21:00:13 <timburke> #startmeeting swift 21:00:13 <opendevmeet> Meeting started Wed Jan 15 21:00:13 2025 UTC and is due to finish in 60 minutes. The chair is timburke. Information about MeetBot at http://wiki.debian.org/MeetBot. 21:00:13 <opendevmeet> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 21:00:13 <opendevmeet> The meeting name has been set to 'swift' 21:00:21 <timburke> who's here for the swift meeting? 21:00:28 <mattoliver> o/ 21:01:10 <mattoliver> I think Jian said he was going to the dentist 21:01:34 <timburke> well, we can get started :-) 21:01:46 <acoles> o/ 21:01:49 <timburke> i didn't put much on the agenda 21:01:54 <timburke> #link https://wiki.openstack.org/wiki/Meetings/Swift 21:02:02 <timburke> first up 21:02:09 <timburke> #topic gate breakage 21:02:12 <kota> o/ 21:02:31 <timburke> a recent bandit release broke out pep8 job 21:02:41 <acoles> boo 21:02:43 <mattoliver> oh yay 21:03:07 <timburke> fortunately the fix was simple and we merged it quickly (thanks acoles!) 21:03:10 <timburke> #link https://review.opendev.org/c/openstack/swift/+/939184 21:03:10 <patch-bot> patch 939184 - swift - CI: Remove B320 and B410 bandit skips (MERGED) - 1 patch set 21:03:16 <acoles> kinda mean to fail because we *excluded* a test that gets removed 21:03:27 <timburke> yeah, i felt similarly :-/ 21:03:44 <acoles> WDYT about my doc follow-up https://review.opendev.org/c/openstack/swift/+/939209 ? 21:03:44 <patch-bot> patch 939209 - swift - CI: document bandit tests by reference - 1 patch set 21:03:56 <timburke> but i didn't bother to submit a bug about it or anything 21:04:09 <JayF> I'll also note a new eventlet landed in global requirements; in oslo.log for eventlet-patched things we had to fix PipeMutex; I have a draft patch for swift's version of PipeMutex here https://review.opendev.org/c/openstack/swift/+/935987 in case you all see any issues; I know swift is a little special so may not be impacted but wanted to surface it. 21:04:09 <patch-bot> patch 935987 - swift - Use stdlib "os" in PipeMutex - 1 patch set 21:04:09 <timburke> approved! yeah, better to just link out to their docs 21:04:33 <timburke> JayF, i saw that! py313 support here we come! 21:04:49 <acoles> OIC you already did +2, thanks 21:05:37 <JayF> FWIW, I don't intend on revising 935987; so if you wanna take it over feelfree, I'll likely abandon it in a few months if it doesn't move :) 21:06:00 <timburke> 👍 21:06:29 <timburke> it's still weird to me that it causes some failures. i should probably dust it off to try to understand more 21:06:59 <timburke> but anyway, if anyone notices any other gate failures where it's just bandit complaining in the pep8 job, feel free to recheck 21:07:21 <timburke> #topic py2 is dead! 21:07:38 <timburke> it's a carryover from last week, but now we've got more folks around to celebrate ;-) 21:07:50 <JayF> timburke: my hunch is that since it unconditionally patches eventlet as part of the new job, and swift only conditionally does it, that there may just need to be some tweaking 21:07:56 <mattoliver> woo hoo! 21:08:06 <JayF> congrats :) 21:08:13 <timburke> the big patch to take out py2 code paths seems to be getting close, too 21:08:15 <timburke> #link https://review.opendev.org/c/openstack/swift/+/853697 21:08:15 <patch-bot> patch 853697 - swift - Remove py2-only code paths - 18 patch sets 21:08:37 <timburke> thanks zaitcev, acoles, and mattoliver! 21:09:22 <timburke> there's still more that can be done, of course 21:09:25 <mattoliver> yeah, it was a big one.. that must've been a tedious patch to write timburke 21:09:47 <mattoliver> thanks for working on it. 21:10:06 <timburke> simplify our super() calls, use subtests, use f-strings... 21:10:24 <mattoliver> We know basically have 3 +2s on it now. So shall I just +A it? 21:10:32 <acoles> I wanna get a `yield from` in somewhere :D 21:10:37 <mattoliver> yeah, we can changes things as we touch files now, which is nice. 21:10:42 <acoles> mattoliver: do it! 21:11:03 <mattoliver> done! 21:11:04 <timburke> i'm definitely looking forward to having a modern codebase 21:11:07 <timburke> \o/ 21:11:43 <acoles> I'm looking forward to not having to nursemaid a py2 dev environment anymore 21:11:55 <timburke> swift isn't the only place we might want to drop some old python versions -- i also put together a pyeclib patch 21:11:56 <timburke> #link https://review.opendev.org/c/openstack/pyeclib/+/938098 21:11:57 <patch-bot> patch 938098 - pyeclib - Drop support for Python 2.7, 3.5-3.8 - 4 patch sets 21:12:48 <timburke> maybe it would've been better for me to split that into "drop py2" and "drop old py3s" -- oh well 21:13:23 <timburke> but there are more patches following it to do some further cleanups in pyeclib, too 21:14:35 <mattoliver> nah it's fine. I feel like its the start of the new year, time to clean some stuff up. We can always put things back. We're cleaning swift, let's clean pyeclib too! 21:16:02 <timburke> the nice thing about dropping the old versions of py3 is that we can switch to using latest image tags when building wheels 21:17:35 <timburke> well, that, and taking out half our unit test jobs 21:17:56 <timburke> all right, last topic i've got 21:18:01 <timburke> #topic swift release 21:18:38 <timburke> i meant to get together an authors/changelog patch last week, but still haven't. hopefully i'll find time this week 21:19:17 <mattoliver> kk, first release without py2! 21:19:23 <timburke> (though i'm a little skeptical -- i've got a downstream release to get out, too, and i'm moving over the weekend) 21:19:41 <mattoliver> well no rush 21:20:02 <timburke> mattoliver, and i'm not even sure it deserves a release note! we announced the last stable release supporting py2 back in *yoga*! 21:20:09 <acoles> it'll be done when it's done 21:20:32 <timburke> #topic open discussion 21:20:41 <timburke> anything else we should bring up this week? 21:21:19 <mattoliver> device id size has become an major issue downstream for us. So looks like ringv2 is bubbling back up 21:21:40 <mattoliver> So I've started reloading and reviewing it again. 21:21:53 <mattoliver> maybe we can include it in the new release too ;) 21:22:06 <timburke> oh yeah! makes me happy i've kept p 834261 reasonably up-to-date 21:22:06 <patch-bot> https://review.opendev.org/c/openstack/swift/+/834261 - swift - ring: Introduce a v2 ring format - 21 patch sets 21:23:55 <mattoliver> We have so many devices in our clusters, be kinda can't expand much more without the abilty to use bigger ints (crazy problem to have). 21:24:50 <opendevreview> ASHWIN A NAIR proposed openstack/swift master: db-replicator: include container name in logs, not just db path https://review.opendev.org/c/openstack/swift/+/933379 21:24:55 <mattoliver> I feel like it was pretty close to landed a year or 2 ago.. so hopefully it wont take too much effort to land it (esp thanks to Tim keeping it up to date). 21:25:16 <timburke> i remember OVH having a similar problem years ago -- i think we've mostly managed to avoid it by buying to maximize TB/drive rather than minimize $/TB 21:25:49 <mattoliver> although I assume me landing the py2 deprecation, we have alot of patches now in conflict :P 21:26:27 <mattoliver> yeah, don't think that'll cut it anymore. So no py2 and soon new ring format! 21:26:36 <mattoliver> 2025 is going to be great! 21:26:38 <timburke> almost certainly :-) it'll be interesting to see what bothers to get rebased the fastest 21:28:44 <mattoliver> that's all I've got. I've been on vacation 21:29:06 <timburke> all right, i think i'll call it then 21:29:17 <timburke> thank you all for coming, and thank you for working on swift! 21:29:20 <timburke> #endmeeting