[Gluster-infra] [Gluster-Maintainers] [gluster-packaging] Fwd: [CentOS-devel] https://blog.centos.org/2020/12/future-is-centos-stream/

[Michael Scherer]

Le vendredi 11 décembre 2020 à 16:01 +0530, Amar Tumballi a écrit :
> On Thu, Dec 10, 2020 at 11:22 PM Michael Scherer <mscherer at redhat.com
> >
> wrote:
> 
> > Le jeudi 10 décembre 2020 à 22:06 +0530, sankarshan a écrit :
> > > What is your recommendation? As in, the next steps from here.
> > 
> > - check if there is c8s image on amazon already
> > 
> > if there is one
> > - switch the image and reinstall the builders (3rd time this week,
> > so I
> > should not stumble like the previous 2)
> > 
> > if not
> > - install centos8-stream-release rpm, dnf upgrade -y, reboot
> > - add the rpm in ansible so it will be here if we reinstall
> > - wait for a proper ec2 image and add it to ansible
> > 
> > Given we had kernel bugs in the past (
> > https://github.com/gluster/glusterfs/issues/1402#issuecomment-666358241
> >  ), I think  faster access to fixes for the CI (or even for
> > production)
> > is a good idea.
> > 
> > 
> 
> Please go ahead! This looks like a good plan to be adhering to the
> centos
> stream... For me,  from outside, centos8-stream is similar to fedora
> rawhide, a moving base.

On the delivery, yeah. But centos-stream is not going to get any
radical update, because things go to RHEL after. There is internal
gating in place to verify the ABI (and kABI) is not broken, and people
aren't going to push "latest upstream" version unlike Rawhide. And the
process of deciding what feature go, and what can be supported is still
the same. 

But for example, something like 
https://bugzilla.redhat.com/show_bug.cgi?id=1762161#c0 would be easier
to correct, since we would have to wait for RHEL release (and/or
exfiltrate kernel rpms from the intranet)


I understand why folks make a fuzz about c8s (our com was not great, to
say the least...), but in practice, that's just swapping RHEL and
Centos order in the pipeline (before, stuff went to RHEL, then got
rebuild by Centos, now that will be the reverse, more or less).

> As long as tests are running fine, we are fine, when it breaks, to
> debug,
> there are more than glusterfs elements to consider now.

-- 
Michael Scherer / He/Il/Er/Él
Sysadmin, Community Infrastructure



-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 836 bytes
Desc: This is a digitally signed message part
URL: <http://lists.gluster.org/pipermail/gluster-infra/attachments/20201211/4ead0507/attachment.sig>