[CentOS-devel] https://blog.centos.org/2020/12/future-is-centos-stream/

Tue Dec 8 19:20:16 UTC 2020
Japheth Cleaver <cleaver at terabithia.org>

On 12/8/2020 10:20 AM, Rich Bowen wrote:
> On 12/8/20 10:15 AM, Dan Seguin wrote:
>> Yeah, the "free" help just grinded to a halt.
>>
>> Rich, you being the messenger, send it upwards to (IBM)?
>>
>> Your missive is going to make a lot of people go elsewhere.
>
>
> FWIW. IBM has not been part of this conversation. Red Hat is an 
> independent entity. (Yes, I'm sure that resulted in eyerolls, but it 
> also happens to be true.)
>
> And, yes, not only is feedback here being relayed to management at Red 
> Hat, most of them are also here reading it the same time that I am.


The "I'm not mad, I'm just disappointed" tagline might be best here. 
It's clear what the broader EL community had wanted for quite some time 
was a way to /actively/ contribute back into the development of 
Enterprise Linux. CentOS, SL, and others were downstream, and rightfully 
so for their design goals. CentOS Stream (this new incarnation, at 
least) seems to be intended as a way to allow direct feedback straight 
from the community without having to battle through the intermediary of 
the Fedora Project, which ... has its priorities elsewhere.

 From that perspective, this announcement is a positive step. However 
coupling it with the early EOL of CentOS Linux 8 is extremely troubling. 
The build mechanisms are in place, the expectations were in place. IBM's 
purchase of RedHat was supposed to mean the ability to do things in a 
/more professional/ manner rather than a less professional one -- and 
that's exactly what this feels like: something unbecoming of what we'd 
expect from the leadership of either company.

Many folks out there delayed their transition away from EL6 due to the 
instability of some of the changes in EL7 (not mentioning any 
system*cough*d reasons) and were just now performing forklifts up to EL8 
with that support expiring. We were just wrapping our heads around the 
operational headaches of Modularity, and weird new locations for Devel 
packages. And now we're forced onto this new treadmill that we didn't 
want and have to place a lot of effort into re-evaluating.

RedHat should know better than to pull the rug out from under others 
like this. This is not about saving a few dollars here and there, and 
it's not about resource allocation explicitly. If things needed to be 
re-justified, there's a larger discussion around consolidating Fedora, 
RedHat, and CentOS build processes together than would likely have been 
more fruitful. This is about removing a community expectation and 
forcing users away from stability, knowing that the primary blockers 
toward the community getting what it wants is the startup costs and 
trust factor around a new rebuild. IOW, there's no real reason for 
CentOS Linux not to continue to the end of RHEL8's support, with this 
new process being used as a fork for EL9 and beyond's development. THAT 
would be appreciated.

Instead, RedHat embraced CentOS under its umbrella, extended it beyond 
its original design goal of a binary-compatible rebuild, and has now 
extinguished the Linux rebuild starting next year.

Sounds like another company we all know and love.

-jc

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.centos.org/pipermail/centos-devel/attachments/20201208/325ed12c/attachment-0005.html>