On Mon, Apr 1, 2019 at 6:07 PM Sandro Bonazzola <sbonazzo(a)redhat.com> wrote:
Il giorno lun 1 apr 2019 alle ore 16:04 Nir Soffer <nsoffer(a)redhat.com> ha
scritto:
>
> Fedora 28 will be EOL in few month, and at this point people are migration to Fedora
29.
>
> I tried yesterday to build engine development setup on Fedora 29 VM, and found that
it does
> not work:
> - some packages are missing on engine side
> - after hacking the ovirt.repo files, pointing to "28" instead of
"$releasever", I got the packages...
> - but then adding a host fail with "Python is missing" on the host, which
is working Fedora 28 host
> used with engine before.
>
> What are our plans for supporting Fedora 29 on engine and host?
>
> I hope that the work we did for Fedora 28 will make it easier to port to Fedora 29.
I would rather target Fedora 30 which should be available 1 month before Fedora 28 EOL.
What do you think?
Our fedora support was (partially-?) broken for a rather long time now, but we
obviously hope to get soon again to a "fully-supported" state. Assuming we
will,
we'll then have, IMHO, to have a clear and public policy about it. In the past,
we supported "every" version, and this obviously has costs. If we do not want
to
pay these costs, might as well officially declare that we support only
even-numbered
releases, and do not skip over the next odd-numbered version only if there is
a clear reason. If, OTOH, we do want to support every version (and even if not),
it might be worth it to enumerate the steps needed to add full support
for a next
version, and consider automating some/much/most of this.
That said, personally I use CSB on my laptop, so do not feel that strongly the
need to upgrade. I only used fedora on VMs. I do consider to change to fedora,
and might do this when I get my next laptop (in several months or so)...
(Also, I think such threads should be on some mailing list, devel@ or
rhev-devel@).
Best regards,
--
Didi