On 3/1/22 14:24, Jonathan Carter wrote:
Hi Phil (and everyone)
On 2022/03/01 15:10, Philip Wyett wrote:
Thank you for the terse response. The two examples i.e. micronews and
the infra list do not sound
that this is scheduled work at all. Better communication from teams
would possibly give better
understanding and the patience of users/developers that is being asked
for.
Our GitLab instance (Salsa), have fallen behind multiple versions. This
is due to a bunch of different hurdles that all came with their own
decisions (I'm going to urge the Salsa admins again to do a write-up
about it).
So what's happening now is point-to-point upgrades between all the
GitLab versions needed on this upgrade path, along with the migrations
between them (which are quite large, Salsa is one of the biggest GitLab
instances out there).
So while a huge amount of pent up maintenance is all happening at once
now, at least regular updates (and security updates) will be able to run
again on short cycles.
(I hope salsa admins don't mind me posting this, but I hope it helps all
our contributors better understand what's going on).
On a practical note, please take note of any uploads you do during this
downtime, and be sure to do a git push along with the tags when Salsa is
back up again.
-Jonathan
Like everyone, I am thankful for the volunteer time given by the Salsa
admins, and it feels great to know Salsa is fully updated.
While I very much appreciate the work, and understand the pain behind
migrating a huge db, I'd say the communication could be perfected.
The initial announcement explained none of the above, and the list
archive storing the announcement was down (because the list archive
needed Salsa to run). People even outside of Debian (like some managing
the Kolla OpenStack CI that uses Debian and extrepo) were affected. They
asked me multiple times what was going on, and it was painful to be
annoying asking on IRC (sorry guys, and thank you formorer and pabs for
replying), and then relaying the small amount of info I could gather. A
status page (anywhere) would have help a lot (not everyone knows how to
read the topic on the #salsa channel). BTW, is there a public status
page for the Debian infra monitoring?
Cheers,
Thomas Goirand (zigo)