21:01:21 <notmyname> #startmeeting swift
21:01:22 <openstack> Meeting started Wed Dec  5 21:01:21 2018 UTC and is due to finish in 60 minutes.  The chair is notmyname. Information about MeetBot at http://wiki.debian.org/MeetBot.
21:01:23 <notmyname> who's here for the swift team meeting?
21:01:24 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
21:01:27 <openstack> The meeting name has been set to 'swift'
21:01:31 <clayg> o/
21:01:33 <m_kazuhiro> o/
21:01:36 <mattoliverau> o/
21:01:39 <kota_> o/
21:01:44 <tdasilva> hi
21:02:00 <rledisez> hi o/
21:02:06 <notmyname> welcome, all
21:02:24 <notmyname> first off, thanks to mattoliverau for running the last couple of meetings while I was traveling
21:02:43 <mattoliverau> nps, glad your back :)
21:03:35 <notmyname> big stuff going on is still s3api, py3, and losf. so let's cover those and then get to miscellaneous stuff
21:03:42 <notmyname> #topic ongoing work
21:04:13 <notmyname> I know for us at swiftstack, we want the s3api patches landed so we can start using that instead of swift3
21:04:18 <notmyname> #link https://wiki.openstack.org/wiki/Swift/PriorityReviews
21:04:22 <timburke> yes we do :-)
21:04:35 <notmyname> timburke: how do they look? what do you need?
21:05:39 <timburke> should be good, as i recall. just gotta rebase the long-running uploads one to avoid that bad that broke in-process func tests
21:05:52 <notmyname> ok
21:06:37 <notmyname> if you're looking for something to help out on, grabbing one of those patches to review would be very helpful
21:06:51 <timburke> #link https://review.openstack.org/#/dashboard/?S3+API=status:open+file:%255Eswift/common/middleware/s3api/.*+project:openstack/swift
21:07:03 <timburke> (for a broader view of s3api patches)
21:07:07 <notmyname> thanks
21:07:10 <notmyname> I've also seen a lot more py3 patches lately, and it looks like zaitcev may have some extra help
21:07:25 <clayg> timburke: thanks!
21:07:29 <notmyname> he's not here right now. mattoliverau, you've been interested in the py3 work, too. anything to report there?
21:07:57 <mattoliverau> lots of good intention but not as much work as I'd hoped.
21:09:03 <mattoliverau> I'll find some time today to look at some reviews. Maybe py3 while we have zaitcev's convinced others to help ;)
21:09:20 <mattoliverau> wow, my english is bad... you get the idea
21:09:41 <clayg> haha
21:09:50 <notmyname> tdasilva: who is Steap?
21:09:57 * mattoliverau moves from phone to laptop so can use a real keyboard.
21:09:57 <tdasilva> Cyril
21:10:01 <notmyname> great!
21:10:23 <notmyname> welcome Steap!
21:10:37 <notmyname> we were just talking about the ongoing py3 work in swift
21:11:05 <Steap> oh
21:11:09 <Steap> so everyone is sad now?
21:11:15 <notmyname> lol
21:11:23 <notmyname> Steap: it looks like you're helping pete out with py3 work? I've seen a lot of patches lately
21:11:49 <Steap> notmyname: yep, I took a quick look and sent a bunch of patches
21:11:56 <Steap> Tim is helping as well
21:12:03 <Steap> I don't think I can port everything, though
21:12:17 <Steap> Some of the issues are really Swift-specific, and I do not know Swift enough to fix all tests
21:12:39 <notmyname> what's your general strategy? are you taking a module at a time or are you looking at a particular compat area across all modules (eg print/print())?
21:13:15 <timburke> yeah, that's been my concern for a while now -- we need people that know swift making sure that the tests are still testing what we think they're supposed to be testing
21:13:18 <Steap> a bit of both, but mostly the first one
21:13:27 <Steap> we need to fix one test_*.py file
21:13:33 <Steap> add it to tox.ini so that it never gets broken again
21:13:38 <Steap> and do that for all test files
21:13:45 <timburke> yay ratchets!
21:13:48 <Steap> which is what we did with all other OpenStack projects
21:14:03 <notmyname> sounds good
21:14:44 <notmyname> are there specific ways the rest of us can help? should we have some sort of "hey I need some swift-knowledge dropped here" or anything like that?
21:14:48 <Steap> timburke: yeah, it's easy to just break tests so that the y"work" and we do not want that :)
21:15:06 <clayg> if we delete all the tests are we py3 compatible?
21:15:34 <notmyname> :)
21:15:41 <Steap> clayg: :)
21:15:57 <kota_> lol
21:16:09 <timburke> the sooner we can get to running func tests against a py3 server, the better i'll feel
21:16:27 <kota_> tests will be py3 compatible, but... the code...
21:16:37 <clayg> 👍
21:16:46 <notmyname> timburke: yeah
21:17:01 <Steap> timburke: there are still bugs about the func tests?
21:17:39 <notmyname> Steap: ok, for the short-term (ie next 7 days), what is needed to make progress on py3 in swift?
21:17:39 <timburke> well... i know our func tests have a habit of sending non-RFC-compliant request lines...
21:17:50 <timburke> which breaks under py3
21:18:05 <timburke> see https://github.com/python/cpython/pull/7932
21:18:07 <Steap> notmyname: merge everything that has been submitted :)
21:18:13 <tdasilva> timburke: are there any of the services that have been completely ported to py3?
21:18:31 <Steap> notmyname: then, I can probably do a bit more work, but I'll need to team up with someone who knows the specifics of Swift
21:18:49 <Steap> unfortunately, the latest bugs are always the hardest
21:19:07 <timburke> tdasilva: don't think so? i wouldn't feel confident in saying that any have. but i know i can bring up a py3 proxy and have *some* func tests passing
21:20:59 <notmyname> timburke: we had a gist or etherpad of py3 hard problems, but I can't find a link now. do you have it handy? I think it would be good for Steap to see
21:21:24 <timburke> https://gist.github.com/tipabu/833b03a865dba96e9fa2230b82f5d075 but i haven't looked at it in a bit
21:21:44 <notmyname> Steap: from the swift team side, it's been mostly zaitcev and mattoliverau who have said they'll be able to work on it, with some additional help from timburke
21:22:09 <Steap> I'll probably work with zaitcev when he's back next week then :)
21:22:41 <timburke> feel free to reach out if you feel stuck before then, though :-)
21:23:06 <Steap> hehe
21:23:07 <Steap> sure :)
21:23:18 <tdasilva> Steap: rledisez is also in France so he could be a good person to ask for help too
21:23:21 <notmyname> Steap: definitely. and take a look at timburke's gist above for some identified issues. it's a bit dated, but mostly still correct, I think
21:24:03 <mattoliverau> +1, work's kept me busy, but I'll attempt to review py3 patches today.
21:24:17 <timburke> i feel like 9/10 times the worst part of porting is dealing with WSGI strings...
21:24:23 <mattoliverau> (sorry I was off dealing with a naughtly toddler.. fun times)
21:24:47 <timburke> figuring out when they're in use, and where we ought to convert between them and native strings or bytes
21:25:48 <notmyname> ok, I want to move on to a different topic. anything else for py3 right now?
21:26:27 <mattoliverau> Yeah, we should write about that, or some "rules" and put them in the review guidleines or something, so we as reviewers and devs do the right thing in regards to py3 (though I guess we'll have tests)
21:27:12 <notmyname> mattoliverau: yeah, i think just writing it down somewhere will be helpful to reference for everyone as we do the port
21:27:29 <notmyname> Steap: thanks for stopping by. I know it's late for you :-)
21:27:40 <notmyname> rledisez: anything to share about losf this week?
21:27:48 <tdasilva> Steap: thanks for your help!
21:27:48 <Steap> notmyname: you're welcome :)
21:28:04 <rledisez> notmyname: nothing new
21:28:21 <notmyname> rledisez: ok. please give us an update when you've got soemthign :-)
21:28:32 <rledisez> sure!
21:28:35 <notmyname> #topic other things...
21:28:52 <notmyname> let's see.. what else is going on?
21:29:15 <notmyname> https://bugs.launchpad.net/swift/+bug/1806500 has a patch that should be landed soon. it's a bug that affects enabling sharding
21:29:17 <openstack> Launchpad bug 1806500 in OpenStack Object Storage (swift) "sharder fails when drive is unmounted" [High,Confirmed]
21:29:58 <notmyname> any other bugs or patches to mention for people to be aware of?
21:30:19 <notmyname> kota_: clayg: rledisez: anything to bring up?
21:30:44 <clayg> nothing from me, your bug with the sharder is new and people trying to shard have probably hit it
21:30:47 * kota_ was busy in the last couple of weeks so nothing new.
21:31:05 <notmyname> ok, last thing I want to talk about is releases
21:31:17 <kota_> I'll have time to be back to review on some patches in the next week.
21:31:27 <notmyname> it's been a long time since we did an upstream release (august 22)
21:31:28 <rledisez> just this which fix a bug I introduced myself some times ago :)
21:31:28 <rledisez> https://review.openstack.org/#/c/611614/
21:31:29 <patchbot> patch 611614 - swift - Fix SSYNC concurrency on partition - 5 patch sets
21:31:50 <timburke> kota_: think you'll get a chance to revisit https://review.openstack.org/#/c/592231/ this week?
21:31:51 <patchbot> patch 592231 - swift - s3api: Include '-' in S3 ETags of normal SLOs - 6 patch sets
21:32:05 <clayg> rledisez: oh cool, that sounds great!
21:32:11 <notmyname> rledisez: oh! thanks for bringing it up
21:32:19 <kota_> timburke: maybe in the next week. is that ok for you?
21:32:35 <timburke> yeah, should be fine. thanks!
21:33:04 <notmyname> rledisez: if nobody has reviewed that by monday, please start bugging us in irc
21:33:22 <rledisez> notmyname: no need to say it twice, will do ;)
21:33:30 <notmyname> :-)
21:33:56 <notmyname> back to releases... here's the general openstack release schedule for this cycle: https://releases.openstack.org/stein/schedule.html
21:34:11 <rledisez> otherwise, we are more and more thinking about "global" coordination in replication/reconstruction process. maybe based on global task queue. for now, this is just thinking
21:34:30 <notmyname> there isn't any particular deadline looming, but I would like to cut an upstream release when the s3api patches land
21:35:03 <notmyname> and the sooner that happens, the better. just because (1) there's been a lot that's landed and (2) it's been a while since we've tagged a release
21:36:25 <notmyname> does anyone else have any questions or comments for this week's meeting?
21:37:54 <mattoliverau> nope, and a release once the s3api patches land sound good :)
21:39:07 <notmyname> then let's call the meeting done :-)
21:39:21 <notmyname> thank you everyone for coming. and thank you for your contributions to swift
21:39:24 <notmyname> #endmeeting