21:00:47 #startmeeting swift 21:00:48 Meeting started Wed Oct 14 21:00:47 2015 UTC and is due to finish in 60 minutes. The chair is notmyname. Information about MeetBot at http://wiki.debian.org/MeetBot. 21:00:49 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 21:00:52 The meeting name has been set to 'swift' 21:01:00 who's here for the swift meeting? 21:01:04 o/ 21:01:05 o/ 21:01:06 Hello! 21:01:08 hi 21:01:09 here 21:01:10 hi 21:01:10 o/ 21:01:10 o/ 21:01:12 o/ 21:01:16 o/ 21:01:19 o/ 21:01:47 welcome, everyone :-) 21:02:01 o/ 21:02:14 helllo 21:02:22 the summit is quickly approaching 21:02:30 o/ 21:02:31 I'm supposed to push a schedule by the end of the week 21:02:51 we've all been working on https://etherpad.openstack.org/p/tokyo-summit-swift as a place to coordinate topics 21:02:52 #link https://etherpad.openstack.org/p/tokyo-summit-swift 21:02:55 hi 21:03:09 so let's see what's there and rank them so we can figure out what the schedule will look like 21:03:14 #topic summit planning 21:03:17 funny how the summit goes from "long ways off" to "oh crap" in the span of a week 21:03:23 no kidding :-) 21:04:17 here's how I think we should do this: go down the list of topics listed, give each a priority (high medium low, 1..10, somehting). then I'll sort them for scheduling later 21:04:31 and there are some interesting thigns (IMO) at the bottom that aren't yet fleshed out 21:04:51 to recap, we've got 2 fishbowl sessions, 12 workroom sessions, and a meetup all day friday 21:05:03 fishbowls are large rooms 21:05:24 workrooms are smaller and won't have a prominant topic on the printed schedule 21:05:29 hello 21:05:35 the meetup is completely ad hoc 21:05:46 * cutforth regrets getting caught in a hallway conversation 21:06:04 any questions on how the summit will look before we talk about the topics? 21:06:38 notmyname: it seems like the fishbowls might have some contention - be interesting to see how that works out 21:07:04 (also, I think it's definitely possible to schedule continuous sessions for something that needs more time) 21:07:23 I felt that in vancouver the mental context switching every 35 minutes was pretty rough 21:07:38 ok, first up 21:07:38 notmyname: +2 21:07:56 mark seger diagnostics sessions in a workroom 21:08:00 relative priority? 21:08:13 let's do from 1 to 10. 10 is highest 21:08:30 notmyname: can't relatively weight w/o a full list? 21:08:54 ok, on a scale of 1 to 10, how much to you want to have this as a scheduled topic? 21:08:54 :-) 21:08:55 notmyname: can we stack rank or something? 21:09:20 clayg: yeah, I'll do that after we have a number. or if we make it though all of them in this meeting we can do that at the end 21:09:25 notmyname: depends on the synopsis? I feel like I missed a link to a etherpad or something... 21:09:27 * notmyname votes 3 21:09:36 https://etherpad.openstack.org/p/tokyo-summit-swift 21:09:39 sorry i was late ;) 21:09:59 7 21:10:11 how about a 5 ;) 21:10:23 * acoles has been scrolling ahead on etherpad 21:10:24 5 21:10:34 5 21:10:38 jrichli 5 21:10:43 5 21:10:46 ok, 5 21:10:49 5 21:10:54 next up 21:11:01 ops feedback fishbowl session 21:11:03 10 21:11:05 10 21:11:07 10 21:11:07 10 21:11:09 10 21:11:11 10 21:11:11 8 21:11:14 8 21:11:14 10 21:11:15 2.7182818 21:11:16 yeah, that one is obvious :-) 21:11:20 10 21:11:26 10 21:11:33 torgomatic: yep, notmyname did not specify integers 21:11:37 cschwede: just curious, why 8? 21:12:00 and there's a note from someone on it about potentially doing a longer session 21:12:23 torgomatic: it'll be fine until we get a complex vote :) 21:12:39 hehe 21:12:48 notmyname: to be able to vote something different higher, but still ensuring it gets enough points to be included ;) 21:12:56 cschwede: heh ok :-) 21:13:21 seeing as we don't actually have any other proposed fishbowl sessions yet, it's very likely to happen :-) 21:13:25 next up 21:13:31 symlink from hrou 21:13:44 7 21:13:46 8 21:13:47 7 21:13:51 9 21:13:54 7 21:13:57 5 21:14:06 9 21:14:09 8 21:14:11 8 21:14:14 8 21:14:26 9 21:14:29 ok, I'll write down 8 21:14:36 next 21:14:37 hmm, what is cschwede saving his 10 for? :) 21:14:45 production keymaster from jrichli 21:14:57 acoles: next summit ;) 21:15:10 notmyname: 10 21:15:15 9 21:15:18 9 21:15:21 o/ (sorry now in waiting room) 21:15:21 7 21:15:22 9 21:15:25 7 21:15:27 7 21:15:29 8 21:15:32 7 21:15:33 8 21:15:34 9 21:15:58 7 21:16:09 8 or 9, writing down 9 21:16:11 mattolivearu: we are now voting (rating) for each item. 21:16:13 next 21:16:28 remaining encryption items from jrichli 21:16:29 9 21:16:31 9 21:16:32 9 21:16:36 9 21:16:37 9 21:16:39 9 21:16:40 I guessed that thanks kota_ 21:16:41 9 21:16:42 10 21:16:43 9 21:16:47 8 21:16:50 8 21:16:55 would it be easier to just pick the one topic of the 15 listed that we we don't think will fit into the 14 slots we have and will have to bump out to the ad hoc? 21:17:05 Lol 21:17:12 :) 21:17:12 ok, 9 as well. and these should be scheduled together 21:17:19 clayg: too easy! 21:17:20 next 21:17:26 global cluster improvement from kota_ 21:17:29 10 21:17:32 9 21:17:32 9 21:17:35 10 21:17:39 9 21:17:42 10 21:17:43 7 21:17:44 9 21:17:46 9 21:17:50 7 21:17:50 9 21:18:03 clayg: been thinking a lot about that pigeonhole principle lately, haven't you :P 21:18:04 Everyone votes the same each time 21:18:11 we have 12 slots, actually. and we may choose to let one topic take more than one slot. or maybe "this afternoon we're covering all these things" 21:18:18 anchoring 21:18:57 next up 21:19:06 storage tiering discussion from m_kazuhiro 21:19:13 7 21:19:16 6 21:19:17 7 21:19:17 10 21:19:19 10 21:19:19 9 21:19:21 10 21:19:23 6 21:19:26 9 21:19:26 7 21:19:29 8 21:19:31 2 21:19:32 more of a split on this one :-) 21:19:38 6 21:20:04 * notmyname writes down 7 21:20:11 next 21:20:24 role-based access control with oslo.policy from ho 21:20:27 8 21:20:29 500 21:20:33 8 21:20:33 lol 21:20:37 briancline: :-) 21:20:38 i mean 10 21:20:41 10 21:20:43 typo ;-) 21:20:44 9 21:20:47 8 21:20:48 briancline: lol 21:20:48 * torgomatic knows some of those words 21:21:02 10 21:21:04 8 21:21:10 8 21:21:17 i'll go with the status quo 21:21:39 lots of 8s and a 500 which makes the average pretty close to 475 21:21:45 :D 21:21:46 \o/ 21:21:51 yes! 21:21:51 lol 21:21:54 notmyname: how is 2 + 12 != 14? 21:21:55 We can vote that high! Awesome 21:22:02 ho: congratulations! you've won 21:22:11 clayg: just separating fishbowl from workroom. 21:22:13 hehehe 21:22:14 clayg: hex 21:22:27 next up 21:22:34 clayg: no wait, something else 21:22:36 using keystone session object in swiftclient 21:22:36 duh 21:22:50 acoles: octal 21:22:51 5 21:23:04 notmyname: depends if anyone can tell us how 21:23:06 what's a session object 21:23:11 there you go 21:23:12 acoles: lol 21:23:12 Don't know what that really is.. So.. 6? 21:23:15 clayg: :D 21:23:19 6 21:23:23 keystone has an object thing that does a lot of the auth validation 21:23:27 retries and all that 21:23:29 heh - maybe that means we *should* talk about it? 21:23:32 6.5 ish 21:23:36 including keeping state between requests 21:23:39 Ahh OK 7 then :p 21:23:59 7 21:24:04 notmyname: it doesn't have a facilitator ( joeljwright ?) 21:24:08 it's supposed to make things a lot better if you're using keystone. also, I think everyone else is doing it (or nearly), so if we don't they won't think we're one of the cool kids 21:24:16 clayg: not gonna be in Tokyo :( 21:24:18 can we discuss the keystone v4 as well then (grouping keystone topics)? 21:24:23 notmyname: so I am 10 for it happening but not sure how far we would get in a session 21:24:25 clayg: bad news is that joeljwright and timburke aren't in tokyop 21:24:30 acoles: yeah 21:24:32 cschwede: v4?! 21:24:34 digi-joel 21:24:44 it's like "Keystone for Dummies^WSwift Developers" 21:24:51 acoles: https://review.openstack.org/#/c/181393 21:24:54 lol 21:25:11 i said v4 because it will break current clients 21:25:17 cschwede: oh that, right, thanks 21:25:30 cschwede: we'll just do that one over sake 21:25:41 \o/ 21:25:44 ... and tears 21:25:49 :-) 21:25:55 notmyname: sounds like a plan! :) 21:25:56 ok, next up 21:26:18 swiftclient docs: improve the docs there, including making sure people know which part of swiftclient to use 21:26:35 +2*32 21:26:43 ^ that's as big as I go 21:26:47 Alex should be at this summit so I'll drag her along 21:26:49 clayg: but even though he's not going to be there, joeljwright helped by adding some of these swiftclient topics 21:26:55 mattoliverau: great 21:26:57 8.5 21:27:03 9 21:27:07 7 21:27:13 9 21:27:35 notmyname: that's because joeljwright is gentleman and a scholar 21:27:57 ...and has impeccable taste in clothes 21:27:57 briancline: but it's... it's... DOCS! 21:28:03 around...7? 21:28:05 heh 21:28:13 there's a start here https://review.openstack.org/#/c/223319/ 21:28:20 ok, next up 21:28:35 oh yeah, speaking of swiftclient stuff, clayg are we at that point where i should nag you about how it's almost the summit and the client still isn't (that much) better? 21:28:38 clayg: i almost had a 5 because of other topics, but i said the same thing in my head 21:28:46 container sync from eranrom (discuss patches) 21:28:53 timburke: YUP! 21:28:58 9 21:29:01 9 21:29:02 9 21:29:06 9 21:29:13 9 21:29:14 9 21:29:15 briancline: likes container sync better than docs ;-) 21:29:21 Lol 21:29:24 well.... idk... like 6 or 7 21:29:24 8 21:29:33 maybe 8, probably 7 21:29:35 7 21:29:42 7 21:29:44 notmyname: that's how you can tell i'm a corporate drone 21:29:48 lol 21:29:49 it's a lot - that's the point 21:30:12 9s and 7s. I'll write down 8 21:30:19 next up 21:30:33 EC topics from acoles (and I hope others too) 21:30:36 1 21:30:38 0 21:30:40 10 21:30:41 ROFL 21:30:42 lol 21:30:44 lol 21:30:44 2**32 21:30:50 10 21:30:50 100 divided by 10 21:30:53 10 21:31:00 9 21:31:03 9 21:31:03 9 21:31:08 9 21:31:17 Lower is better right? :p 21:31:22 8 21:31:35 All of this has been golf scoring actually 21:31:37 mattoliverau: in your hemisphere maybe ;) 21:31:42 next up 21:31:44 lol 21:31:46 ROFL 21:31:47 container sharding 21:31:49 10 21:31:55 9 21:31:58 10 21:32:01 minwoob: you can't say 10 to everythign ;-) 21:32:01 10 21:32:03 9 21:32:04 10! 21:32:06 9¾ 21:32:08 9 21:32:09 8 21:32:12 9 21:32:16 9 21:32:20 notmyname: I've only said it twice! :) 21:32:28 8 21:32:44 This is the summit where we go back to poc #1 :p 21:33:06 next up 21:33:12 pipeline improvements/auditing 21:33:16 9 21:33:22 n/a 21:33:36 I'd be fine with this one on friday 21:33:44 oh - proxy pipelines 21:33:50 That's what I said in the ehterpad :) 21:33:52 (although I think it's a great improvement!) 21:34:04 All pipelines really but proxy is the main 21:34:06 mattoliverau: oh yeah! 21:34:27 i like this but...around 8? 21:34:28 idk... 6 21:34:41 7 21:34:42 i'm not sure everyone wants that. 21:34:44 8 21:34:48 priority as "friday" 21:34:50 seems like there are more important topics? 21:35:00 10 - cause I put it in, but was more thinking of making a group interested on friday 21:35:04 next is the same 21:35:10 swift3 compatibility 21:35:30 9 21:35:33 i like to this priority as friday, too :-) 21:35:43 notmyname: idk, if timburke was going to be there I might try to rib him into talking about s3_compat repo and progress 21:36:10 5 21:36:16 8 (just don't do it while I'm involved in pipelines) 21:36:25 I want to leave this one as friday for now, until we see what else we have 21:36:27 friday 21:36:36 ok, next up 21:36:39 Yeah friday 21:36:55 anti-myth replica part placer from clayg 21:37:02 10 21:37:04 10, restoring! great! 21:37:18 10 21:37:21 10 21:37:21 1/x 21:37:30 1/0 21:37:40 10 21:37:42 10 21:37:46 ok, 10. I got it :-) 21:37:47 acoles: ^ yeah i think i'll get rid of most of those by toyko 21:37:47 oh, can we use fish numbers? 21:37:51 Going into doctors now 21:37:53 10 because graphs and charts 21:38:12 blmartin_: probably mostly gifs 21:38:17 Noooooooooooo 21:38:18 cats? 21:38:25 ok, those are the ones that were fleshed out. next up are general ideas, some need to be expanded 21:38:26 blmartin_: of c... ^ exactly 21:38:29 eg this one 21:38:34 hummingbird update 21:38:54 or combine with hummingbird sync protocol unification 21:38:58 redbo_: dfg: hurricanerix: nadeem_: do eet! 21:39:12 mmmmmm... get the syncs in sync - so sexy 21:39:14 oops, by 1/x i meant something that graphs very differently. fuzzy brain 21:39:44 * notmyname wonders if any of those people are online 21:39:47 i defer to dfg & redbo =) 21:39:55 hurricanerix: no YOU do it 21:39:57 hurricanerix: will you be there? 21:40:00 DEW EET! 21:40:01 notmyname: yp 21:40:04 yup 21:40:07 yay! 21:40:08 ok, done 21:40:08 i'd actually really like to hear about the progress on that 21:40:10 2**32 21:40:14 We can do that. There'll be some stuff in our talk, but we have a lot of data we can share. 21:40:20 9 21:40:30 9 21:40:41 hello? 21:40:48 9 21:40:52 redbo_: I might watch the video of the business track session - but I'm not going to participate if I can be obnoxious and ask questions 21:40:53 dfg_: just say 10 ;) 21:41:13 9 21:41:14 I assume you meant "can't" :) 21:41:18 redbo_: dfg_: yeah, somehting not a presentation but figuring out how we can unify some of the backend stuff from hummingbird and listen to some of the progress there 21:41:37 10 21:41:43 dfg_: oh good :-) 21:41:44 dfg_: good job 21:41:45 nadeem told me to write 10. i trust him 21:41:50 lol 21:41:58 lol 21:41:58 dfg_: srly, nadeem_ is whicked smart 21:41:58 as oppose to the rest of you jokers 21:42:00 dfg_: I'll expect the rough draft first thing monday 21:42:07 ;-) 21:42:20 ok, next one 21:42:26 chaning policy on an existing container 21:42:31