On Tue, Dec 15, 2020 at 12:30 PM Phelps, Matthew <mphelps@cfa.harvard.edu> wrote:

On Tue, Dec 15, 2020 at 1:00 PM Mike McGrath <mmcgrath@redhat.com> wrote:

I'd also just add that while I find Johnny's characterization of what happened accurate, Ljubomir took a couple of leaps that I don't think existed.  Red Hat decided not to continue paying actual money for what was actively harming us and no longer providing the value that it once did.  No one, not even the board, could force Red Hat to continue paying for this project which was just not working for us. 

Thanks for admitting that the reason Red Hat did this was financial. This BS about it being "a better way for Community input into RHEL" is just that, BS.

Ah, actually I didn't do that.  RHEL is and has been doing fine.  Don't confuse "value" with revenue.  CentOS Linux no longer served any purpose at Red Hat and I'll flip it back around as I did in the previous email.

Why should Red Hat, or any company, continue to pay for something that isn't working out?

If RedHat needed to justify or clarify the investment it was making in CentOS -- as a reminder: *after* taking the independent project under its wing and letting others snuff themselves out as superfluous -- then the professional thing to do would have been to go to the larger community about it.

Present options, such as the rebuild being spun back out. Or discuss mechanisms for increasing CentOS->RHEL conversions. Or solicit direct funding options to get free-riders to contribute directly to CentOS Project expenses while still keeping a firewall in place.

 

Can we stop with the charade that this is supposed to be a good thing for the CentOS community? It's not. It was never intended to be. It's a punishment for us getting "free Red Hat" all these years.

I don't think anyone's said that.  This is a massive change and disruption for the existing CentOS community.  90% of the community (by our estimates) will be able to stay on CentOS 7 until 2024 just as they expected.  We made sure the 10% on CentOS Linux 8 didn't continue to grow (thus trying to minimize impact).  We aren't punishing anyone and the fact that two other clones have already popped up is a testament to that.

No, the fact that two other clones have popped up is a testament to OSS communities' ability to cope with events. The unexpected churn from having the distro pulled out is absolutely a punishment because it creates a great deal of work for all of us to return to the operational status quo with no real benefit.

Direct question: If the CentOS Project (via the Board), secures funding for expenses relating to the rebuild, does it get to continue CentOS Linux?

 

Well, you all see the reaction this has garnered around the world, and it's all negative except for the Red Hat employees trying to convince us it's a good thing. Nice try.

Actually, things took an interesting turn around Thursday.  Once people understood what we actually announced much of the press has been very positive, and now that the shock has worn off, we're seeing quite a lot of support.

"What you actually announced" was that CentOS+CR was going to be used internally for testing against future minor releases, and that any ideal of a binary-compatible rebuild was going away.

I'm sure there's support from internal RH teams that for some reason didn't have access to internal RHEL minor release betas. I can't imagine who else this benefits in any way shape or form (except Oracle, Amazon, and promoters of Debian-derived distributions).

"RedHat EL Stream" is a useful thing, and whether that's a Preview (post-QA), a Beta (intra-QA), or Rawhide (pre-QA), there's a place for it as an official way to provide feedback. But it's entirely orthogonal from the "North American Enterprise Linux Vendor" rebuild project.

We all know differently. And we are all now making influential choices that will hurt Red Hat. 

I don't mean to sound cold here but if you really want to talk about the business side of this....  If you don't have a budget and don't end up finding a home in our coming low-cost or free offerings (Fedora, CentOS Stream, UBI, or RHEL for developers, CI, Open Source, edu, mom/pop shops, etc).  Then what choices are you talking about?

Anyone even tangentally associated with OSS is aware of the free rider problem. As written above, there were plenty of ways to approach this without forcing CentOS Linux to be canned prematurely in the middle of a major release support cycle, immediately after EL6 had gone EOL, and after RedHat's entry had seemingly removed the need for other projects to continue operations. Frankly, the lack of goodwill demonstrated here places both "free" RHEL and UBI into suspect categories.

-jc