AlmaLinux No Longer Aims For 1:1 Compatibility With RHEL

[Duve]

CentOS Stream is compatible with RHEL.

In fact, Red Hat encourages that developers use CentOS Stream for development because if it works on CentOS Stream it will work on RHEL.

I think that is the thing that got missed in the all of the mess.... RHEL is dependant on CentOS Stream, the sources should hit Stream before anywhere else (the only exception to this is embargoed bug updates, but that is used sparingly for something serious). You want RHEL sources, they will suggest that you look at CentOS Stream for them... they will be there.

[pWe00Iri3e7Z9lHOX2Qx]

So exactly what CentOS Stream is doing, but instead Stream actually contributes to upstream instead of being leeches feeding of RedHat and giving nothing in return.

What?

Fedora -> CentOS Stream -> RHEL.

By "upstream" do you mean various upstream projects or do you mean upstream to RHEL? CentOS Stream is upstream RHEL. I'm not sure what you mean by "so exactly what CentOS Stream is doing". None of the rebuilds can do exactly what CentOS Stream is doing because its' the only one that is RHEL's upstream.

[pWe00Iri3e7Z9lHOX2Qx]

So I'm here, I hate the technical direction that Canonical is going. I hate the commercial direction that RHEL is going. I guess I'm going to move to SUSE... If they fuck it up I might just push to move everything to fucking Windows server, at least Microsoft is consistant and won't rugpull support without warning.

Debian? That's the most "stable" offering that isn't beholden to a corporate overlord. I do hope that a few big companies start donating more to support Debian LTS and Extended LTS efforts. It would be nice for the community to at least ensure that all important packages get 5 years of security backports.

And for people who say that Oracle/Alma/Rocky are just leechers, turns out Red Hat is also "leeching" from the entire Linux ecosystem. If they want to have an OS that can't be cloned I guess they should just make a closed source one and stop fucking around.

RHEL development is actually more open / transparent after all of this through CentOS Stream.

And I know that Red Hat is one of the largest, if not the largest individual contributor to that ecosystem. But turns out their "leechers" also are, Oracle in particular is a huge contributor.

One way or another, I think this is a pretty short sighted move. Free clones are a great way to onboard new users, bringing more certification money, more mindshare and more sysadmins pushing to use the thing they already know how to use in their companies. For companies its a perfect development target, without having to mess with licensing their potentially short-lived dev machines. All in all I think long term this will erode RHEL as the standard "professional" Linux.

Not when the clones create paid support organizations that undercut Red Hats' rates. Red Hat employs thousands of people and spends many millions of dollars per quarter to build RHEL. Someone else creating some build automation and offering cheaper support is the part that is BS to me.

[andyprough]

Red Hat employs thousands of people and spends many millions of dollars per quarter to build RHEL.

Let's not get carried away - they aren't even a separate entity any longer, they're just a small line item on IBM's income and loss statements. All of those people work for IBM, and those millions of dollars are nothing more than a rounding error on IBM's financial reports.

[Quackdoc]

I think that is the thing that got missed in the all of the mess.... RHEL is dependant on CentOS Stream, the sources should hit Stream before anywhere else (the only exception to this is embargoed bug updates, but that is used sparingly for something serious). You want RHEL sources, they will suggest that you look at CentOS Stream for them... they will be there.

it has been stated multiple times by multiple sources the centOS stream will not have the same degree of support that RHEL and RHEL clones have, if you use centOS stream, this means I think 4-5 years of support that packages will be missing out on. this is not insignificant.

[cynic]

Red Hat employs thousands of people and spends many millions of dollars per quarter to build RHEL. Someone else creating some build automation and offering cheaper support is the part that is BS to me.

RH is spending millions of dollars just to contribute to software worth billions.

Let's not forget that altough the RH contribution to the whole echosystem is very precious (and probably vital), they're still contributing to software written by other beside them.

It's not like RH wrote everything they're selling support to.

[Knghtbrd]

Good.

Now Red Hat needs to finish turning the screws by taking a page out of the NFL's playbook.

[snip anti-Linux trolling]

Red Hat should do the same thing and say that going forward no one is allowed to claim they are or are not binary or bug-for-bug compatible with RHEL, Red Hat or any reference to Red Hat's software.

If these people want to build a Linux based business then let them do it on their own, no more free ride.

I think you don't understand that a statement of compatibility does not violate trademark law. Much to IBM's chagrin back in the 1980s when they tried to bury Compaq any way they could. Too bad, we won't see people saying their stuff is "compatible with the douchiest corporate Linux distribution."

IBM Red Hat is now profanity as far as I'm concerned. Any client considering it, I will simply tell them: They promise ten years of support and then kill it a year later. You'd be better off with SUSE if you need enterprise-level support. If you don't, Debian—Debian's support model is not contingent on a profit motive, and it never will be.

[pWe00Iri3e7Z9lHOX2Qx]

Let's not get carried away - they aren't even a separate entity any longer, they're just a small line item on IBM's income and loss statements. All of those people work for IBM, and those millions of dollars are nothing more than a rounding error on IBM's financial reports.

Well, Red Hat certainly isn't the majority of IBM's revenue. But software is IBM's largest financial reporting segment. They made $5.9B last quarter in that segment ($14.3B total). Red Hat contributes over $1B of that (I'm too tired right now to figure out a closer approximation by looking at all the RH quarterly growth numbers in IBM's financial statements since the acquisition). So probably ~20+% of the revenue in their largest segment is from Red Hat. Certainly not a "small line item". IBM was always going to try to extract as much value out of Red Hat as they could, otherwise they wouldn't have splashed out $34B to buy them. Especially with the global economy looking shaky, I bet they are worried now more than ever about potential customers opting to just use Alma / Oracle / Rocky and cut out a large expense to Red Hat to help cut costs and weather the storm. Interestingly, and possibly not coincidentally, all of this is taking place after IBM's projected growth for Red Hat this fiscal year softened a bit last quarter.

[Ananace]

What annoys me with this whole debacle is how badly RedHat are shooting themselves in the foot, especially in regards to smaller companies and developers working on third-party software.
Without a bug-compatible free version available, there's nothing - and no, CI-incompatible license-restricted "free" dev licenses don't help - to onboard them onto the fully supported and paid version.

We're probably going to have to move a bunch of our RHEL servers at work to something like SLES instead, since they end up in a fun intersection between running community developed software - which already leans towards abandoning RHEL due to the changes looking like they'll be breaking their support possibilities, while at the same time also requiring our enterprise-grade backup software and its precompiled kernel modules to work.

[pWe00Iri3e7Z9lHOX2Qx]

IBM Red Hat is now profanity as far as I'm concerned. Any client considering it, I will simply tell them: They promise ten years of support and then kill it a year later. You'd be better off with SUSE if you need enterprise-level support. If you don't, Debian—Debian's support model is not contingent on a profit motive, and it never will be.

I'd love for both SUSE and Debian to grow market share in big ways as a result of this, and for waaaaaaay more companies to help fund Debian LTS and Extended LTS efforts. Competition is good for all of us. Debian has a fairly unique position in terms of the size of the project and a "stable" focus that isn't downstream of a for profit corporation. The cynic in me doesn't see the 800 pound gorilla in this space losing meaningful mass though.