21:01:40 <timburke> #startmeeting swift 21:01:40 <opendevmeet> Meeting started Wed Jan 11 21:01:40 2023 UTC and is due to finish in 60 minutes. The chair is timburke. Information about MeetBot at http://wiki.debian.org/MeetBot. 21:01:40 <opendevmeet> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 21:01:40 <opendevmeet> The meeting name has been set to 'swift' 21:01:50 <timburke> who's here for the swift team meeting? 21:01:55 <mattoliver> o/ 21:02:00 <kota> o/ 21:03:26 <acoles> o/ 21:03:27 <timburke> as usual, the agenda's at 21:03:29 <timburke> #link https://wiki.openstack.org/wiki/Meetings/Swift 21:03:47 <timburke> first up 21:03:52 <timburke> #topic PTGs 21:04:03 <timburke> there are two currently scheduled! 21:04:23 <timburke> first will be another vPTG in late March 21:05:01 <timburke> then there will be another in-person one as part of the summit in mid June 21:05:42 <timburke> summit's in Vancouver; early bird pricing for it ends mid-February 21:05:59 <mattoliver> oh nice, I hope I'll be able to go, Vancouver is great 21:06:18 <timburke> my assumption is that i should for sure get things rolling for the vPTG 21:06:30 <acoles> has it been confirmed/announced that the summit will also host a PTG ?- I saw a survey but no confirmation 21:07:14 <timburke> acoles, my understanding is yes, it's confirmed 21:07:16 <timburke> #link https://lists.openstack.org/pipermail/openstack-discuss/2023-January/031645.html 21:07:46 <timburke> "So, an in-person PTG will take place Wednesday and Thursday! June 14 -15th." 21:08:25 <kota> will the swift room be located there (Vancouver)? 21:09:05 <timburke> if people are going, i'd love to have a room for us 21:09:19 <kota> nice 21:10:19 <timburke> if having a room makes it easier for people to get approval to go, i'd consider going even on my own dime 21:10:59 <timburke> so my homework for next week: 21:11:11 <timburke> 1. get doodle up for vptg times 21:11:21 <timburke> 2. get etherpad up to collect topics 21:11:57 <timburke> 3. talk to employer about going to vancouver 21:12:43 <timburke> speaking of broader openstack things... 21:12:55 <timburke> #topic next release name 21:12:58 <mattoliver> I'm happy to handle the etherpad again if you want 21:13:10 <timburke> that'd be great, thanks mattoliver 21:13:32 <timburke> the coming release is antelope 21:13:39 <timburke> following that will be bobcat! 21:13:58 <timburke> #link https://lists.openstack.org/pipermail/openstack-discuss/2023-January/031666.html 21:14:27 <timburke> i think i like the animal naming convention we've landed on for the second trip through the alphabet :-) 21:14:57 <acoles> could the 's' release be .... 'swift' :D 21:15:00 <mattoliver> pretty cool were at the 2nd round through the aplhabet. 21:15:16 <mattoliver> should be ;) 21:15:24 <kota> hope it 21:15:47 <kota> OpenStack Swift version Swift. hmm... :D 21:16:49 <timburke> on to patches! there are a handful that have been kicking around for a bit and i wanted to draw attention to them again 21:17:12 <timburke> #topic cleaning up deprecations 21:17:16 <timburke> #link https://review.opendev.org/c/openstack/swift/+/851100 21:17:26 <mattoliver> cool, I've been away, so a bunch of patch reminders will be greatly appreciated :) 21:17:45 <timburke> we've had py310 unit tests for a bit now 21:18:09 <timburke> and following https://review.opendev.org/c/openstack/swift/+/869600 we've got py310 func tests in the gate 21:18:51 <mattoliver> oh I remember this one, it was a follow up, to some of the py310 stuff, So I'll take a look at this today. 21:19:20 <timburke> cleaning up the deprecation warnings seems desirable in part because we've seen warning emissions lead to longer job times (potentially even causing job timeouts) 21:19:34 <timburke> thanks mattoliver! 21:20:03 <timburke> it should also put us in a better position to support py311 and beyond 21:20:20 <mattoliver> nice 21:20:29 <timburke> speaking of... 21:20:40 <timburke> #topic testing under py311 21:21:00 <acoles> I'll look at it again 21:21:44 <timburke> i haven't tried to get a gate job for it up yet, but i *have* at least started running unit tests locally -- and it looks like our sqlite greening will need some updating 21:21:47 <timburke> #link https://review.opendev.org/c/openstack/swift/+/866051 21:23:30 <timburke> iirc, unit tests should pass on python 3.11.1 with that 21:26:18 <mattoliver> ok interesting, seems to be just overloading the execute I guess. I'll put it on my list 21:27:17 <timburke> i'm a little concerned about what impact that patch could have on earlier versions of python, though -- as best i can tell, the earlier versions would have Connection.execute eventually call self.cursor().execute, and i'm not entirely sure how things will react if *both* Connection.execute and Cursor.execute have been greened like that 21:28:44 <mattoliver> is there more info on the bug, I guess I can just run py3.11 without this patch and poke around 21:29:07 <timburke> i suppose i could be a little more defensive and only add the greened Connection.execute on py311+ ... but that doesn't seem great either... 21:29:50 <timburke> yeah, i think you'll probably just want to run unit tests and see how they fail. focus on the db/backend tests 21:30:42 <mattoliver> kk 21:31:55 <timburke> #topic own-shard-range epoch reset 21:31:58 <timburke> #link https://review.opendev.org/c/openstack/swift/+/809969 21:32:19 <mattoliver> oh this is kinda one of mine, busy day for me :P 21:32:43 <timburke> i think i can be the one promising to review it at least :-) 21:33:25 <timburke> just want to keep it on people's radars -- clayg was +2 on it a while back, but then we let is slip into conflict 21:33:49 <timburke> i need to refresh my memory about what (if any) lingering concerns i had 21:33:58 <mattoliver> kk, I'll take a look at the state of it at least. 21:33:59 <mattoliver> me too 21:34:37 <timburke> #topic nested shard range lookups 21:34:40 <timburke> #link https://review.opendev.org/c/openstack/swift/+/863562 21:34:41 <mattoliver> I think we wanted to maybe update some log lines to they were better. 21:34:59 <mattoliver> otherwise this patch that was suppose to lock the problem down seems to have fixed it as we haven't seen the problem again since carrying this downstream 21:35:12 <mattoliver> oh this one isn't mine! 21:35:16 <mattoliver> yay 21:35:22 <timburke> heh 21:35:36 <mattoliver> is this the flat shard range tree in memcache? 21:35:46 <timburke> yup 21:35:50 <mattoliver> I guess I should just follow the link and stop being lazy 21:36:10 <timburke> we've been doing some profiling in prod recently, and finding the update target keeps showing up as a hot piece of code 21:37:40 <timburke> some of that comes from deserializing a bunch of json, some of it's instantiating a bunch of ShardRange objects, some of it seems to be instantiating a bunch of Timestamp objects -- this patch ought to help with all of those as i recall 21:39:33 <timburke> that's all i've got 21:39:38 <timburke> #topic open discussion 21:39:47 <timburke> anything else we ought to discuss this week? 21:40:57 <acoles> sorry, I have been too distracted to work more on the nested shard lookups 21:41:11 <mattoliver> nah, I'm still ramping back up after holidays. I have rebased the tracing stuff. So if anyone ever gets bored you can go review and fix my code :P 21:41:20 <acoles> if anyone can nudge it forwards from WIP, please do 21:41:26 <mattoliver> or just want to play with tracing, also would be useful 21:41:35 <mattoliver> acoles: I'll try to take a look this week. 21:41:48 <acoles> :thum 21:42:02 <acoles> lol 👍 21:42:17 <mattoliver> it's kinda important, and probably something I understand somewhat, so I should probably review it :) 21:42:57 <mattoliver> I've been playing with a shrink older than patch 21:43:12 <mattoliver> just so we can make some movement on older shard containers that really need a shrink 21:43:31 <mattoliver> got the basic stuff all plumbed and some initial tests, feel like it's working. 21:44:02 <mattoliver> Just so we can start making progress while we still perfect early-active srhinking 21:44:26 <mattoliver> #link https://review.opendev.org/c/openstack/swift/+/867758 21:44:55 <timburke> nice -- i'll try to take a look 21:44:56 <mattoliver> needs more tests in code and need to build up more confidence in it. 21:45:39 <mattoliver> ta 21:46:37 <timburke> all right, i think i'll call it for this week 21:46:50 <timburke> thank you all for coming, and thank you for working on swift! 21:46:53 <timburke> #endmeeting