Rocky Linux claims to have found 'path forward' from CentOS source purge | Hacker News Reader

Rocky Linux claims to have found 'path forward' from CentOS source purge

theregister.com

14 pointsmmrezaie21 comments

Honestly, I’m tired of this. We were going to use Rockylinux as a base for the new version of our research cluster at a University used by researchers from all over the world. We are a tightly funded shop with money allocated as parts of research grants. We were running off of CentOS because a few tools that researchers use were built “for” RHEL by the commercial vendors. Since Redhat now considers users of CentOS “freeloaders” and “leeches”, we have to move on to a better distribution that doesn’t have greedy commercial interests and works with the open source community in the spirit of open source including redistribution freedom. To be fair, even though we were using CentOS at no cost, we have reported a fair share of bugs to RH that I hope to believe benefited the overall community.

We have settled on making the effort to move to Debian for the compute nodes.

One thing we could not break out of is FreeIPA :-( We are not brave enough to run FreeIPA on Debian or find a completely new alternative in Debian that wouldn’t break us while we are already changing so much. So, we are going to pay for RHEL for one server to run IDM.

I will tell the researchers that come to us needing RHEL compatibility to pay for RH license out of their research grants, or spend the time and work with me to make their stuff work on Debian(I can already see what choice they will make).

I hope the fine folks and management gurus at IBM/RH rejoice with their short term profit at the cost of long term loss of faith and ruin.

> We have settled on making the effort to move to Debian for the compute nodes.

I’d donate to Rocky and stick with it, if it were my project.

I would happily donate to Rocky if I thought it would really sustain it. But, honestly, they are too dependent on RH and finding ways to subvert their restrictions (and considering RH is pretty boldly clear that they dont want to give away the sources) makes it a whack-a-mole game that is not sustainable in my opinion. Been burnt with CentOS 8 once, and I'm not willing to go through it again. Atleast I know Debian wont go this way.

> But, honestly, they are too dependent on RH …

With enough support Rocky may grow by itself, without RH code. Like what is happening with OpenSearch and Elastic, and OPNsense and pfSense. And my gut feeling is that will happen.

> We are not brave enough to run FreeIPA on Debian

are these redhat projects really so crappy that it is something to be scared about to attempt to simply run them on another linux distribution?

Not that freeIPA is scrappy. on the contrary, it has been very good. Properly documented, and does the job well. The issue is finding another alternative for it as cohesive as freeipa in the Debian ecosyystem. FreeIPA is tightly integrated into RH ecosystem that it is prone to trouble when introducing changes to the environment it runs at. Can I make it run on Debian - Sure? But, will I trust such a setup for the critical element of our stack? NO.

You can get up to 16 RHEL licenses for free.

You could always try talking to Red Hat Sales and dump a list of your submitted bugs at the same time. Say you need a really good deal or you're moving to Debian. The worst they can do is say no.

[Author of the original article here]

> So, we are going to pay for RHEL for one server to run IDM.

With a free developer account you can run up to 16 copies of RHEL, with up to 8 cores each, in production, without paying.

Have you tried reaching out to them about obtaining an Academic Research subscription? I believe they offer those. Can’t speak as to how much they are though.

Like for the vast majority of use cases, unless you really need the bugs present in RHEL, you should be even better served by CentOS Stream than you were on old CentOS.

It's just a few patches ahead RHEL, with packaged builds using the exact same Infra and QA process.

It's not RHEL N+1 beta as some disingenuous people have argued

> Setting aside the large numbers of angry people who don't really understand how open source licenses work, our impression is that the core issue here is that there are an awful lot of people who feel that simply because this is Linux, they have some kind of right to get it for free. Unfortunately, they don't. That is not what the "free" in Free Software means, and it never was.

Something about this irks me. I realize there are some differences between FOSS, FLOSS, and open source. And I realize sometimes there’s some ambiguity there. But I’m still not sure I agree.

Would it be possible to:

1. create a free dev account

2. get the sources

3. merge changes into rocky repo

4. have the account terminated by the Hat

5. goto 1?

Subversive, but mostly in compliance with GPL and Red Hat EULA... Ish. Kinda.

But I probably made an error here. Who can point it out?

4. Have the user terminated by IBM lawyers.

without sufficient caffeine or context, this line reads like a cyberpunk plot.

William Gibson! Lawyers are also execution branch in any corporate vassalage area. They carry the Licence, pocket Bible and FN Five-seven.

Possible sure. Also illegal and would get the company shut down.

I'm not sure if this is illegal.

The developer (john doe) honours the EULA up to the point where the sources are on his computer and cancels hist account (or has RedHat do that). From there, it's just GPL code that can be re-destributed freely (save for RedHat trademarks and licensed assets).

The company (rocky, alma, springdale, oracle) is in it's full right to accept any changes and merge them into their GPL code base.

Sure the ethics are questionable, but I suppose the IBM lawyers would have a hard time prohibiting this in the EULA while simultaneously honouring GPL and welcoming contributions from anonymous developers.

But I suppose I'm just too dumb to see that my cunning plan isn't very cunning at all. :-)

> IBM lawyers would have a hard time

Never underestimate your enemy. (c) Sun Tzu, The Art of War

How does RedHat know said user is the one who distributed the changes? If the changes are given to Team Rocky on CD via USPS, there would be know way to know which RedHat user is the one giving changes to Rocky.

Red Hat wouldn't even care to identify the user: Rocky Linux, the Rocky Enterprise Linux Foundation, CIQ, etc would be brought to trial for accepting those sources.

Related: GPLv2, Red Hat, and You

https://news.ycombinator.com/item?id=36517045

Reply on news.ycombinator.com