21:00:16 <timburke> #startmeeting swift 21:00:16 <opendevmeet> Meeting started Wed Jan 17 21:00:16 2024 UTC and is due to finish in 60 minutes. The chair is timburke. Information about MeetBot at http://wiki.debian.org/MeetBot. 21:00:17 <opendevmeet> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 21:00:17 <opendevmeet> The meeting name has been set to 'swift' 21:00:23 <timburke> who's here for the swift team meeting? 21:00:29 <kota> o/ 21:01:39 <timburke> maybe it'll just be us :-) 21:01:52 <kota> okay 21:01:53 <timburke> as usual, the agenda's at 21:01:55 <timburke> #link https://wiki.openstack.org/wiki/Meetings/Swift 21:02:19 <timburke> first up, the main thing on my mind lately... 21:02:26 <timburke> #topic eventlet 21:02:34 <mattoliver> o/ 21:02:40 <mattoliver> Sorry im late 21:02:45 <timburke> no worries! 21:03:38 <timburke> so there's been ongoing tumult -- with p 904147 we started testing with latest eventlet, and our gate promptly broke 21:03:39 <patch-bot> https://review.opendev.org/c/openstack/requirements/+/904147 - requirements - Update eventlet to 0.34.3 (MERGED) - 3 patch sets 21:04:06 <timburke> the good news is, there's a fix already merged against eventlet 21:04:18 <timburke> #link https://github.com/eventlet/eventlet/pull/890 21:04:40 <mattoliver> Oh cool 21:05:00 <timburke> but it seems like until there's an upstream release and a fresh u-c change to use *that*, our gate's broken 21:05:27 <mattoliver> Do we know how long they'll be? 21:05:27 <jianjian> it's Tim who fixed it 21:06:28 <timburke> there was also an attempt to revert the up-rev in p 905905 but it looks like the cross-project testing will prevent that from merging, despite the approvals 21:06:28 <patch-bot> https://review.opendev.org/c/openstack/requirements/+/905905 - requirements - Revert "Update eventlet to 0.34.3" - 1 patch set 21:06:50 <timburke> mattoliver: fingers crossed on a day or two 21:07:11 <zaitcev> Hold on, this is not everything. I saw a fix on our part, which I thought was good, but it didn't work. 21:07:12 <jianjian> is our broken gate caused by merged 904147? that one passed gate before merging it. 21:08:03 <zaitcev> https://review.opendev.org/c/openstack/swift/+/905796 21:08:04 <patch-bot> patch 905796 - swift - Work with latest eventlet (again) - 1 patch set 21:08:34 <timburke> jianjian, yeah -- the cross-project testing job should have caught the breakage, but it doesn't run with an xfs temp dir, so some tests get skipped 21:09:01 <timburke> unfortunately, all of the tests that would have caught it are end-to-end tests that require xfs 21:09:57 <timburke> zaitcev, yeah, that patch would have had us working with eventlet 0.34.2 -- then after it was merged, there was a 0.34.3, with fresh breakage 🙄 21:11:35 <timburke> and actually, i misspoke -- it looks like we need to have (1) a new eventlet release, (2) a requirements change to pull in the new release, *and* (3) a small fix to opt in to some old eventlet behavior 21:11:46 <timburke> #link https://review.opendev.org/c/openstack/swift/+/905796 21:11:46 <patch-bot> patch 905796 - swift - Work with latest eventlet (again) - 1 patch set 21:12:37 <timburke> oh! sorry, zaitcev, that's the patch you were linking -- i was thinking of p 904459. ugh. 21:12:38 <patch-bot> https://review.opendev.org/c/openstack/swift/+/904459 - swift - Get tests passing with latest eventlet (MERGED) - 2 patch sets 21:14:22 <timburke> in an effort to avoid this particular fire drill in the future, i also wrote up a new test to exercise the header-parsing bug that was missed but without requiring xfs 21:14:25 <timburke> #link https://review.opendev.org/c/openstack/swift/+/905997 21:14:26 <patch-bot> patch 905997 - swift - tests: Exercise current eventlet breakage off XFS - 1 patch set 21:16:27 <timburke> separate from our particular gate breakage, it looks like there's some interest in up-revving eventlet minimum requirements across all of openstack 21:16:31 <timburke> #link https://lists.openstack.org/archives/list/openstack-discuss@lists.openstack.org/thread/7PJZ4WYRS2B3PIZ5KYFL6V3JWYL4V5FQ/ 21:16:45 <timburke> the particular patch for swift is 21:16:48 <timburke> #link https://review.opendev.org/c/openstack/swift/+/905945 21:16:49 <patch-bot> patch 905945 - swift - bump eventlet to latest version that support pytho... - 1 patch set 21:17:18 <timburke> ...which i've got more than a couple reasons for opposing ;-) 21:20:10 <timburke> all this is to say, i think we're stuck waiting on other people to unstick our gate at the moment, so don't worry about issuing rechecks for now 21:20:43 <jianjian> got it 21:20:54 <kota> kk 21:21:01 <mattoliver> and the question raises again, when will we drop py2 support ;) 21:21:13 <timburke> unless we wanted to fix the affected tests to stop issuing both content-length and transfer-encoding... 🤔 21:22:21 <timburke> i'll admit, i hadn't actually looked into whether the tests were intentionally testing that specifically, or just doing it accidentally 21:23:38 <timburke> mattoliver, yeah -- but even then, we declare py36 support, too. maybe it'd be reasonable for us to drop that at the same time? either way, i know i've got some environments that still (regrettably) need py2, so... 21:24:11 <mattoliver> kk 21:24:37 <clarkb> I mentioned this in the tc channel earlier today but it might be worthwhile seeing if eventlet is interested in third party ci from projects like swift 21:24:49 <clarkb> then hopefully we catch these things before they even merge upstream 21:25:06 <timburke> as an eventlet maintainer, i certainly wouldn't be opposed ;-) 21:25:25 <mattoliver> luckily we have timburke but not a bad idea :P 21:26:34 <timburke> any other questions or comments on eventlet or the state of our gate? 21:27:03 <clarkb> timburke: I think the first thing to get going would be that new non voting job for swift that installs eventlet from source. Then we can configure zuul to run that same job against PRs to eventlet if the eventlet projects agrees (sounds like there is already interest) 21:27:07 <clarkb> thats all I had 21:28:01 <timburke> clarkb, good next step! i can work on getting that together 21:29:01 <timburke> next up 21:29:12 <timburke> #topic part-number api 21:29:22 <timburke> #link https://review.opendev.org/c/openstack/swift/+/894570 21:29:23 <patch-bot> patch 894570 - swift - slo: part-number=N query parameter support - 86 patch sets 21:29:29 <timburke> #link https://review.opendev.org/c/openstack/swift/+/894580 21:29:30 <patch-bot> patch 894580 - swift - s3api: Support GET/HEAD request with ?partNumber - 94 patch sets 21:30:10 <timburke> i know i said i'd take a look last week, but i didn't get to it. will try again this week 21:31:06 <mattoliver> yeah me either, I got sidetracked on namespaces and then had yesterday off 21:31:19 <timburke> thanks, mattoliver. if anyone else wants to take a look before merging, now's the time :-) 21:31:36 <timburke> but on that note... 21:31:43 <timburke> #topic get_namespaces api 21:32:00 <timburke> the main two patches merged! good job jianjian and acoles! 21:32:23 <timburke> and mattoliver! i forgot you originally wrote the proxy patch :-) 21:33:34 <timburke> i think acoles was still hoping to have one more patch; might be good for us to look at it while it's all fresh in our heads 21:33:34 <timburke> #link https://review.opendev.org/c/openstack/swift/+/901335?usp=dashboard 21:33:35 <patch-bot> patch 901335 - swift - proxy: only use listing shards cache for 'auto' li... - 20 patch sets 21:33:35 <jianjian> yeah, finally they are all merged 21:33:35 <mattoliver> lol 21:34:15 <jianjian> that's a good follow-on, Mat and I are looking at it. 21:35:14 <mattoliver> I really like the 3rd patch thats left. But its a pretty big refactor. I think its a big win. Just we just want to be sure we covered all cases. So will test it a bunch today in my vsaio 21:35:50 <timburke> 👍 21:35:59 <timburke> all right, that's all i had 21:36:04 <timburke> #topic open discussion 21:36:13 <timburke> is there anything else we should bring up this week? 21:37:44 <mattoliver> I'm really close to pushing up a new version of sharding sync points patch that has tests and is much better.. I'll get it up today incase anyone is insterested in having a look or play. 21:38:01 <mattoliver> But it's not a high priority patch so no rush 21:38:52 <mattoliver> #link https://review.opendev.org/c/openstack/swift/+/905064 21:38:52 <patch-bot> patch 905064 - swift - wip: shard replication sync points - 3 patch sets 21:39:50 <mattoliver> oh and timburke thanks for +a ing p 829476 21:39:50 <patch-bot> https://review.opendev.org/c/openstack/swift/+/829476 - swift - cli: add --sync to db info to show syncs - 3 patch sets 21:40:07 <mattoliver> seems we didn't land it like we thought.. I dropped the ball there 21:40:16 <mattoliver> thanks all I got 21:40:19 <timburke> sure! seemed silly not to have it :-) 21:42:30 <timburke> and i'll make sure i nudge it along once the gate's fixed -- as well as anything else! so don't be afraid to keep reviewing and voting until then 21:42:57 <mattoliver> kk 21:44:18 <timburke> all right, i think i'll wrap this up early then 21:44:30 <timburke> thank you all for coming, and thank you for working on swift! 21:44:34 <timburke> #endmeeting