21:00:23 <timburke> #startmeeting swift
21:00:24 <openstack> Meeting started Wed Aug 21 21:00:23 2019 UTC and is due to finish in 60 minutes.  The chair is timburke. Information about MeetBot at http://wiki.debian.org/MeetBot.
21:00:25 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
21:00:27 <openstack> The meeting name has been set to 'swift'
21:00:45 <timburke> who's here for the swift meeting?
21:00:57 <alecuyer> o/
21:01:02 <clayg> o/
21:01:04 <rledisez> hi o/
21:01:14 <mattoliverau> o/
21:01:18 <clayg> alecuyer: been a minute; hope you're doing well
21:01:31 <alecuyer> yep, doing good thanks!
21:02:06 <timburke> agenda's at https://wiki.openstack.org/wiki/Meetings/Swift
21:02:23 <timburke> first up, some openstack news
21:02:31 <timburke> #topic U cycle release naming
21:02:54 <timburke> the poll's done! sorry i forgot to mention it last meeting
21:03:02 <timburke> #link http://lists.openstack.org/pipermail/openstack-discuss/2019-August/008624.html
21:03:16 <timburke> winner is: Ussuri!
21:03:46 <timburke> might still change, of course; still has to be vetted by the foundation and all
21:04:17 <timburke> named after...
21:04:18 <timburke> #link https://en.wikipedia.org/wiki/Ussuri_River
21:05:07 <timburke> on to updates
21:05:24 <timburke> #topic symlink-backed and s3 versioning
21:05:58 <timburke> one of the pre-reqs landed: https://review.opendev.org/633094
21:05:58 <clayg> thank you timburke and thiago for the comments, i'll get to work!
21:05:59 <patchbot> patch 633094 - swift - Allow "static symlinks" (MERGED) - 29 patch sets
21:06:11 <clayg> oh yeah, any last questions about hardlinks?
21:06:47 <timburke> clayg's been doing a great job on https://review.opendev.org/633857
21:06:48 <patchbot> patch 633857 - swift - symlink-backed versioned_writes - 16 patch sets
21:07:14 <timburke> anythign we should be keeping in mind there? upgrade considerations, interactions with other features, etc?
21:07:19 <clayg> did I manage to get that where it runs with use_symlinks = true/false in the gate?
21:07:36 <timburke> yup, works like a champ
21:07:47 <clayg> well, use_symlinks requires symlinks in your pipeline obvs - but it'll ignore the option if it doesn't find it
21:08:06 <timburke> it's nice that you've got some tests that are use_symlinks-specific so i can look for the skips
21:08:22 <clayg> oh good
21:09:43 <timburke> clayg, i guess my main question above is, if you're rledisez, do you turn it on?
21:10:24 <rledisez> timburke: that's a good question
21:10:25 <timburke> and for rledisez, what can/should we do and communicate to help you make that decision?
21:10:58 <rledisez> and the answer is at that point I didn"t look much into it which is bad, so I can't really answer
21:11:14 <clayg> well there *is* the difference with the bytes reporting of new data all moving under the x-[history|versions]-location
21:11:32 <timburke> ...which has impacts on container quotas
21:12:40 <timburke> i *think* it makes them a little easier to reason about, though? *shrug*
21:13:53 <timburke> clayg, anything else we can do for you? or are you just blocked on reviews?
21:14:07 <clayg> i *was* blocked on reviews - but I got two just today!
21:14:23 <clayg> so i'm good for now
21:14:43 <timburke> #topic py3
21:15:05 <timburke> we're down to one more patch to get all of func tests ported! https://review.opendev.org/675227
21:15:06 <patchbot> patch 675227 - swift - py3: Finish porting func tests - 4 patch sets
21:15:44 <zaitcev> Those are mostly easy. I'm still a little concerned about a mixed cluster and migration
21:16:28 <timburke> there's one pre-req (https://review.opendev.org/671167) to getting most of probe ported https://review.opendev.org/671333
21:16:29 <patchbot> patch 671167 - swift - py3: fix up listings on sharded containers - 4 patch sets
21:16:30 <patchbot> patch 671333 - swift - py3: (mostly) port probe tests - 2 patch sets
21:16:46 <timburke> zaitcev, true -- i think there's going to be some headaches there...
21:17:41 <timburke> we've come a long ways, though -- thanks everyone for helping out!
21:18:03 <zaitcev> Maybe we can have Matt look at https://review.opendev.org/#/c/671167/ , track down the types for all the variables again, in case.
21:18:04 <patchbot> patch 671167 - swift - py3: fix up listings on sharded containers - 4 patch sets
21:18:40 <timburke> mattoliverau, got time for it?
21:19:08 <mattoliverau> oh yeah of course, will look at it today!
21:19:15 <timburke> thanks!
21:19:23 <timburke> #topic lots of small files
21:19:30 <timburke> alecuyer's back \o/
21:20:00 <alecuyer> hope everyone's doing good :)
21:20:20 <timburke> fwiw, we here at swiftstack have been doing some testing with it
21:20:57 <alecuyer> yes Romain told me ! That's great. He also told me you had issues, not so great :/ if you can point me at stacks or a way to reproduce I'm happy to try and see what happens
21:21:49 <timburke> i should harangue swifterdarrell into coming back on IRC and talking more about it
21:22:13 <alecuyer> I saw one paste where it would fail to rename a file to #d
21:22:51 <timburke> i was just looking that one up :-)
21:22:52 <timburke> http://paste.openstack.org/show/757015/
21:23:37 <timburke> i should also try to get you a merge from master soon... sorry, it's been a bit
21:24:14 <timburke> is there anything the rest of us can help with, or that we should know about?
21:24:19 <alecuyer> I think that's the one yes, so I have not seen that before and don't know how that would happen but hm I'll re-read the code there
21:25:04 <alecuyer> well quickly: we're finally moving a couple of machines to use native golang leveldb (vs wrapped C++ lib). hasn't blown up yet, haven't seen if the CPU usage decreases as much as I hoped. I'll tell next week!
21:25:17 <alecuyer> and the grpc to http patch "works" (passes functests)
21:25:23 <timburke> \o/
21:25:41 <alecuyer> I still need to write "streaming like" functions for a couple of ancillary functioons (for listing all quarantined objects for example)
21:25:44 <alecuyer> that's it from me
21:26:23 <timburke> fwiw, i think we'd done some more investigation on the traceback and found that there were some concurrent writes going on, so the non-durable file got deleted before it could be made durable
21:26:51 <timburke> (i wasn't the one doing the investigating, though, so this is all second- or third-hand; sorry)
21:26:57 <alecuyer> Ok I'll have to think what happens in the regular diskfile case and see if we match behavior. happy to discuss it out of meeting if you wish
21:27:13 <alecuyer> sure thanks for the info anyway :)
21:27:15 <timburke> sounds good
21:27:27 <timburke> #topic sharding
21:28:03 <mattoliverau> I've pushed a new version of the cleaving empty shard range patch. Which timburke has already looked at.
21:28:26 <mattoliverau> Haven't really looped back round to auto sharding stuff yet.
21:28:37 <mattoliverau> So not too much to report
21:28:53 <timburke> mattoliverau, sorry, i've hijacked your auto-sharding slot a bit -- i should probably change the title in the meeting agenda ;-)
21:29:14 <timburke> i think clayg had a look too
21:29:18 <mattoliverau> no prob at all. It's keeping me thinking about it. so all good :)
21:29:37 <clayg> just trying to wrap my head around it
21:30:04 <mattoliverau> yeah the cleaving functions become a little inception like.
21:30:35 <timburke> i'll see if i can take a deeper look soon, but it probably won't be this week
21:30:39 <mattoliverau> Esp when you want to add a "that doesn't count" down a rabbit hole :P
21:30:44 <mattoliverau> nps
21:31:55 <timburke> i also wanted to call out https://review.opendev.org/#/c/675014/ -- i think it could improve root container responsiveness/availability a good bit
21:31:56 <patchbot> patch 675014 - swift - Latch shard-stat reporting - 4 patch sets
21:32:48 <zaitcev_> Interesting
21:33:09 <timburke> that's all i've got
21:33:15 <timburke> #topic open discussion
21:33:53 <mattoliverau> I'll add it to my list of reviews to try and get to today after a billion meetings
21:34:09 <zaitcev_> I asked about this before, and you sent me to URL that I promptly lost... But now I'm looking here: https://releases.openstack.org/train/schedule.html  and it does not say when the branches are going to be made for projects.
21:34:09 <timburke> thanks mattoliverau :-)
21:36:12 <timburke> zaitcev_, looks like for stein, thebranch was created at the same time that i did the release: https://github.com/openstack/releases/commit/d3f7e561980edfa841077724b4e9aafbe2443cb3
21:36:50 <timburke> notmyname did a similar thing for rocky: https://github.com/openstack/releases/commit/3c44c84dd14d4336c6c8e1b2606048662baae99e
21:37:20 <timburke> queens had a gap of like 5 days
21:37:54 <zaitcev_> Is there going to be an RC1 on our part? Or do we just tell them clone a random commit from the trunk and use that?
21:38:57 <timburke> that's more or less been our existing procedure. would an RC help you? i could look into it... or just make more releases
21:39:05 <zaitcev_> Naah
21:39:25 <zaitcev_> Well, my manager wants to micromanage. But I'm fine
21:40:03 <clayg> 🤣
21:41:19 <zaitcev_> One other question.
21:41:35 <zaitcev_> I see you filed https://bugs.launchpad.net/ossa/+bug/1840507
21:41:36 <openstack> zaitcev_: Error: malone bug 1840507 not found
21:41:36 <timburke> i mean, we've had ~50 commits since 2.22.0 ... new features, py3 fixes, slo fixes, symlink fixes, correctness improvements... i could totally justify a 2.23.0 ;-)
21:41:53 <timburke> heh. yup
21:42:50 <zaitcev_> Weren't we supposed to migrate to some other bug tracking site for Stein, called "Storyline" or some such. Has that happened or not? Or you're just defying OpenDev?
21:43:21 <timburke> storyboard. i've still got concerns about performance
21:45:51 <timburke> i'm not really opposed to moving, but this feels like a notable degradation from launchpad to me: https://storyboard-dev.openstack.org/#!/project/openstack/swift
21:46:18 <zaitcev_> "Firefox detected a potential security threat and did not continue to storyboard-dev.openstack.org. If you visit this site, attackers could try to steal information like your passwords, emails, or credit card details."
21:46:32 <zaitcev_> GJ
21:47:09 <timburke> eh, just add the exception -- you weren't going to put your credit card on there anyway ;-)
21:48:48 <timburke> anything else to bring up?
21:49:21 <zaitcev_> not for me
21:50:13 <timburke> all right
21:50:25 <timburke> thank you all for coming, and thank you for working on swift!
21:50:32 <timburke> #endmeeting