philosophical question:
What more that being alive can be to exist?
or vice versa 
Being tied to a hospital bed, aka “Vegetable” is existing…
Walking around, is living…
Tough comparison, but sums it up!
Of course they aren’t, but you continue to argue against points that nobody’s making. Let’s suppose Stream turns out to be (as many of us suspect it will, though Davide doesn’t seem nearly as concerned) completely unsuitable. In that case, the devs have a few options:
- Maintain their own RHEL8-compatible distro
- Use someone else’s RHEL8-compatible distro
- Switch to a completely different distro
The first is, as you observe, not impossible–but it is hard. It’s a lot more work than they’ve been doing, and unless they get significantly more resources, it isn’t going to work well.
The second takes more front-end work than the first, and relies on whoever’s supporting that distro to continue doing so.
The third takes even more front-end work, also relies on whoever’s supporting that distro, but what seems to be the leading candidate is in fact community-supported, with no corporate interest controlling it.
2 Likes
Yes I see, we are starting to talk about different things,
What I am saying is just that RH can’t impeach anyone from creating another Cent-OS style solution, especially from a Legal point of view.
Now NS’s dev/bosses decisions are another matter.
And I might repeat what many have already said, using Stream is going to be a no go because it is a rolling distro. , by essence incompatible with the Server World, and because of the sustainable “stream” of updates … that would make the NS’s team under way too much pressure trying to keep up with the said updates … maybe it is actually the goal of RH ? making “us” “drop it” because if being tired of fighting …
Helloo… we’re talking about business/enteprise grade distro. Even a consumer one like Ubuntu is funded by a company. GPL is enough for avoiding copyright backlashes or license fee gates for accessing technology.
But support, bugfix, security patches and supporting new hardware don’t come fast enough without payed developers.
Even this project, which currently rely on CentOS, without the money, developer, openness of Nethesis for loosing money (less customers due to public access) and gaining information, bug reports, user cases, hints for developments and the exceptional case of huge contributions from a few guys about development, analysis, addons… The project could not be here.
GPL helps to all of this job not to be “wasted” if tomorrow a company decision will cut the rope. NethServer as concept is born in Mitel as eSmith. Maybe during the story of company this distro was not considered that interesting, therefore SMEServer born. And now, with part of the concepts, is now NethServer. Thanks to GPL.
But few years ago SMEServer was (still is, IMHO) in a development freeze for lacking of vision and funding.
Due to current goals, features and uses, CentOS is providing quality of code, patching, security answers that are not easily accessible without paying software. And, as far as i know, this kind of support level is really tough to reproduce without adequate developers and funds.
2 Likes
AlmaLinux wird ein freier, quelloffener, Community-getriebener, 1:1 binärkompatibler Fork von RHEL 8 (und zukünftigen Versionen) sein. Anwendern von CentOS verspricht das Unternehmen, dass Lenix einen unterbrechungsfreien Weg zur Konvertierung bestehender CentOS-Server bieten wird, ohne dass es zu Ausfallzeiten kommt oder irgendetwas neu installiert werden muss. Das Unternehmen behauptet sogar, dass Sie in der Lage sein werden, ganze CentOS-Serverflotten mit einem einzigen Befehl zu portieren, ohne dass eine Neuinstallation oder ein Neustart erforderlich ist.
Translated from german:
AlmaLinux will be a free, open source, community-driven, 1:1 binary-compatible fork of RHEL 8 (and future versions). For CentOS users, the company promises that Lenix will provide a non-disruptive way to convert existing CentOS servers without experiencing downtime or having to reinstall anything. The company even claims that you will be able to port entire CentOS server fleets with a single command, without the need to reinstall or reboot.
Hi
Their text has a title of AlmaLinux, but in the text still refers to Lenix (CloudLinux)…
Not quite polished yet, WIP is still visible…
It’s ready when it’s ready, and that’s when we can start testing stuff…
My 2 cents
Andy
CentOS started life as a fork of RHEL, minus the RH proprietary bits and branding.
CentOS is to be no more. Save for the marketing ploy of keeping the name and tacking on Stream, what was once CentOS is no more. The new CentOS Stream could just as well be called Fedora Stable. Or Cindy. The point being - when the purpose and implementation is changed, keeping the name is fluff.
RIP CentOS. Period. Full stop. That’s all, folks!
So, now what is being proposed as Rocky Linux (I understand the sentiment, but I wish he would have gone with the much more business sounding RockOS) and by others, is basically to restore the practice that WAS (as in, shall be no more) CentOS - that is, to produce RHEL minus the RH proprietary bits and branding.
What’s the panic, then? If, over the next three years, this reincarnation path is endorsed by the C in CentOS, then what exactly is the difference between moving (as was being done) from CentOS 7 to what is currently CentOS 8, and moving from CentOS 7 to (just to give it a name) Rocky Linux 8?
CentOS 8 was RHEL 8 minus the RH bits.
Rocky Linux 8 will be RHEL 8 minus the RH bits
Hmm…
I think, while the CentOS community is rightly angered by the breaking of the LTS for CentOS 8, there’s not so much to worry about… PROVIDED Rocky Linux (or one of the other RHEL forks) succeeds before 2024.
Now, the real worry would be - what if IBM/RH were to decide to shut down both CentOS and RHEL? Then, in that Fedora-CentOS-RHEL path, you would have only Fedora!
Perhaps what is REALLY needed is an ACTUAL community-driven enterprise distro? Is there a Debian enterprise version? I don’t know.
1 Like
This is right, but what would Rocky Linux 9 be? We have a stable Open Source CentOS 8 Base, but there wouldn’t be a stable CentOS 9 Base.
Hi Michael
AFAIK, all “clones” of RedHat will be using RHEL as basis, just as Centos did. They don’t plan to “clone” Centos, but to replicate what Centos did. Clone the RHEL from Source, minus the RHEL specific Bits…
This includes RockyLinux and whatever CloudLinux’s Lenix (temp name) or whatever it’s going to be called… Oracle already does this for RHEL7, RHEL8 and will continue with RHEL9…
If NethServer decides on going such a route, it would be possible and not too much work.
But it would still be dependent on RedHat/IBM corporate whims…
For that reason, I’d also advocate a look at eg Debian or FreeBSD as alternatives, both Distros without Corporate Influence, and both have been longer around than Red Hat!
Note: I’d personally have no problems with either, I use both for 20+ years now. And do note my wording: “I’d also advocate a look at eg Debian or FreeBSD as alternatives…”. Not more, not less!
My 2 cents
Andy
I was thinking about the future of RHEL, will it be open source anymore, or will they change to a close source? It’s IBM now.
If it will not be open source at the future all folks can’t use the new version anymore.
Perhaps they use CentOS Stream as an open source platform, change some important parts of the code and make it close.
Of course these are only misgivings from me.
1 Like
1 Like
The very large majority of it, at least, has to be open-source–RH didn’t write most of that code; they used other GPL’d code, packaged it, and released it. That was open-source before they got it, and it will continue as open-source. Their own stuff, well…
Even if it remains nominally Open Source, there are always “dirty” tricks they could use to discourage repo usage…
- A 1-2 day Repo Maintenence downtime just after RHEL relaease updates…
- Maybe a few “Typos” thrown in randomly so it doesn’t fully compile.
- Or a random BLOB thrown in (Closed source Binary, there by error, of course…)
As I said, I don’t trust them anymore.
Big Blue ain’t clean, they’ve been known to play dirty, and claim innocence…
(December 2020 - “Some user” put up the 2029 LTS Support date… Centos has always mirrored RHELs LTS dates…)
My 2 cents
Andy
Forget CentOS. It’s dead (at least, as you’re thinking of it.)
As I see it (with RL meaning Rocky Linux or whatever distro assumes the “built from RHEL, community version” mantle):
RHEL 7 => CentOS 7
RHEL 8 => CentOS 8 (this is the end of the CentOS line)
RHEL 8 => RL 8 (ostensibly, an equivalent to CentOS 8)
RHEL 9 => RL 9
etc
If the RL community, over the next three years, matures, then it simply resumes the work that the now-defunct CentOS used to do - that is, take RHEL and produce RL.
RHEL is mostly open source, so it will be hard for IBM/RH to “kill” it. They CAN elect to stop maintaining it, though. There are a plethora of OS alternatives out there and they could decide to switch to Debian or SLES or Slackware or to assemble an all new distro from scratch or whatever.
However, having worked in multiple supercomputing environments that deploy literally thousands of RHEL and/or CentOS nodes, I can tell you this much - dropping RHEL is not something that can be done with the flip of switch, not even by IBM. (I can also tell you, the sites relying on CentOS are scrambling for a new “not Stream and can’t afford RHEL” option.) IBM CAN flip a switch and announce they WILL be dropping RHEL and immediately refocusing development to “something else”. That said, practically and contractually the result is not an event but a long, slow process to migrate the customer base from RHEL to that “something else”. (Have you considered why IBM didn’t shorten the CentOS 7 EOL to eoy 2021?)
In the meantime, RHEL would go into a stagnant EOL state, but the core of the code would still exist in open source. The question then would be, who would have the resources to take it up. Probably not the community. Recompiling and releasing RHEL sans proprietary bits is one thing. Driving the development of RHEL itself is another, much different, much larger task, even if compared to what it was back in RHEL 1 days. So maybe HP would pick it up? Dell? Who? Somebody big, that’s for sure.
Anyway, I find the whole “can’t trust IBM/RH” thing to be reactionary, driven by the decision to drop CentOS for Stream (not the same things, as I’ve said.) It’s not so much of an “out of the blue! oh no! panic now!” change, but a three year warning to replace CentOS. Either replace it with another distro, or (more rationally) use the remaining CentOS 7 LTS period to replace CentOS 7 it with another RHEL 8-derived build. The latter being essentially a new “community” continuing that work that the now-dead CentOS used to do - taking RHEL minus proprietary bits and making a community release.
Furthermore, anyone paying attention has long known there was a piece missing in this process:
Fedora => RHEL => CentOS
How does the chaos that is Fedora become a stable RHEL? By this:
Fedora => [ IBM/RH resources tame Fedora => RHEL ] => CentOS
Except, there weren’t resources in that taming process to also (formally) keep up with newer developments, leading to all the extended repositories for RHEL and CentOS.
By pulling the resources formerly spent as CentOS (which, by this time, the C stands for Corporate more than Community) and using them for Stream, IBM/RH are looking to resolve the shortfall.
Fedora => Stream => RHEL => “community”
(Note that I say Stream and not CentOS Stream. They are not the same things and re-using the CentOS name is something akin to car company’s that attempt to revive model names from a half-century ago. It’s familiar. It evokes emotions. But 2021 Fords and 1961 Fords have little in common. CentOS Stream is not CentOS! CentOS as we knew it is dead!)
Anyway, the end result should be a better RHEL, and that should yield a better community version. It’s just that IBM/RH now says “the community version is back where it belongs, as a community responsibility - and you have three years to get there.”
It’s not like the process “How to get from RHEL to community version” is dark magic. If you can build a house, you can build another house.
There is magic involved, though, that being: Can the community come together to create a community that can create a community release? (Just look at SME Server to see a dysfunctional community.) Or are the naysayers claiming the community is now dependent on the C=Corporate entity to do their work for them?
We should forget the calls to start over from Debian or Ubuntu or Slackware blah blah blah. What is needed is what Rocky Linux (RockOS, please!) and others are doing.
Without all the panic and hysteria and drama.
Hi Scott, to begin with consider myself pretty distro-agnostic… do not really care that much.
Tend to disagree with your optimism an other solid community can be formed that easily. Especially something which involves rebuilding a dirstro without adding value to it.
Problem I see lack of involvement in the long run because community members cannot have a lot of input. I mean at the end it’s not that exiting to rebuild without development.
IHMO this was one of the challenges CentOS faced a few years back which makes me a bit more pessimistic.
Without referring to technical issues regarding Debian, see upsides in the structure of the project.
Debian is not meant to fork and rebuild; rather to build-up-on. Lots of smaller Debian based dirstro’s just mirror (or even directly point to) Debian repositories.
From this perspective it may be a good fit for Nethserver too 
2 Likes
That’s a world of difference from what a stable enterprise version would require. It would be impractical in a actual large environment to keep up with that model. A large enterprise wants a core that changes once, maybe twice, annually. If they need something special, that is why they have devops. When you have thousands of boxes, multiple OSes, users running dozens of solvers, attaching to multiple storage solutions, running over mixed fiber and ethernet connections, the headaches of global WAN transfers of TB/PB data sets, etc etc etc, the last thing you want is a core OS that’s always in flux. (For example, at one site the decision to migrate from RHEL 4 to 5 took so long, so much testing, so much mitigation planning, that by the time they made the change, it was to the penultimate release of 5.)
While the goals of a Rocky Linux are admirable, I think I conveyed that it’s not simple. The community that existed 10 or 20 years ago does not exist now. It will be a greater challenge to find qualified free labor than it used to be. (Due to level and type of interest, not financials.) But, if (big IF) it can be done, then what is the current CentOS userbase should end up no worse off. But, as you say, there is a reason IBM took over CentOS (and before that, Red Hat, and before that it was already a major player in Linux - so in a way, you could conclude that IBM has been a responsible steward for many years already.)
Me, I’ll hide and watch, occasionally chime in. However, in my retirement, I content myself with Mint and a very few Windows apps under Wine. And those will eventually disappear. Linux app options are still not quite on par with Windows, but they get closer all the time. And for those times when I have to live in a corporate world, if they want Windows then I give them Windows (apps) 
And where I used to keep hundreds and then dozens of SME Server installations going, I now have only one Koozali and one Nethserver. Eventually, that will become two of one type. You can take the fact that I’m in this forum as an indication of which one will become two
1 Like
Comic relief time…
Do you think that Rocky Linux is safe from Mr Sly Tradermark Claims?
Yes.
Unless Rocky Linux starts making boxer movies starring a guy named Rocky.
Or Stallone starts up a Rocky Balboa linux distro.
In which case, RockOS is better anyway.
That is, until Dwayne Johnson gets involved
some news on RH
4 Likes