21:00:25 <timburke> #startmeeting swift
21:00:26 <openstack> Meeting started Wed Mar 20 21:00:25 2019 UTC and is due to finish in 60 minutes.  The chair is timburke. Information about MeetBot at http://wiki.debian.org/MeetBot.
21:00:27 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
21:00:29 <openstack> The meeting name has been set to 'swift'
21:00:38 <timburke> who's here for the swift meeting?
21:00:41 <zaitcev> o/
21:00:55 <mattoliverau> o/
21:01:05 <rledisez> o/
21:01:25 <clayg> o/
21:01:35 <tdasilva> hi
21:01:53 <mattoliverau> we know that kota_ and notmyname wont be here
21:01:56 <timburke> so we've got a pretty short agenda; perfect for me to try my hand running a meeting :-)
21:02:02 <timburke> #link https://wiki.openstack.org/wiki/Meetings/Swift
21:02:14 <timburke> first up: release!
21:02:23 <timburke> #topic Swift release
21:02:42 <timburke> this'll be our last release for Stein
21:03:12 <timburke> i put a handful of patches that i think would be good to get in on the priority reviews page
21:03:15 <timburke> #link https://wiki.openstack.org/wiki/Swift/PriorityReviews
21:03:23 <timburke> and i think clayg has already approved one!
21:03:56 <timburke> https://review.openstack.org/#/c/633671/ seems like it must be hurting zigo; i'd really like to make him whole
21:03:57 <patchbot> patch 633671 - swift - Fix decryption for broken objects - 8 patch sets
21:04:02 <clayg> the easy one
21:04:22 <timburke> https://review.openstack.org/#/c/634507/ was tangentially related to that
21:04:22 <patchbot> patch 634507 - swift - encryption: Be more tolerant of errors while decry... - 2 patch sets
21:04:45 <timburke> and https://review.openstack.org/#/c/571905/ fixes a bunch of bugs and already has a +2
21:04:46 <patchbot> patch 571905 - swift - Fix quoting for large objects - 5 patch sets
21:05:12 <timburke> beyond that, we've got a bunch of great stuff currently working its way through the gate
21:05:24 <clayg> OMG p 571905 scares me - do we ALWAYS need to quote when we make a subrequest with a path built by hand like that... probably!?
21:05:24 <patchbot> https://review.openstack.org/#/c/571905/ - swift - Fix quoting for large objects - 5 patch sets
21:05:47 <timburke> clayg, 99% sure, yeah
21:06:08 <timburke> and once it all lands, we'll tag a release! unless anyone has patches to advocate for that i'm not thinking of
21:06:42 <timburke> so... anybody got patches that should land ahead of a release?
21:06:46 <clayg> how long do we have to get that all landed?  you and kota fixed up p 637662 for me
21:06:47 <patchbot> https://review.openstack.org/#/c/637662/ - swift - Simplify empty suffix handling (MERGED) - 4 patch sets
21:07:14 <timburke> yeah, i was going to add that, then i realized i could just approve it ;-)
21:07:19 <clayg> but you know p 644691 hasn't made sufficient sacrifice to the zuul
21:07:19 <patchbot> https://review.openstack.org/#/c/644691/ - swift - Remove uncalled function - 1 patch set
21:08:18 <timburke> all right, i'll assume the current patchlist is pretty good then
21:08:38 <mattoliverau> I'd have to look at other patches, but they look good.
21:08:39 <timburke> feel free to message me later to know that i'm wrong ;-)
21:08:58 <timburke> next up, updates!
21:09:07 <timburke> #topic losf
21:09:12 <clayg> @timburke wait... how long do we have?
21:09:44 <timburke> oh, right: final deadline is end of the week next week, but with notmyname out of town next week, i'm treating friday as deadline
21:09:47 <clayg> @timburke do we have to land everything on the page?  oh... no just that top catagory - ok
21:10:03 <clayg> 👍
21:10:50 <timburke> hmm... with kota_ and alecuyer not here... i guess i'll let rledisez offer losf updates?
21:11:20 <rledisez> timburke: i'm not completly aware of everything, but I know alecuyer was working on tests
21:11:32 <rledisez> he wrote an ansible playbook for zuul
21:11:42 <rledisez> and I can't really say more
21:11:43 <timburke> https://review.openstack.org/#/c/644879/
21:11:44 <patchbot> patch 644879 - swift (feature/losf) - WIP - start working on ansible playbooks for LOSF - 1 patch set
21:12:09 <timburke> cool! love that we're moving toward tests running with it in the gate
21:12:41 <timburke> only other agenda item was py3 updates
21:12:44 <timburke> #topic py3
21:13:20 <timburke> i've seen a bunch of stuff from zaitcev and mattoliverau lately (feel free to offer your own updates)
21:13:40 <timburke> we've got a (nonvoting) py37 unit test job now
21:13:54 <zaitcev> I don't know, I was away for a week on customer cases. Didn't do a thing for py3 until today.
21:14:08 <mattoliverau> I've done some reviews and attempting to write a py3 port patch: https://review.openstack.org/#/c/643855/
21:14:08 <patchbot> patch 643855 - swift - py3: port staticweb middleware - 4 patch sets
21:14:28 <timburke> and once one of the patches in the recheck-cycle pands, py37 will even pass!
21:14:29 <mattoliverau> probably still needs some work
21:14:40 <mattoliverau> timburke: \o/
21:15:09 <timburke> mattoliverau, yeah, i'm hoping to look at it today -- i'm wondering if the func test failure may actually be legit
21:15:52 <mattoliverau> oh yeah, I should probably look more closely at the func tests. I'll poke today.
21:16:03 <timburke> i'll work on getting https://review.openstack.org/#/c/642520/ rebased, too, and hopefully by next week i'll have a follow-up to run that test in the gate
21:16:04 <patchbot> patch 642520 - swift - Get functional/tests.py running under py3 - 8 patch sets
21:16:29 <mattoliverau> nice
21:16:49 <mattoliverau> getting func tests would make the py3 stuff feel more stable :)
21:17:02 <timburke> i think that's about all i've got for py3
21:17:09 <timburke> #topic open discussion
21:17:28 <clayg> I think this was the most well run meeting evar - you're hired
21:17:35 <zaitcev> I'm actually looking at mattoliverau's staticweb thing... It's getting a 412 in what I presume a py2 environment, which is really strange.
21:17:57 <timburke> clayg, glad you approve :-) also, too late if you *don't* like it :P
21:18:18 <mattoliverau> zaitcev: I could also have done something wrong, /me is learning this crazy py3 stuff ;)
21:18:48 <mattoliverau> we now know that it's notmyname and kota_ that slow down the meeting ;P
21:20:06 <timburke> so, i know https://review.openstack.org/#/c/601950/ has been going a while; zaitcev has a +2, rledisez still seemed to have some questions... i wonder if that would be a good thing to try to get in after the release
21:20:06 <patchbot> patch 601950 - swift - Enable to configure object-expirer in object-serve... - 11 patch sets
21:20:24 <timburke> i feel like we've been talking about it forever
21:20:52 <zaitcev> It's Kazuhiro's thing, isn't it. Seems fine, although it's only a stepping stone to the distributed expirer, if I remember right.
21:21:04 <zaitcev> Oh, BTW
21:21:05 <clayg> whoa - after the release - that's like next week business
21:21:39 <timburke> right. in part, i want to make sure that he can move forward on that; the general task queue seems like a useful thing
21:21:49 <zaitcev> Did you guys see what I dregged out from my pile of old patches - patch 70513
21:21:49 <patchbot> https://review.openstack.org/#/c/70513/ - swift - Add Guru Meditation - 3 patch sets
21:21:57 <timburke> i did just notice that!
21:22:12 <timburke> seems interesting
21:22:17 <zaitcev> Imagine a patch so old that creiht reviewed it
21:22:43 <timburke> notmyname just lol'ed
21:22:53 <timburke> literally; he's over my shoulder now
21:23:11 <zaitcev> And we never merged it, because Swift almost never hungs, so there is no need for a tool to diagnose hungs.
21:24:05 <timburke> ...then torgomatic needed to write https://github.com/swiftstack/python-stack-xray ...
21:24:05 <rledisez> zaitcev: so, it can help when process are hanging? if so, i'm interrested
21:24:42 <zaitcev> But this week I had a customer case come in with updaters hanging for no reason. I suspect a bug in eventlet in case when os is not monkey-patched, so we fork while holding eventlet's internal pipes, and then the forked processes do something to each other over that pipe. Or something. Either way it hungs.
21:25:01 <clayg> idk, lp bug #1820893 sure feels like a hung replication object-server - allthough all signs just point to a forever growing backlog of REPLICATE requests 😞
21:25:02 <openstack> Launchpad bug 1820893 in OpenStack Object Storage (swift) "Unable to limit number of REPLICATE requests per drive" [Undecided,New] https://launchpad.net/bugs/1820893
21:25:50 * clayg is trying to think if he's ever seen hung object-updaters...
21:26:02 <zaitcev> Ouch, I wish I knew python-stack-xray was a thing.
21:26:43 <timburke> it takes a bit to learn how to read the stacks, but it's been super-helpful for me before
21:27:53 <timburke> i'll try to take a look at the meditation patch, at least get an idea of what it's doing
21:28:01 <zaitcev> If you have a moment.
21:28:10 <timburke> anybody else have anything to bring up?
21:30:17 <timburke> all right, i'm'a call it
21:30:39 <timburke> thank you all for coming, and thank you for working on swift!
21:30:47 <timburke> #endmeeting