There are plenty of cases where infrastructure can be split into "critical" and "nice to have".

In HPC, it is reasonably common for people to run RHEL on the scheduler, database and login nodes (totalling maybe a dozen servers) but then use CentOS (now Rocky) for the compute nodes, which might be hundreds of nodes. Those customers simply can't justify the cost of RHEL everywhere, even with more recent changes which make it cheaper than it has been historically in this use-case.

To me it sounds like the pricing model is not optimal. I guess in these situations, they could pay RH "a bit more" in exchange for a legal RHEL license covering those extra nodes. But if there was no such offer then it sounds like a lost opportunity.

(but I'm a complete outsider, I just share my views and I do acknowledge I may be missing something, so feel free to elaborate further)

I think you are right - the problem is that "a bit more" really ought to be a small uplift on the needed licenses, rather than the current inadequate discount on compute nodes (along with hoop-jumping to convince sales people you aren't trying to scam them).

Boeing, up to at least 2016. A large amount of RHEL, and a screamingly HUGE amount of CentOS.

Sometimes support - for the OS or from applications - is actually useful. I worked at a shop once that ran CentOS for most things but RHEL for the database servers, because the vendor only officially supported that.

Additionally as their needs grows and the free version doesn't meet their needs out of the box and the non-free version does they'll be more apt to buy into it given their previous experience with their products.

So the reason for not using RHEL everywhere is the pricing model? I wonder whether it'd be possible for RH to optimize it so that choosing RHEL 'd still be the best option for them now that CentOS doesn't provide the same level of stability.

> I wonder whether it'd be possible for RH to optimize it so that choosing RHEL 'd still be the best option for them now that CentOS doesn't provide the same level of stability.

What you may not be aware of, or have forgotten, is the free versions provided by other companies that provide package compatibility and similar support (as in updates and security fixes) that CentOS did.

Oracle Linux, Alma Linux, Rocky Linux, and Amazon Linux. Quite possibly even more.

Red Hat will have a hard time undercutting them just by reducing their pricing, which is why they're using tactics like a free number of subscriptions/licenses for small organisations and developers and trying to cut off access to the source code.

In our use-case, needing to handle licensing in any form for additional nodes we scale up in non-production would break our processes. We run RHEL on some core infrastructure with equivalents in our pre-production but CentOS everywhere else to give a ‘close enough but free’ infinitely scalable environment to our devs. As the person in charge of the existing relationship with RHEL, there is no amount of discounting or special deals that would make the extra complexity of registration worth it. We need an unencumbered freely deployable OS for our non-production, and for us the result of this change is likely to be a re-platforming to something like Amazon Linux and a seperate support contract or line item on our AWS relationship, rather than giving RedHat any more money and time.

For us, this means we stop doing business with RedHat.

It is usually some combination of needing paid support if updates break things, and also generosity. Trust and generosity define the choice of paid support organization.