more easily available. Otherwise this will be a big boon to Canonical and Oracle as people either jump ship entirely or jump ship to a free distro with closeness to RHEL and an in-place commercial upgrade path. Of course, *todays* terms with Oracle Linux are nice and favorable, but it would be a bit risky to presume they won't screw you down the line (they are Oracle after all, very fond of intimidating made up invoices for software you don't even use).

    • I guess the question they must be asking themselves, how many of those scared into Ubuntu would have ever paid for RHEL anyways, if CentOS continued to be supported? If the answer is almost none, then they didn't lose anything, and possibly gained some RHEL subscriptions from those using CentOS today who don't want to jump ship (perhaps they have some RHEL, and some CentOS, now they'll convert to all RHEL).

      • by Bert64 ( 520050 )

        It's not just about immediate users, its about mindshare and experience.
        Someone who uses CentOS gains experience of RHEL, if they move to an environment with more money to spend they might buy RHEL. If that user instead used Ubuntu they would be unlikely to use RHEL later.
        Similarly some smaller shops might start with CentOS and then buy RHEL later as their business expands.
        If they close things up, the barrier to entry is higher and they lose these potential customers.

        • by Junta ( 36770 )

          To expand on this.

          Back in the early 2000s, RedHat was pretty much 'the' distribution to run on desktop and servers. Then they did the commercial-only RH and the clearly alpha testing Fedora Core. This more than anything else opened the door for Ubuntu to sweep the desktop linux market and in turn become the defacto distribution for a lot of these desktop users as they developed server side applications that started requiring Ubuntu in some form server side.

          In my niche, there was a pretty even split between

          • by Wolfrider ( 856 )

            > Now suse is almost unheard of

            --SuSE is more targeted toward European business:

            https://linuxhint.com/redhat-v... [linuxhint.com]

            • by Junta ( 36770 )

              Yes, that was the case in my niche as well, but within the niche, even in Germany and other European installs, most aren't SUSE anymore that used to be suse some time back. Might not be universal across other market segments, but at least in our long time suse customers that were mostly europe, most have moved to rhel at some point.

              Interestingly enough, one of those customers jumped ship *because* IBM bought RH and they wanted to buy from IBM... for whatever reason.

      • I suspect you're right, but the problem is, without the Centos pathway to RHEL, those people that didn't pay are now more likely to pay Canonical than they ever were to pay Redhat/IBM (assuming lots will now jump ship onto Ubuntu).

        I suspect the bean counters looked at how much it cost to build and maintain the Centos product and decided it wasn't worth it. However, by making the EOL time for Centos 8 *shorter* than Centos 7, they've just pissed a load of people off. All Centos 8 users now need to move off t

    • by ufgrat ( 6245202 )

      Sorry, but it will be a cold dark day in the punitive afterlife of your choice before I trust Oracle with anything that isn't tightly wrapped up in contracts-- or have you not had to bother with licensing Java or MySQL lately?

      • by Junta ( 36770 )

        As I mentioned, long term Oracle is a bad choice of vendors, as Oracle customers are Oracle's favorite victims of completely imaginary invoices and threats of court if you don't pay for things you don't even use.

        However, I have already seen it happen, a lot of shops talking about migrating to Oracle, and in my business, yesterday we had three customers upon getting the centos-announce email ask about our thoughts on Oracle Linux instead of CentOS. That's just the ones that would bother to ask.

        Oracle is goin

  • CentOS Blog Link (Score:5, Informative)

    by TemporalBeing ( 803363 ) <bm_witness@@@yahoo...com> on Tuesday December 08, 2020 @05:03PM (#60809304) Homepage Journal
    From CentOS Themselves: https://blog.centos.org/2020/1... [centos.org] Sad, that essentially makes CentOS useless.
  • Comment removed based on user account deletion
    • by Junta ( 36770 )

      Oracle had 8.3 out before CentOS did.

      Oracle is very CentOS-like in how they let you get the distro and update it.

      Oracle can get enterprise support without a reinstall or 'transition' script to make it into a proper distro (support from Oracle instead of RedHat though).

      Is it a risk that Oracle will screw over a future edition, because they are oracle, but IBM *appears* to be screwing over CentOS *today* without a clear announcement on what next.

      Canonical promises support on Ubuntu LTS, so there's another ent

      • Canonical promises support on Ubuntu LTS, so there's another enterprisey option.

        Ubuntu LTS promises 5 years of security updates + another 5 years if you pay.
        RHEL/CentOS 7 promised 10 years of security updates + several (7?) years for paying RHEL customers.

        Redhat/CentOS was more attractive for the "if it ain't broken, don't fix it" philosophy. For example, every time I update my laptop to the next Linux Mint LTS version, it takes me a weekend to get everything running again because some configuration file format and command-line tools that scripts depend on changed. (I make notes every

  • Oracle Linux (Score:4, Informative)

    by innocent_white_lamb ( 151825 ) on Tuesday December 08, 2020 @05:44PM (#60809444)

    Some of the folks on the Centos Mailing List are debating the merits of moving their installations to Oracle Linux.

    Yes, I know. Oracle.

    But apparently (and I never knew this until today) Oracle Linux is actually free to download and install and use. They just charge for support, like Red Hat does.

    So if what one needs is a free rebuild of RHEL, Oracle Linux might be the thing to have.

    • I've been using Oracle Linux (basically RHEL 7) for a few years now. They have their customized kernel that you can opt out of, beyond that it works exactly like RHEL. The only reason I switched back to Centos for new installs is because an Oracle account (free) is needed to download the ISOs, and every time I log in with it they will call/email/send passenger pigeons for months afterward trying to get me to buy support.

  • by caseih ( 160668 ) on Tuesday December 08, 2020 @05:46PM (#60809450)

    Most of us used CentOS because it was stable, supported, and cheap, relying on RedHat to do the work for free as it were. I'm sure many of us would be willing to pay some amount for RHEL updates for small business and even home installations. If RedHat offered some kind of lower paid tier to replace CentOS many of us would jump on that. Maybe with different branding to differentiate it from RHEL enterprise offerings.

    If not, I'm sure we'll see the resurgence of forks like Scientific Linux.

    • Even in a larger company, CentOS became easier to deploy after virtualization became the norm. Looking into using somebody else's computers, aka the cloud, makes it straightforward to deploy CentOS rather than RHEL - license tracking and usage can become a nightmare when the number of instances are dynamic. Privative Software that the company uses only supports few OS, and even then, few variants (patch levels). To add injury to this situation, if something does not work on CentOS, somebody in this company
      • by b0bby ( 201198 )

        I'm in a small company, and for us this is huge - we don't have a licensing department, we just want to deploy and not worry about it. CentOS has been great for us - super stable and no headaches. We'll probably switch to Ubuntu LTS which I'm sure will also be great, but there will be a slight learning curve.

  • by Koatdus ( 8206 ) on Tuesday December 08, 2020 @05:52PM (#60809478)

    It is called Fedora.

    People use CentOS because they want a stable RedHat style Linux but don't need the support.
    Realistically, once you have used Linux for a couple of years how much support do you need.

    • by ufgrat ( 6245202 )

      If I understand it correctly, Fedora is "unstable", but in an "expected to work" way. CentOS stream is now RHEL Beta, and All Shall Bow to RHEL. This is right up there with them killing docker-ce because they want you to run OpenShift, which no matter how good it may be, I keep typing without the "f".

  • There are a lot of large organisations who leverage Centos so and like me they will be doing homework over the next few months to find an alternative.
    The real problem is that the cost of RHEL doesn't make sense for a lot of businesses. I might pay $10 a year for a server OS but I don't see thousands of dollars in value for the alternative. One of the places where I used to work at one stage set up a project to migrate to RHEL from Centos, it was rolling along until someone in senior management actually said

  • I read that somewhere!

  • Made the leap from Centos 6 servers over the Centos 8 hoping I would be good until May 2024 on full updates and 2029 for maintenance. Now l look and everything ends next December. Wish they warned us sooner and maybe I could have avoided that. Been happy with Centos since the beginning and deployed RHEL for customers.

    I had been looking at Ubuntu LTS before the switch and probably the way I will go. Too bad what happened Scientific Linux otherwise that would have been the way to go for me. I had actuall

    • by ebvwfbw ( 864834 )

      Hang in there. RedHat sometimes announces these things and then changes their minds. Not sure why they'd support Centos 7 and not 8. Other than 7 has a LOT of drivers that people need. I was shocked that whole entire lines of old hardware was eliminated. Like a PCI-X raid. AMI raid, P series like one that would work in a DL380, Supermicro, Dell. In the case of the DL380 I found an repo site that had DUD files that support it. I think it's elrepo that has a whole bunch of stuff that they stopped supporting w

  • by Maelwryth ( 982896 ) on Wednesday December 09, 2020 @02:04AM (#60810780) Homepage Journal

    That was probably a mistake. I can see the benefit of having a Fedora LTS sitting halfway in between Fedora and RHEL but the point of CentOS was that it was the community version of RHEL, both proof of Red Hats commitment to open source and a gateway to RHEL.

    Those projects that started off using CentOS knew that if it ever came time to sell or scale up then there was a company standing ready to take over support. I have heard a couple of people saying that the CentOS project was a one way thing with people essentially just leaching off RHEL for free. I reckon that is bullshit. RHEL is built on open source software largely built and tested by the community. If anything it was the other way around.

    Anyway, It looks like CentOS is dead as it is no longer an enterprise operating system. I look forward to seeing what Stream is like and what the fork of RHEL looks like as well. Thank you to the CentOS team as well which according to Linux Unplugged was essentially just two people...

  • by whitroth ( 9367 ) <whitroth@@@5-cent...us> on Wednesday December 09, 2020 @12:42PM (#60812238) Homepage

    I've been on the general CentOS mailing list since '09. I'm waiting for one single link to a post warning that this was even being considered, before it was presented as a fait accompli.

    It's certainly *not* the spirit of FOSS.

  • What beats me is how IBM got functional control of CentOS.org, a non-profit, and got it sing and dance to IBM's tune.

    If it can do it with one .org, what about GNU.org?

"I shall expect a chemical cure for psychopathic behavior by 10 A.M. tomorrow, or I'll have your guts for spaghetti." -- a comic panel by Cotham

Working...