21:00:52 <timburke> #startmeeting swift
21:00:52 <opendevmeet> Meeting started Wed Aug 20 21:00:52 2025 UTC and is due to finish in 60 minutes.  The chair is timburke. Information about MeetBot at http://wiki.debian.org/MeetBot.
21:00:52 <opendevmeet> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
21:00:52 <opendevmeet> The meeting name has been set to 'swift'
21:01:04 <timburke> who's here for the swift meeting?
21:01:15 <mattoliver> o/
21:02:05 <timburke> as usual, the agenda's at
21:02:08 <timburke> #link https://wiki.openstack.org/wiki/Meetings/Swift
21:02:16 <timburke> first up
21:02:22 <timburke> #topic vPTG
21:02:53 <timburke> just a reminder that we've got the ptg scheduled in a couple months
21:04:01 <timburke> i still need to get that etherpad up -- i got as far as opening a tab with the URL i wanted, but didn't update it at all, and now i've got to find the tab again :P
21:04:09 <timburke> next up
21:04:16 <timburke> #topic liberasurecode release
21:04:17 <mattoliver> Lol
21:04:28 <timburke> so the new backend is merged!
21:04:35 <mattoliver> Yay!
21:04:48 <timburke> next step toward people actually getting to use it is to get it into a release ;-)
21:04:56 <timburke> i threw together some notes
21:05:02 <mattoliver> Great idea
21:05:04 <timburke> #link https://review.opendev.org/c/openstack/liberasurecode/+/958118
21:05:31 <mattoliver> I'll review them today
21:06:03 <timburke> thanks -- i was thinking i ought to try to get them merged and cut a release by the end of the week
21:06:45 <timburke> then i ought to start on doing the same for pyeclib -- after making sure that its wheels would include the latest and greatest libec :D
21:07:04 <timburke> speaking of releases...
21:07:14 <timburke> #topic 2.35.1 stable release
21:07:44 <mattoliver> Releases releases everywhere!
21:07:48 <timburke> i got those aws-chunked patches merged
21:07:56 <mattoliver> Nice
21:07:57 <timburke> and got some notes up for them, too
21:08:00 <timburke> #link https://review.opendev.org/c/openstack/swift/+/956831
21:08:32 <opendevreview> Merged openstack/swift master: Revert "Remove unused MANIFEST.in"  https://review.opendev.org/c/openstack/swift/+/950615
21:09:20 <timburke> i'm torn about whether to push for that ahead or behind a 2.36.0 intermediary release; mostly i just want to make sure users can still use boto3 et al out of the box
21:09:57 <timburke> heh -- the gate beat me to the next topic
21:10:08 <timburke> #topic bringing back MANIFEST.in
21:10:38 <timburke> more downstream packagers are affected than just nvidia
21:11:12 <mattoliver> lol
21:11:33 <timburke> thanks mattoliver for figuring out that the issue comes from trying to build sdists and wheels from a bare tarball with no git history
21:12:00 <mattoliver> Yeah no probs... took me longer to figure out then I'd like to admit :P
21:12:31 <timburke> i'll have to try to remember to include that in my personal packaging tests then next time there's a patch like that
21:13:01 <timburke> (like when i get around to rebasing https://review.opendev.org/c/openstack/pyeclib/+/956612 )
21:13:42 <timburke> next up
21:13:50 <timburke> #topic gate issues
21:14:22 <timburke> recently our py37 jobs all started failing without providing and logs to try to diagnose
21:14:48 <mattoliver> well that's useful of them :P
21:15:26 <timburke> acoles jumped in -infra and got some help -- the issue was that they were using ubuntu-bionic nodes which don't support zuul's new default ansible version (11)
21:15:42 <timburke> we've got fixes merged now to master
21:15:44 <timburke> #link https://review.opendev.org/c/openstack/swift/+/957959
21:16:12 <timburke> where we just get off bionic -- we can use pyenv to install py37 on jammy
21:16:22 <timburke> and also stable/2025.1
21:16:30 <timburke> #link https://review.opendev.org/c/openstack/swift/+/958002
21:16:32 <mattoliver> oh nice
21:17:24 <mattoliver> So we don't need the nodeset anymore?
21:17:31 <timburke> there, things got a little trickier: py36 could get installed via pyenv, but when we went to install tox, it dragged in a too-new packaging that required py37+
21:18:03 <timburke> so for that stable branch, we bring py37 up to jammy and pin the ansible version back to 9 for the py36 jobs
21:18:13 <timburke> mattoliver, not on master ;-)
21:18:47 <timburke> we still explicitly specify a nodeset in the base job -- at some point i should try bumping it up from jammy to noble
21:18:49 <mattoliver> wow, so it was ansible pulling it up to 3.7
21:19:42 <mattoliver> Oh that makes sense, being set in some parent somewhere.
21:20:12 <timburke> i still need to fix last year's stable branches, though. and there we've got to deal with py27, too :-(
21:20:30 <mattoliver> oh bugger, yeah
21:20:33 <timburke> speaking of branch maintenance...
21:20:37 <timburke> #topic unmaintained branches
21:20:57 <timburke> we've currently got 6 unmaintained branches, going back to victoria
21:21:11 <timburke> i propose we eol all of them
21:21:17 <timburke> #link https://review.opendev.org/c/openstack/releases/+/958101
21:21:56 <timburke> there have been a handful of patches against them -- but as best i can tell they were all to do with trying to fix CI
21:22:49 <mattoliver> yeah makes sense.
21:23:00 <mattoliver> wow there are a some old ones there.
21:23:01 <timburke> half of them haven't been touched all year. of the two that have open patches against them, the patches were also trying to maintain CI -- and none of them have it passing yet
21:23:32 <timburke> i sent a note to the mailing list about it this morning
21:23:34 <timburke> #link https://lists.openstack.org/archives/list/openstack-discuss@lists.openstack.org/thread/3CJPUXP5TVI2DMCSHONCDM44BFHCJXIZ/
21:23:43 <timburke> but i don't expect any objections
21:23:54 <timburke> new swift is best swift
21:24:27 <mattoliver> +100
21:24:50 <timburke> all right, i think that's all i've got
21:24:55 <timburke> #topic open discussion
21:25:03 <timburke> anything else we want to bring up this week?
21:26:05 <mattoliver> I've been playing with ringv2 followups.. but nothing to really review yet. Addressing some of your comments from 2 years ago :P But keep a look out for that
21:26:48 <timburke> can do. i know i owe some other folks reviews, too
21:27:39 <mattoliver> just playing with adjusting size of the past tables when we ajust ring size. And also removing removed devs from history (so a new reused dev_id on a new device doesn't get all confused).
21:27:41 <mattoliver> *confusing
21:28:00 <mattoliver> Will push something up today
21:29:03 <mattoliver> otherwise, I can't think of anything that interesting to bring up, we talked about MANIFEST.in so I think I'm good for now :)
21:30:14 <timburke> good call on removing devices from the history -- if we got them out of the ring entirely, presumably they aren't coming back
21:30:53 <mattoliver> yeah that's my thought. And don't want to go looking for them
21:31:52 <timburke> all right, i think i'll call it early then, and get back to re-organizing the cli tools i want to add to pyeclib
21:32:07 <timburke> thanks for coming, and thank you for working on swift!
21:32:13 <timburke> #endmeeting