• 1 Post
  • 98 Comments
Joined 11 months ago
cake
Cake day: July 22nd, 2023

help-circle
  • All good points and I appreciate and enjoy the discussion.

    In my view, CentOS Stream is already a lot more of a “community” distro than the original CentOS was.

    This is possibly a semantic point, but for me, a community distro is owned and operated by the community without any corporate control. All the points yonu make are true and valid, but ultimately, Centos is owned by a very large corporate entity that could stop it whenever they want to and nobody else can do anything about that.

    Some examples of community owned distros are Debian, as well as Rocky and Alma Linux. Both of the latter have commercial arms, but are are fully independent legal entities owned by the distro. Rocky is owned by Rocky. This point was particularly important because that’s what the community thought Centos /was/, but it turned out that Redhat owned Centos. I don’t think either of the new distros would have been as trusted if the same thing that happened to Centos - a corporate entity ultimately deciding what happens - could have happened to them. When abandoning a sinking ship, it’s prudent to check you’re not boarding another with a big hole in it.

    I did happen to look follow Rocky’s path closely, and our company chose it to migrate our doomed Centos8 machines to, because our developers didn’t have time to rebuild everything for Debian in that particular window. That decision was largely based on that legal standpoint because we didn’t want Centos repeating on us. It was also reassuring that Rocky was founded by Greg Kurtzer, who founded Centos and had that project effectively stolen from him, and he least of anyone wanted the same thing happening. (BTW, Rocky was named after the other co-founder of Centos, who has since died - a nice gesture)

    My cynicism of Redhat and their motives are real and may be misplaced, but I don’t think they’re done piddling in the EL swimming pool just yet. I adored the company once and had nothing but respect for what they achieved. But that was then and this is now.




  • RH had taken over the Centos project and Board by that time. You’re right that Centos was already circling the drain in terms of resources (I remember waiting many weeks for point releases), but the way they did this was brutal and poorly communicated.

    And remember those downstream ‘rebuilds’ only appeared to fill the vacuum caused by Centos disappearing. That they’re both doing very well does make you question whether Centos could have been sustained in its traditional form. (As opposed to Stream, which is only of benefit to Redhat and those in its testing cycle)



  • Rocky and Alma are RHEL alternatives and are absolutely aimed at the enterprise. Fedora merging with either of these projects would be super surprising indeed. It would make no sense whatsoever.

    It would make a lot of sense to Rocky and Alma though - as if RHEL went there would be a huge vacuum and their models would be impossible. I know there was a lot of talk in both companies when the source was paywalled about building directly from Fedora’s sources (Alma may actually be doing that, I’m not sure). Both R & A have significant user bases, both Enterprise and Community, and there would be considerable desire to keep the wheels turning. Some sort of collaboration (or just downstreaming directly from Fedora) feels inevitable as a choice if that were to happen.

    The “community” enterprise option from Red Hat is not Fedora, it is CentOS Stream.

    Centos Stream is not community by the way - it’s entirely owned and run by Redhat (AIUI, They took over the name from its community origins and replaced the board with its own employees. The vote to end traditional Centos (which was community run) was given as an ultimatum with a great deal of bad feeling) Stream’s purpose is as an upstream staging area for new releases of RHEL. Redhat state it’s not suitable for production use, so it’s of no real benefit to anyone that isn’t part of that test cycle. (In some defence of Redhat here, Centos was struggling with low resources for a long time before this and point releases often took weeks or even months to appear behind RHEL)

    RHEL don’t publish sales figures afaik, so they’re the only ones who could say whether they’re up or down. I’m just one guy who’s worked in a mostly EL based world which has been negatively affected by these decisions, so I’m keeping half an eye. I could be completely wrong, but the facts we do know aren’t healthy for someone wanting to enter into a business relationship with them, which is what a corporate company does when choosing a supported distro like RHEL.

    And yes, I am quite cynical - you’re right to point that out. I also hope I’m wrong. If I’m not, I have a lot of confidence that the world will continue with or without RHEL, but yes, it would be a big loss to the FOSS contributions they have made and continue to make - as well as a lot of good people losing their jobs.




  • You say that like it was a small thing, but small things don’t create such bad feeling, cause most of the Centos volunteer team to resign, create off two entirely new distributions (Rocky and Alma). The subsequent paywalling of RHEL sourcecode and its accompanying spiteful communications make it clear where Redhat’s focus is - or, rather, isn’t. People judge companies by what they say and do, and I and many others are deeply concerned for the future of RHEL after the IBM takeover and are moving away from it.

    I think there is a lot of nostalgia about the great work that Redhat did (and still does, at a smaller scale) and are overlooking what it’s become but RHEL as a business product is not the force it once was. I think it’s entirely possible that Redhat/IBM will simply pull the plug on RHEL and the entire EL universe will need some serious remapping if its to survive.

    (Was a Centos user, still maintain 180 EL servers, am quite aware of the FUD, much of which originated and still does in the other direction from Redhat and its employees. The Centos 8 announcement came just after I’d manually migrated 60 vms to it, which then needed migrating again to another distro - so this did cause us some significant work and cost.)


  • I would not encourage anyone to join the EL universe as I don’t consider it as stable as others.

    TLDR; Redhat’s being absorbed into IBM and they don’t care about RHEL. RHEL (in my view) is dying a slow death. Without RHEL, there is no Fedora or Centos Stream. There’d also be no Rocky or Alma, as things currently stand.

    (Although if that happened, I’d not be surprised if the users of Fedora merged with Rocky and Alma in some form of new and fully independent distro - we’ve already seen how well such disasters can be worked around)

    Longer reasoning: Redhat, in my view, have made some unpredictable and frankly terrible decisions over the past few years with RHEL which have caused a great deal of concern in the business sector about its stability as a product. (Prematurely ending Centos 8 six years early, paywalling the source code, and more recent anti-rebuilder steps. They also treated the community team working for Centos appallingly throughout these leading to many resignations.) Further more, these were communicated without warning or consultation and have sometimes come across as petty and spiteful, rather than as professional business decisions.

    IBM bought Redhat shortly before this happened, mostly for its cloud services. It seems from the outside that RHEL is being squeezed. There have been two major rounds of layoffs. In all, this paints a picture of a company that is in decline and we’ve seen a reduction in contributions to the excellent work done by Redhat in the foss world. IBM have a long history of buying and absorbing companies - I don’t see why Redhat would be any different and RHEL doesn’t make enough money.

    Our company is moving away from EL and I know of several others who are doing so. We’re all choosing Debian.




  • Linux: 1995, Sco (At work), then got a copy of Slackware on a Cover-CD around 2000. Shortly after found Debian and have been using that at home exclusively for over two decades, now onto desktops and laptops as well as a couple of home servers. (I use EL distros, Ubuntu and OpenSuse at work nowadays)

    Longer history: 1981: ZX81. 1985, Dragon 32. 1988 Amstrad CPC. 1991 an XT. 1992 A 386 sx25 with 1mb ram, and so on.




  • IRC’s not as popular as in its heyday, and while once it was the main choice for multi-playing gaming chat (Quakenet et al), that’s largely gone elsewhere, but it’s still very good for certain technical channels.

    IRC has also proved to be remarkably resistent to commercialisation, mostly due to the users. Even when one of the biggest networks, Freenode, got taken over by a drug addled mentalist Reference who started insisting all all kinds of strange things, the users just upped sticks and created a new network. A bit of fuss, but the important stuff stayed the same and it’s continued much as before as a new network, Librenet.


  • Others have answered your question - but it may be worth pointing out the obvious - backups. Annoyances such as you describe are much less of a stress if you know you’re protected - not just against accidental erasure, but malicious damage and technical failure.

    Some people think it’s a lot of bother to do backups, but it is very easily automated with any of the very good free tools around (backup-manager, someone’s mentioned timeshift, and about a million others). A little time spent planning decent backups now will pay you back one day in spades, it’s a genuine investment of time. And once set up, with some basic monitoring to ensure they’re working and the odd manual check once in a blue moon, you’ll never be in this position again. Linux comes ahead here in that the performance impact from automated backups can be prioritised not to impact your main usage, if the machine isn’t on all the time.