On Dec 6, 2022, at 19:32, Jeffrey Walton <noloader(a)gmail.com> wrote:
And dnf-system-upgrade every 6 months is a small price to pay for
Fedora (
https://docs.fedoraproject.org/en-US/quick-docs/dnf-system-upgrade/).
You get Red Hat processes and stability with modern software. Its a
win-win.
I suppose it’s a win if you’re the one being paid to fix broken systems. :/
I am fine with your suggestion for the modern containerized platforms or cloud service,
using agile methodology, and constantly spinning out new versions. Release early and
release often. Not everyone has that workflow.
It all boils down to risk assessment.
Greg K-H has a good point and what he describes is one of the risks with using enterprise
Linux. The kernel in RHEL doesn’t get all fixes in the upstream. On one hand, one of those
bug fixes might end up being a security hole down the road. But on the other hand, that is
a risk that Red Hat takes and is responsible to fix. This is why you pay for RHEL.
In your model, every bug, breaking api change and orphaned package is your problem and
it’s up to you to fix whatever code you use to run your service.
Some companies would rather pay another company to take that risk than shoulder it
themselves.
I’m glad that Fedora is the basis of future RHEL. I will be able to tell what will work
and will not in new releases. I’ll get valuable experience in the new software and maybe
even get a hand in guiding the path to future releases. I will have the same basic OS on
my laptop as what I’m running in production. I am also happy to know that people use
Fedora Server in production as an alternative to other distros.
I just don’t agree with your assessment that enterprise Linux has no value. Not everyone
has the headcount or technical ability to use Fedora for their infrastructure. And I think
it is disingenuous to suggest otherwise.
--
Jonathan Billings