21:00:00 <timburke> #startmeeting swift
21:00:01 <openstack> Meeting started Wed May 15 21:00:00 2019 UTC and is due to finish in 60 minutes.  The chair is timburke. Information about MeetBot at http://wiki.debian.org/MeetBot.
21:00:02 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
21:00:04 <openstack> The meeting name has been set to 'swift'
21:00:10 <timburke> who's here for the swift meeting?
21:00:12 <mattoliverau> o/
21:00:16 <kota_> hi
21:00:22 <patchbot> I'm here!
21:00:23 <tdasilva> hello
21:00:56 <clayg> o/
21:01:01 <alecuyer> hello o/
21:01:39 <timburke> #link https://wiki.openstack.org/wiki/Meetings/Swift
21:01:55 <timburke> agenda's got a bunch of stuff we want updates on, and now everybody's had a little break from the PTG so there may have even been some progress ;-)
21:02:11 <timburke> but first, what do we think about a libec release?
21:02:19 <timburke> #topic liberasurecode release
21:03:06 <tdasilva> timburke: what's in it?
21:03:13 * mattoliverau hasn't been keeping an eye on libec. I assume it's been a while. So long as there are some patches landed we want to get out, then yes.
21:03:41 <timburke> zaitcev's been great about getting some warnings cleaned up, and it looked like someone (prometheanfire, maybe?) was bumping into some issues that are already fixed on master
21:04:15 <mattoliverau> opps only the first 1/2 of that was suppose to be emoted -ETOOEARLY :P
21:04:29 <tdasilva> found this fix: https://github.com/openstack/liberasurecode/commit/7e97b2f808e1115425eef527899b27528f05af3a
21:04:38 <timburke> there hasn't been *that much* code that's landed, but it's also been... 9 months or so since our last release
21:05:14 <kota_> oic. that's long time.
21:06:13 <timburke> tdasilva, the other patch of interest (in my view) is `3564b71 Remove get_chksum to hush warnings (by Pete Zaitcev, 3 months ago)`
21:06:47 <timburke> so, i'm thinking we ought to tag a 1.6.1
21:07:05 <tdasilva> +1
21:07:09 <mattoliverau> crc32 and swift on s390 seems like a good enough reason. Less warning is good too.
21:07:11 <mattoliverau> +1
21:07:29 <timburke> and once we get around to reviewing and landing the quadiron stuff, we can do a 1.7.0
21:07:59 <timburke> but the nice thing is, we can do these basically whenever we want. this was mostly just a heads-up
21:08:11 <timburke> on to updates!
21:08:15 <timburke> #topic updates
21:08:33 <timburke> #topic tempurl restrictions
21:08:38 <mattoliverau> cool
21:08:45 <timburke> mattoliverau, have you had a chance to look at this yet?
21:09:25 <mattoliverau> so I almost pushed something up last night. but just fiddling with some SLO tests and want to check with symlinks
21:09:43 <mattoliverau> I still need to play with symlinks, but can push what I have first
21:10:31 <mattoliverau> So when using tempurl prefixes + PUT if you head other containers objects and you always get a 401, if the object exists or not
21:11:20 <timburke> is the same true for HEADing objects in the same container but with a different prefix?
21:11:30 <mattoliverau> when you allow X-Object-Manifest you can put what you want (items starting with prefix or not) and a GET to the DLO will get the contents. Which seems like the initial bug..
21:12:16 <mattoliverau> I think so, I think I have a func test for that too.. but will confirm.
21:12:17 <timburke> hmm...
21:13:00 <timburke> well, good to know. we'll keep an eye out for more tests!
21:13:13 <timburke> #topic migration to storyboard
21:13:32 <mattoliverau> Will push up some initial tests today :)
21:13:59 <timburke> i know we asked to get a test migration going so we could start playing with the new system -- has that happened yet, or are we still waiting?
21:14:05 <mattoliverau> I saw diablo_rojo join when the meeting started
21:14:31 <mattoliverau> I haven't heard anything, but will follow up with her today (if she's around).
21:14:32 <timburke> \o/ diablo_rojo!
21:14:39 <timburke> sounds good
21:14:54 <timburke> that was where i expected it to be, just wanted to check
21:15:04 <mattoliverau> if summoning diablo_rojo doesn't work, what about diablo_rojo_phon?
21:15:07 <diablo_rojo> o/
21:15:10 <mattoliverau> \o/
21:15:13 <diablo_rojo> Haha
21:15:25 <kota_> lol
21:15:45 <diablo_rojo> So.. last week was not as productive as I would like it to have been, but I literally wrote the test migration down on my todo list for this week.
21:15:54 <diablo_rojo> I will have it done by the next meeting.
21:15:58 <mattoliverau> cool
21:15:59 <timburke> yay! thanks diablo_rojo!
21:16:03 <mattoliverau> thanks!
21:16:23 <diablo_rojo> I can go greate the project now in the dev storyboard if you want to make stories now.
21:16:32 <diablo_rojo> It wouldnt interfere with the migration.
21:16:33 <timburke> i don't think anyone's actually getting impatient for it, i just want to make sure it doesn't fall off our radar :-)
21:16:46 <timburke> so, up to you
21:16:47 <diablo_rojo> timburke, I appreciate that more than I can explain :)
21:17:02 * diablo_rojo logs into the dev site now to make the project + project groups
21:17:09 <diablo_rojo> projects + project group
21:17:17 <diablo_rojo> wrong placement of plurality
21:17:30 <timburke> cool, thanks!
21:17:46 <diablo_rojo> I'll drop in the swift channel and ping when its done
21:17:52 <timburke> excellent
21:17:52 <diablo_rojo> shouldn't take me too long though
21:17:57 <timburke> #topic container building
21:18:02 <mattoliverau> diablo_rojo: kk, thanks again
21:18:52 <timburke> so we had containers building... and pushing to https://hub.docker.com/r/openstackswift/saio ...
21:19:24 <kota_> yey
21:19:46 <timburke> but currently it's busted. -infra's got a patch to fix it... but it brings up a question i had about this:
21:19:59 <kota_> (no overview available message i see)
21:20:15 <tdasilva> kota_: yeah, we need to fix that
21:20:16 <timburke> should an inability to build and/or publish container prevent us from merging code?
21:21:01 <timburke> (the infra fix is https://review.opendev.org/#/c/659296/ which is all black magic to me)
21:21:01 <patchbot> patch 659296 - opendev/system-config - Pin skopeo to unbreak skopeo+bubblewrap - 3 patch sets
21:21:08 <mattoliverau> oh interesting. I'd say for the -latest image.. no.
21:21:10 <tdasilva> timburke: IMO no, I was wondering the same thing if we should make those jobs non-voting
21:22:21 <mattoliverau> but if/when we get it tied into the release tooling, that should be important.. thought that doesn't prevent code landing. But is when it is a "critical" component
21:23:03 <timburke> tdasilva, i think i'd be ok with that... but yeah, i've got the same thought as mattoliverau about if/when this gets tied into tag creation
21:24:56 <tdasilva> timburke: but even that I'd say can happen out-of-band. Ideally we should be able to go back and build container images for our past releases, no?
21:25:31 <mattoliverau> let's say yes? but to be honest I don't know how the release triggering works.
21:25:41 <tdasilva> I agree it's a super nice to have, but i'm not super concerned about it. the tooling doesn't exist yet, so we can cross that bridge when we get there???
21:25:43 <mattoliverau> But that's something we can donfirm when getting it set up
21:25:53 <tdasilva> right
21:25:54 <mattoliverau> tdasilva: +1
21:26:26 <timburke> yeah, i'd hope so. getting it lined up with an appropriate version of upper-constraints might get interesting (particularly if we want to build images for pre-u-c tags), but surely we can sort it out as a manual thing
21:26:28 <tdasilva> I should have added: the tooling doesn't exist yet AFAIK. -infra said they would work on it after PTG
21:26:55 <timburke> all right, sounds like we're on board with making container stuff non-voting
21:26:58 <tdasilva> timburke: yeah, I was thinking that was the only caveat based on the conversation we had
21:27:17 <timburke> #topic py3
21:27:34 <timburke> test/unit/proxy/ is done!
21:27:50 <timburke> that brings us up to ~83% of unit tests running under py3
21:27:52 <mattoliverau> \o/ nice
21:28:08 <timburke> most of the rest is in obj/ iirc
21:28:22 <timburke> though there are a few straggling middlewares
21:28:59 <timburke> on the func test front, we got some pre-reqs landed, but still no gate job yet
21:29:41 <timburke> can anyone volunteer to review (some or all of) the chain starting at https://review.opendev.org/#/c/642520/ ?
21:29:41 <patchbot> patch 642520 - swift - Get functional/tests.py running under py3 - 16 patch sets
21:31:31 <timburke> :-/
21:31:32 <mattoliverau> I've been distracted with catching up with work stuff. But will try and loop back round to py3 today or tomorrow. So I'll volunteer to take a look, at atleast some
21:31:41 <timburke> thank you mattoliverau!
21:31:57 <tdasilva> timburke: i'll get an environ setup and at least start running tests on those patches
21:32:23 <timburke> thanks tdasilva
21:32:30 <timburke> separately: at the PTG i set a goal of getting all unit and func tests passing by the end of the month. we're now halfway through May; how are people feeling about the feasibility of that goal?
21:33:38 <mattoliverau> depends on the functests. though there are less of them then unit tests.
21:34:02 <mattoliverau> So will reserve my comment until I start reviewing it :P
21:34:10 <timburke> fair enough :-)
21:35:01 <timburke> i knew it was a fairly aggressive goal, so in my head i put the odds at around 50/50 going in
21:35:56 <timburke> i just want to get a feeling for how well i can forecast such things
21:36:31 <mattoliverau> nah it's a good kick in the pants. if we need to push it out fine. but lighting a fire under us I think helps. especually in something like py3, where its tedious
21:37:05 <mattoliverau> wow, my spelling sucks this morning. sorry :)
21:37:28 <timburke> only other thing i've got on py3 is that i'm going to start prodding eventlet about doing another release. when we declare support, i'd really rather not have to tell people "just go grab whatever's on eventlet master" :P
21:38:11 <mattoliverau> good idea. how receptive are they as a community?
21:38:35 * timburke shrugs
21:38:47 <mattoliverau> ie, do we need to start prodding now, last week, or the week before our release :P
21:38:54 <timburke> they take my patches, and seem reasonably responsive
21:39:00 <mattoliverau> cool
21:39:25 <timburke> i figure i ought to start prodding now, so i can *also* start working the openstack/requirements side sooner rather than later
21:40:07 <timburke> #topic lots of small files
21:40:19 <timburke> kota_, alecuyer, rledisez how goes it?
21:40:48 <kota_> now we got the gate running against to losf backend!
21:40:56 <timburke> \o/
21:40:59 <mattoliverau> nice
21:40:59 <alecuyer> yes! :)
21:41:01 <kota_> everything works fine for now.
21:41:24 <timburke> "for now" seems ominous :-)
21:41:31 <kota_> #link https://trello.com/b/xhNxrcLX/losf
21:41:56 <kota_> alecuyer is still digging some dependency issues
21:42:25 <alecuyer> still grpc issues yes
21:42:32 <kota_> and i've started on some docs and... cleanuping.
21:42:52 <alecuyer> the latest version does run functests but it fails with concurrency (real workload). I'll post more details about it
21:43:16 <alecuyer> so I think I should work on moving to HTTP (I started). And, writing unit tests
21:43:45 <kota_> with the work to get the gate job, I found some issues like where we should manage the logs, process naming, etc...
21:44:32 <kota_> so I want to try the works, with stable gate ;)
21:44:42 <kota_> alecuyer: +1
21:45:17 <timburke> that all sounds great! is there anything the rest of us should be doing to help?
21:46:01 <kota_> not so much to me for now, alecuyer do you?
21:46:43 <alecuyer> No, not yet I think
21:47:00 <timburke> all right. keep us updated!
21:47:07 <timburke> #topic open discussion
21:47:16 <timburke> does anyone have anything else to bring up?
21:50:37 <timburke> well, i started to break up https://review.opendev.org/#/c/648263/ into something that might be useful and *not* have the same risk of creating dark data
21:50:38 <patchbot> patch 648263 - swift - WIP: s3api: Make multi-deletes async - 5 patch sets
21:50:46 <timburke> #link https://review.opendev.org/#/c/659162/
21:50:46 <patchbot> patch 659162 - swift - Add operator tool to async-delete some or all obje... - 3 patch sets
21:51:35 <timburke> just as a heads up
21:52:01 <mattoliverau> cool
21:52:10 <timburke> i'd be very interested in alecuyer or rledisez's opinion on it (since i'm claiming it would be useful for an operator)
21:52:36 <alecuyer> I'll look at the patch tomorrow ,
21:52:55 <timburke> thanks alecuyer!
21:52:58 <alecuyer> see if we can try it, and yes it would be useful :)
21:54:35 <timburke> eventually, i'd kinda like to leverage it into a kind of container-reaper, and let clients do something like DELETE /v1/AUTH/test/c?recursive=true
21:54:58 <timburke> well, that's all i've got, and it looks like we're getting to the end of our slot anyway
21:55:10 <timburke> thank you all for coming, and thank you for working on swift!
21:55:15 <alecuyer> thank you
21:55:31 <timburke> #endmeeting