14:00:00 #startmeeting glance 14:00:01 Meeting started Thu Aug 18 14:00:00 2016 UTC and is due to finish in 60 minutes. The chair is nikhil. Information about MeetBot at http://wiki.debian.org/MeetBot. 14:00:02 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 14:00:04 The meeting name has been set to 'glance' 14:00:09 o/ 14:00:09 #topic roll call 14:00:12 \o 14:00:14 hi bunting 14:00:22 nikhil: Hello! 14:00:24 o/ 14:00:26 o/ 14:00:33 Hello all. 14:00:36 o/ 14:00:39 welcome all 14:00:45 We've a short agenda today. 14:00:47 o/ 14:00:49 let's get started 14:00:54 #topic agenda 14:00:56 #link https://etherpad.openstack.org/p/glance-team-meeting-agenda 14:01:18 if any one needs to add one or two items, we still have space/time 14:01:33 but you can choose to do so in open discussion as well 14:01:45 #topic Community Images Updates 14:02:21 Yesterday, the champions/liaisons of this features met on #openstack-glance and determined the state 14:02:43 from the looks of it, the implementation is looking to be a tad more complicated than it was expected 14:02:57 but nothing we can't achieve in the next couple of months 14:03:31 also, there's this issue of making sure client release matches with the release of the API as we will expect API impact 14:03:51 client behavior needs to be adjusted according to what's required and what's preferred 14:03:59 but all of that will be possible when API is ready 14:04:17 Plan of action on this feature is being updated at: 14:04:19 #link https://etherpad.openstack.org/p/community-images-work-items 14:04:38 The tentative timeline is API targetted just before the summit 14:04:49 nikhil: so expectation is that it will be ready to merge when Ocata opens? 14:04:56 and client, tests outside source tree and docs for o-2 14:05:31 Jokke_: not sure, we've about 6 weeks for Ocata to officially start 14:05:38 but ocata opens much before 14:05:50 i.e. right when n-3 is cut (which is in 2 weeks) 14:06:06 it will not be done in 2 weeks. 14:06:08 so, I don't think API will be ready in 2 weeks 14:06:15 Ocata opens when RC1 is cut, not N-3, right 14:06:17 ? 14:06:58 Jokke_: I think the stable/newton is cut at n-3 and backports are done starting/in rc-1 too 14:07:26 it will be hard to get this in even if we get a FFE 14:07:35 oh, that's new ... and totally crap plan. Oh well, we'll see how it works out. 14:08:19 I would not be comfortable to try to get this in for Newton anymore 14:08:22 Jokke_: can you please elaborate on your dislike about the plan? may be we can reconsider alternate plan for community images 14:08:36 nikhil: I was more worried about the release 14:08:42 I see 14:09:14 normally there has been stabilization between Rel-3 and RC1 to keep focus on fixing stuff rather than doing new 14:09:39 if that period is gone, this will just shift everyones focus away 2 months early 14:09:44 Jokke_: I will confirm the dates on the cut of stable/newton, but at this point that's where am at. 14:10:16 noted 14:10:51 I was able to keep up with a few updated to the project team guide, and ML emails but I would like to point out references before I claim authority over dates. 14:10:56 updates* 14:11:08 so my assumption for "by start of Ocata" was expectation of ~6-7 weeks from now, not 2 ;) 14:11:38 Jokke_: yeah :) it's a tricky understanding 14:12:02 Jokke_: officially ocata is supposed to start at the summit week (At least that's what I've been communicated before) 14:12:34 but technically, any commit that doesn't go in stable/newton will be Ocata 14:13:46 and that's what I really meant by Ocata (the release notes, updates to docs and communication to packagers/ops) will need to be done accordingly. However, from a developer perspective ocata features may not officially start until after 6 weeks. (as there are restrictions on what can get in until then) 14:14:49 cool ... move on ... sorry for hijacking 14:14:51 Either way, I won't be able to give much attention to CI (Community Images) until after Newton is stabilized. 14:15:16 So, there is expected delay to this development and we will try our best to get it done before the summit. 14:15:34 Please see the updates on the etherpad linked above. 14:15:37 Thanks! 14:15:46 #topic Import Refactor 14:16:07 So, this is the only major feature that needs core attention from here on to newton is released. 14:16:12 We've about 5 commits: 14:16:18 https://review.openstack.org/#/q/topic:import-refactor-newton 14:16:29 I've picked up: https://review.openstack.org/270980 14:16:44 and rosmaita has picked up 2 https://review.openstack.org/267209 https://review.openstack.org/312972 14:16:58 these are WIP and patches will be up when a good commit is up 14:17:11 we are trying to avoid pushing WIP stuff to gerrit 14:17:31 as it could potentially result into red herring (like before in the early phase of the cycle) 14:18:01 * Jokke_ starts stretching 14:18:08 The glance gate has been cleaned for about a month now and only 2-4 other commits look solid candidats for newton 14:18:53 Also, I expect to push another commit/amend to the spec that writes up "declaration of intent" 14:19:21 this will include details from our conversation early in the cycle and at the defcore ML (plus other details like defcore test spec) 14:19:48 oops, not defcore ML but defcore midcycle 14:20:15 Concrete stuff by next meeting. 14:20:24 comments? 14:21:13 waiting for that ... honestly I'm worried about the late time for this, but we'll see 14:22:05 oh well, we'd do only so much with no one dedicated to it the whole cycle and defcore being the bottleneck the whole time. 14:22:34 I forgot to mention, the above 3 commits are expected to cover the Discovery call needed for the import refactor. 14:22:54 nikhil: not saying that. I'm more concerned rushing it in now than the fact that we have had headwind pretty much last 2 cycles now 14:23:22 That's the first phase of the plan, I think there's a few updates required to good work from stuart right after that. 14:23:46 Jokke_: totally get that. I think we can discuss things around your concerns on the declaration of intent 14:24:05 just want to remind everyone to be careful with the reviews. Once we release that endpoint and it's behavior we need to stick with it 14:24:17 +1 14:24:20 the community issues have been resolved to great extent and people will have to think twice if not thrice before blaming glance about anything! 14:24:56 Jokke_: absolutely 14:25:09 particularly the discovery bit 14:26:32 moving on 14:26:41 #topic Releases updates ( nikhil ) 14:27:05 release schedule is: 14:27:07 #link http://releases.openstack.org/newton/schedule.html 14:27:25 Plan is to cut newton-3 by Aug 31st 14:27:34 but really depends on our momentum 14:28:00 "Not official yet" final releases of: 14:28:06 glance_store is 0.17.0 14:28:12 #link https://review.openstack.org/#/c/356857/ 14:28:21 python-glanceclient is 2.4.0 14:28:29 #link https://review.openstack.org/#/c/356855/ 14:28:45 but I hope that they are final (and we don't have to deal with any bugs) 14:28:54 comments/questions? 14:29:38 so we are feature complete for Newton? 14:29:59 In a sense that at least nothing will anymore be merged to the client? 14:30:21 yeah 14:30:34 nothing is WIP in client or store 14:31:14 any spcs that were approved and not implemented will be moved to ocata for implementation (no need for reapproval) . this includes specs from privious cycles. 14:31:26 wow 14:32:07 Jokke_: wow as in bad wow or good wow? 14:32:13 or worried wow :) 14:32:29 (just trying to make sure we don't get last minute surprises) 14:32:30 Not sure yet ;) 14:32:43 Surprised at least ;) 14:32:48 ok 14:33:12 I think we've had good attention for client and store in terms of reviews as compared to their size relative to glance. 14:34:14 moving on 14:34:30 #topic Barcelona Design Summit space needs 14:34:53 So, for Austin we had allocation as : 3fb, 5wr, cm:full (2 half days) 14:35:12 but here's the list of major constraints for Ocata 14:36:09 There's going to be change in the design summit style/conferences. So, Ocata will be *much* shorter cycle -- only 15 weeks between the design summit and feature freeze, including thanksgiving and other end-of-year holidays 14:36:35 ttx has indicated that we don't need to plan much work for the same. 14:37:01 In that perspective our work is also pretty much planned for short release 14:37:19 if we are going to take that import refactoring seriously 14:37:36 The slots are lesses than Austin and they had to squeeze in the design sessions a bit to accomodate ops meetup to not overlap with cross project workshops. 14:37:45 lesser* 14:38:25 So, contributors' meetup will only be maximum one half day (ie. only Friday afternoon) 14:38:45 as the sessions will be scheduled until Friday morning. 14:40:06 #info All: Plan your travel to summit/Barcelona to attend if ops meetup then Tues or if only cross project + project meetups then Tues evening 4pm onward and *back* on Saturday. 14:40:35 * Jokke_ is gonna be there Sun-Sat all booked already 14:41:00 The contributions to the project and activity etc will be taken into consideration before the slots are allocated to us. 14:41:25 We still need to give our preferences to the planners: 14:42:06 they want to be done by Aug 31st so here's my take on what will work for us 14:42:29 2FB, 4WR, 1CM 14:42:46 our request is lesser than Austin and it's not too less. 14:42:50 I'd say even 2+2 would be likely enough for us 14:42:55 sorry, can you expand the acronyms? 14:43:00 Also, like Jokke_ rightly pointed out we've a bit of backlog too. 14:43:18 fishbowl, workroom, contributors' meetup 14:43:21 FB == Fishbowl ( Our traditional largish rooms organized in fishbowl style, with 14:43:21 advertised session content on the summit schedule for increased external 14:43:21 participation. Ideal for when wider feedback is essential.) 14:43:33 thos are (from Wednesday 4pm to Friday noon) 14:43:44 WR == Workroom sessions (from Wednesday 4pm to Friday noon) 14:43:45 Smaller rooms organized in boardroom style, with topic buried in the 14:43:45 session description, in an effort to limit attendance and not overcrowd 14:43:45 the room. Ideal to get work done and prioritize work in small teams. 14:43:56 CM == Contributors meetup (Friday afternoon) 14:43:56 Half-day session on Friday afternoon to get into the Ocata action while 14:43:56 decisions and plans are still hot, or to finish discussions started 14:43:56 during the week, whatever works for you. 14:43:58 2+2+1 I mean 14:45:11 so for work session, I think what we need is: 14:45:16 1 for ops meetup (like in Austin -- may be useful for rolling upgrades) 14:45:17 1 for defcore meetup (casual updates and get to know plus communicate awarenss) 14:45:43 1 for updates on the code for CI and import (half each) 14:46:12 and 1 for anyone who has a idea (we could potentially accomodate new contributors to ensure our team keeps growing and reviews are not stalled) 14:46:31 rolling upgrades? 14:46:35 i don't think the defcore one is necessary? 14:46:58 i think we need 1 wr for image import for sure, dont' think we should split it 14:47:11 and we will need community images (CI) 14:47:23 hemanthm: that was just my take, I don't know how much feedback we still need from ops 14:47:54 ops meetups have always helped put a good perspective 14:48:01 nikhil: I thought you said there is this time dedicated time for the ops meetup so we don't need to use design sessions for that 14:48:17 we are working on three big changes for Ocata: image import, community images and rolling upgrades. And, the ocata dev cycle is much shorter as pointed out earlier 14:48:28 rosmaita: the defcore one was to indicate collaboration between osc/infra/defcore/glance/glare. but upto the team on what they think is a good idea. 14:48:40 Well I hope we wouldn't need community Images as nikhil pointed out that it's best case scenario merged by summit 14:48:41 are we comfortable with tackling those three big changes in a short cycle? 14:49:10 Jokke_: if people are expecting to go to ops meetup then surely we can avoid that. I am just taking the prior experiences into account. 14:49:43 well, I don't think we need to include all the changes that we discuss at the summit in that specific cycle itself 14:49:58 hemanthm: I'm sure we can comfortably take them on and not deliver :P 14:50:09 but if we are going to have specs discussions, then it's best to take every opportunity to have some F2F time 14:50:18 ++ 14:50:45 i think rolling upgrades are important, we need a devoted session 14:50:49 having F2F time reduces the risk of red herrings 14:50:51 true, but it helps to manage/set expectations 14:50:56 I also think we need to make decision at least what we & glare folks are going to do about glance_store 14:50:57 we can't leave it open ended 14:51:12 yep 14:51:18 Jokke_: ^ 14:51:25 we likely need to at least come together and see if we are going to refactor it for both projects or forget about it 14:52:40 nikhil: just noticed we have 8min ... was there something else in the agenda? 14:52:41 hemanthm: yes, for planning we need interested people at the CM, to be able to come to a common understanding about the cycle plans. 14:52:53 Jokke_: nothing from me 14:53:00 we can continue this in open discussion 14:53:05 #topic open discussion 14:53:25 our "short agenda" ;) 14:53:33 :) 14:53:40 he wasn't wrong. 14:53:57 0:) 14:54:19 ty tsymanczyk (I indeed wasn't under the impression this mtg will stretch so long) 14:54:49 hemanthm: and rosmaita : you're currently driving rolling upgrade specs. so I don't want to enforce having a session on you guys. 14:55:10 but asking a slot for it is still better than not asking for it in the first place 14:55:18 tsymanczyk: so quick question. Some acute issues you're fighting with on the community images? 14:55:20 yes, we definitely want a slot 14:55:27 we can always concede our slot later but can't necessarily ask for more 14:55:47 ++ rosmaita 14:56:10 my point earlier is that i think we need a full slot for rolling upgrades 14:56:23 rosmaita: and any objections for it to be shared with ops/sync but I guess we don't have to plan that now 14:56:25 ? 14:56:34 rosmaita: I see 14:56:43 i think it needs to be RG focused 14:56:53 should I start a etherpad for summit planning? 14:57:03 yes!!! 14:57:08 aight 14:57:15 we can discuss all this on the etherpad 14:57:44 #action nikhil: start a etherpad to discuss the slots needed for Barcelona summit 14:58:00 any more items? 14:58:13 nikhil: was there a deadline for submission for design summit slot requests? 14:58:45 rosmaita: it was 31st 14:58:59 rosmaita: latest Aug 31st is when the data needs to be sent from our end 14:59:03 ty 14:59:17 Jokke_: about community images: https://etherpad.openstack.org/p/community-images-work-items 14:59:37 30 secs :) 15:00:00 rosmaita: -glance? 15:00:04 Thanks all! 15:00:22 rosmaita: yt? 15:00:32 yup 15:00:50 rosmaita: want a min or should we chat on the other channel? 15:00:50 nikhil: took me a minute to realize what you meant! 15:00:51 o/ 15:01:03 o/ 15:01:08 ok, let's end this meeting 15:01:14 crowd here for next one 15:01:16 #endmeeting