Kevin Fenzi wrote:
When there were secondary arch koji's it took a number of people
full
time to nurse along koji shadow. Often things would land in primary,
secondary would find a bug later and there would need to be a lot of
coordination and back and forth before things were fixed in both places.
Now all those folks can concentrate on the bugs as they happen in
primary koji, fix them faster and everyone wins.
Well, of course it is less work for the architecture maintainers, because
they offloaded almost all the work onto us package maintainers, who have to
fix the build on architectures we don't give a darn about because they block
our entire builds. And also wait for them to complete, though sadly the main
offender there (armv7hl) has even been declared a primary architecture.
We could all get our work done much faster if the primary Koji were only
x86_64 and everything else (including ARM) were built asynchronously and
taken care of by the handful people that actually cares about the
architecture.
If there were constant outages you might have more of a point, but
this
is the only multi-day one I can think of,
One multi-day outage is one too many.
The outage is delaying the fix for showstopper bugs such as:
https://bugzilla.redhat.com/show_bug.cgi?id=1664181
(a package that is essentially unusable, that we will be able to fix only on
Tuesday, after the outage).
it's over a weekend,
The weekend is the only time where most of us unpaid volunteers have time to
take care of our packages. Fedora is not just Red Hat employees!
In addition, the outage is scheduled to last practically the whole of Monday
in Europe too. (The declared end of Monday 22:00 UTC is Monday 23:00 CET.
And from EET eastwards, the outage lasts the whole of Monday too.)
noarch packages can keep building and if there's a security or
urgent
issue we can just Exclude s390x until it's back up.
But such bogus ExcludeArch is going to create significant extra work for the
package maintainers and/or the architecture maintainers to undo later. So it
is at most a solution for critical security issues.
In short, I think having the other arches in the one koji is a win
for
everyone.
And I have to respectfully disagree. It means more work for almost everyone
(there are disproportionally more package maintainers than architecture
maintainers), slower builds, and additional single points of failure (like
the one single S/390 mainframe that is apparently the physical host for ALL
the s390x "builders").
Kevin Kofler