21:00:07 #startmeeting swift 21:00:08 Meeting started Wed Apr 26 21:00:07 2017 UTC and is due to finish in 60 minutes. The chair is notmyname. Information about MeetBot at http://wiki.debian.org/MeetBot. 21:00:09 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 21:00:12 The meeting name has been set to 'swift' 21:00:14 who's here for the swift team meeting? 21:00:17 o/ 21:00:18 o/ 21:00:20 o/ 21:00:20 o/ 21:00:22 o/ 21:00:29 here 21:00:31 o/ 21:00:44 hi 21:00:45 hi 21:00:46 o/ 21:01:04 hi 21:01:21 clayg_: clayg: ping 21:01:42 tdasilva: ping 21:01:48 ping 'em all 21:02:08 yeah, I've thought about a "meeting ping" function for patchbot 21:02:15 oh dar 21:02:18 *dear 21:02:20 hi 21:02:26 tdasilva: HI! 21:02:33 clayg_: !!!! 21:02:59 ok, let's get started 21:03:02 #link https://wiki.openstack.org/wiki/Meetings/Swift 21:03:06 rough agenda there 21:03:13 but summit stuff first 21:03:19 #topic summit update and FYIs 21:03:33 only a week and a half until the summit in boston 21:03:49 good news is that mattoliverau found sponsorship so will be able to make it 21:03:50 oh dear :'( 21:03:51 \o/ 21:03:57 yay! 21:03:59 :D 21:04:10 thanks to the OpenStack foundation 21:04:11 congrats 21:04:21 bad news is that some others ( jrichli, pdardeau_, others?) won't be there 21:04:36 notmyname: i'll be there (booth duty) 21:04:44 pdardeau_: awesome! 21:04:52 nice! 21:04:55 that's new info! 21:05:07 * notmyname backspaces over american english idioms 21:06:04 so unfortunately there's been staffing changes at some of our employers (rackspace, intel). so (1) if someone unfirtunately has to leave our community because of a new job, let's make it easy and support them 21:06:28 Or just hire me to work on Swift ;) 21:06:31 and (2) if you know people who are hiring great openstack devs, people like mattoliverau will be at the summit and looking 21:07:29 pdardeau_: I'm not sure if you can or are ready to share anything about you and intel. last I heard your work with osic is over but there might be new opportunities elsewhere at intel 21:07:45 but ... i don't want to belabor the point or put people on the spot 21:08:18 the point is, as people go from time to time, let's recognize and thank them for their awesome work, and help them transitiona nd carry on their work in teh community 21:08:18 notmyname: it's complicated 21:08:42 pdardeau_: I'm sure. please let us know if there's anything the rest of us in teh community can do to help 21:08:51 thx 21:08:59 nice 21:09:18 and also, at the summit there will be some great opportunities to encourage new contributors to swift! 21:09:50 there's both the upstream institute training (on sunday? I think tdasilva is helping out there?) and there's a project onboarding later in the week 21:10:19 I was in the middle of getting the onboarding schedule squared away (ie finding where/when it is) as the meeting started 21:10:25 Half day Saturday and all of Sunday :) 21:10:37 diablo_rojo_phon: thanks (upstream institute) 21:10:51 I'll be there Sunday 21:10:55 tdasilva: any more details on the upstream institute? 21:11:13 I know it's at the Sheraton from 9am - 6pm 21:11:21 but that's all i know 21:11:24 ok 21:11:39 oh! I wanted to say this too... 21:11:54 onboarding is https://www.openstack.org/summit/boston-2017/summit-schedule/events/18714/swift-project-onboarding -- wed, 1:50-3:20pm 21:11:54 hopefully there will be newcomers interested in swift and I won't scare them away ;) 21:12:02 timburke: thanks 21:12:27 clayg_: thanks for leading the meeting last week! when I got back online I had a message saying "clayg led the meeting and it was great!" 21:12:31 heard that from others too 21:12:37 clayg_: thank you. it was very helpful 21:12:42 np boss 21:12:46 notmyname: what details do you want about upstream institute? :) 21:13:18 I was really hoping to have something on the ML last week for me to digest async - so I could ask questions this week - i must have missed it tho :'( 21:13:22 diablo_rojo_phon: I had a follow-up from last week's meeting (led by clayg_; i was out) saying there was some question about interest and our team's participation 21:13:31 diablo_rojo_phon: and I think you and tdasilva have talked? 21:13:39 the onboarding session clashes with the discussion about adding ids to logging 21:14:09 tdasilva maybe you talked to ildikov ? We just wanted some representation there in case people have swift questions. 21:14:09 acoles: ah. definitely will need to split up then 21:14:14 I only talked to jungleboyj 21:14:23 Ah. 21:14:34 o/ 21:14:40 acoles: I'm going through the summit schedule to look for those sorts of schedule conflicts now 21:14:47 Hey sorry, forgot to join up. 21:15:06 ok. upstream institute. jungleboyj diablo_rojo_phon: is there anything you need from us at this point? 21:15:32 diablo_rojo_phon: Yeah, I had talked to tdasilva . He will join up. 21:15:34 jungleboyj: was there a ML thread this past week? 21:15:37 just someone to be around on Sunday :) 21:15:46 In case people have swift questions 21:16:01 tdasilva will be our (awesome as always) rep 21:16:03 notmyname: Since tdasilva had volunteered and was in town I didn't go to the ML. 21:16:09 jungleboyj: ah ok 21:16:17 notmyname: Hope that was ok. 21:16:26 great, then. I'm looking forward to meeting up sunday night to see how things went 21:16:27 Thank you in advance tdasilva :) 21:16:39 notmyname: Sounds good! 21:16:45 diablo_rojo_phon: yw 21:16:55 looking forward to it 21:17:15 for everyone, the summit schedule is online at https://www.openstack.org/summit/boston-2017/summit-schedule/ so be sure to mark the stuff you want to be at 21:17:20 this includes all the forum tracks too 21:17:48 a brief work on the swift forum sessions 21:18:12 first, thank you to everyone who added those comments to the topics about "why this session needs to happen" 21:18:33 what that's resulted in is a large block of time on wednesday (IIRC) in one room for swift 21:18:58 there are format topics and time slots scheduled, but I want to do it similarly to some past summits 21:19:27 at the beginning of that time block, we'll write down the scheduled (and unscheduled) topics that need to be covered, and we'll have 4 hours or so to get through them 21:20:13 so they originally scheduled us for a few more time slots on different days, but they realized they gave swift a lot of time slots (this is because of the topic comments!) 21:20:40 oohh..i did notice that some had disappeared 21:20:47 so they were going to take some away to give to other projects, and I was able to give up the "time islands" ones and ensure we got the large contiguous block of time 21:20:59 nice! 21:21:04 ie instead of losing the big block of time 21:21:24 make sense 21:21:24 so the unscheduled topics I mentioned earlier.. those are the things that got dropped 21:21:49 so I do definitely want to get through most of the stuff we had on the etherpad. it will be fast. it will be hard to do. but we can do it 21:22:11 I'll be working on that schedule before next week's meeting, and I plan to have more details then 21:22:17 * notmyname is just starting on it now 21:22:35 ok, so before we move on to other topics, are there any questions about the summit? 21:22:37 plus there will be the hallway and beer track right? 21:22:43 mattoliverau: of course! 21:22:51 Mmmmm beer 21:23:01 mattoliverau: you can sleep on the plane ride home 21:23:06 i've heard about this hacking rooms? 21:23:08 after my last week, I'll need it ;) 21:23:14 notmyname: that is true 21:23:23 clayg_: ah right! I just saw something about that 21:23:40 AIUI the "hacking rooms" are like the reservable rooms at the PTG? 21:23:41 hacking on swift can be a somewhat raucous affair - hopefully we don't upset anyone 21:23:45 that's what it sounded like to me? 21:24:07 clayg_: but if you understood something different, do tell 21:24:16 well find out when we get there - i'm keeping an open mind 21:24:39 any other questions on the summit? 21:24:56 I'll try to make space for the work that we need to do within reason - if we really can't make it work w/o interrupting other people or the general flow - I'll give that feedback to the foundation folks 21:25:59 oh, maybe our time block is tuesday (like I said, I'm just now looking at it) 21:26:02 ok, moving on... 21:26:09 #topic swift 2.14.0 release 21:26:33 again, thanks for the prioritization and triage lists for the bugs and patches that were done last week 21:26:50 the big stuff landed, there's great stuff on master that's unreleased, let's release! 21:27:16 the authors/changelog update should be making its way through the gate right now 21:27:24 when that lands, i'll do the patch to request the release 21:27:33 thank you for your reviews and work on this release 21:27:43 #info swift 2.14.0: the best one yet! 21:28:02 but you know what's better than swift 2.14.0? 21:28:20 swift 2.15.0 21:28:22 2.14.1! 21:28:30 :D 21:28:32 timburke: yep! 21:28:44 notmyname: is there an openstack release coming up? or did the new scheduling fix it so the next openstack release isn't right on top of the forum? 21:28:53 i'm just going to assume we'll need a minor bugfix ;-) 21:29:00 clayg_: the latter. no openstack release coming up 21:29:02 ... and it's Pike right? we're working on Pike? 21:29:07 yup 21:29:13 timburke: and taht would make it better :) 21:29:20 https://wiki.openstack.org/wiki/Swift/PriorityReviews has some high priority bugs on it 21:29:58 as a reminder, there's a ton of stuff people are interested in working on (in addition to the bugs) 21:30:09 some of them are more "improvements for existing stuff", dunno - i was hoping to get more feedback on that - but maybe I accidently picked the right stuff? 21:31:04 the two most important things in swift are container sharding and improving rebalance. if you're not looking at bugs listed on priority reviews, you can be most helpful by helping with these 21:31:39 oh yeah, I should give sharding some love before I loose my clouds. 21:31:50 :'( 21:31:57 mattoliverau is obviously the world's leading expert on container sharding, but timburke and others have been learning in case we lose matt 21:32:30 if you want to help out on container sharding, the best way is to help improve test coverage. (1) it needs more tests (2) it will help you learn it 21:32:32 timburke and acoles are awesome, just saying 21:33:10 one of the things i've been meaning to do for a while is to make a probe test harness... 21:33:30 if you want to help out with improving rebalance, clayg_ is your starting point. and https://etherpad.openstack.org/p/swift-rebalance is a rough guide of current thinking 21:33:40 and once you learn it you can make it better. fresh eyes and all that. 21:34:12 notmyname: there's also zaitcev's patch to fix PUTs - which is related to golang 21:34:40 ah, right. yeah, you and I need to write up a consolidated view of the different parts of that work 21:34:58 typey typey 21:35:02 yup 21:35:12 or in clayg_'s case, pencil and notepad ;-) 21:35:18 werd 21:35:28 oh, and looking back at the priority reviews: should i go ahead and squash in my fixups for https://review.openstack.org/#/c/302494/ ? i think i'd be ready to +2 with that... 21:35:29 patch 302494 - swift - Sync metadata in 'rsync_then_merge' in db_replicator 21:35:38 ok, any questions on all that? the release, the big things being worked on? 21:35:46 i feel bad that it had languished for like a year before i looked at it :-( 21:36:07 :-( 21:36:36 timburke: IMO, yes, squash changes 21:36:48 kota_: is daisuke still around to help out on that? 21:36:53 do you know? 21:37:08 notmyname: i don't think so 21:37:11 ok 21:37:13 mattoliverau: for the record, thank you so much for your incredible work on sharding, we are indebted to you 21:37:21 yeah, I only occasionally see him in irc any more 21:37:22 indeed! 21:37:24 acoles: +1 21:37:29 acoles: mattoliverau: +1 21:37:37 notmyname: i just talked him yesterday but he said, not able to do it 21:37:40 says the people who are yet to run it :P 21:37:40 acoles: +1 21:37:44 timburke: so yeah, squash away 21:37:49 YW 21:37:50 kota_: ok 21:38:11 notmyname, timburke: i'll try to look at the patch to help something 21:38:16 kota_: thanks 21:38:23 thanks kota_ 21:38:36 #topic follow-up: patch 456205 21:38:37 https://review.openstack.org/#/c/456205/ - python-swiftclient - WIP: Use keystone session when possible 21:38:53 it was on the agenda as a follow-up from last week 21:39:11 but the patch doesn't look like it's moved much/any and lots of tests fail and it's WIP 21:39:13 so.... 21:39:14 ? 21:39:27 yeah, so I've been looking at that patch… I think it highlights how big the change to generic keystone client would be… 21:39:45 I don't think jaosorior has time to work on it 21:40:04 i lost track of the link to the related/earlier attempt/change 21:40:11 but my understanding is that this will be required to use keystone in the new devstack world order? 21:40:57 oh wow 21:41:03 notmyname: keystone certainly want the generic stuff to be used and not the version specific clients 21:41:55 joeljwright: what do we need to do here, then? 21:42:31 it's a big change 21:43:07 the basic idea is in that patch, but it't a lot of work to move to keystone sessions 21:43:18 and it changes some of the client behaviour 21:43:32 we need to do it 21:43:37 but it will take a while 21:43:40 ok 21:44:01 could you outline what's needed with any details you know and link to it on the ideas page? 21:44:16 I think that would help the rest of us keep track over the weeks as we go 21:44:25 or trello or whatever you like 21:45:01 sure, I'll write down what I understand the situation to be 21:45:12 perfect. thanks 21:45:19 #topic open discussion 21:45:27 what else do we need to talk about this week in the meeting? 21:46:46 obviously not much :) 21:46:56 careful... we're in danger if ending 10 minutes before last week's meeting 21:47:08 might make clayg_ feel bad if we end so early ;-) 21:47:15 nah i'm good 21:47:20 heh, ok :-) 21:47:38 thank you everyone for your work on the code and on the community. you are why swift is great 21:47:46 I'm looking forward to seeing (most of) you in boston 21:47:50 until next week.... 21:47:54 have fun peeps 21:47:58 \o/ I can finish breakfast 21:47:59 ...or 30 seconds in #openstack-swift 21:48:03 #endmeeting