14:01:37 #startmeeting glance 14:01:38 Meeting started Thu Sep 1 14:01:37 2016 UTC and is due to finish in 60 minutes. The chair is nikhil. Information about MeetBot at http://wiki.debian.org/MeetBot. 14:01:39 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 14:01:40 #topic roll call 14:01:41 o/ 14:01:42 The meeting name has been set to 'glance' 14:01:47 o/ 14:01:48 o/ 14:01:49 o/ 14:01:51 o/ 14:01:56 let's give a couple mins 14:02:04 o/ 14:03:07 #topic agenda 14:03:13 https://etherpad.openstack.org/p/glance-team-meeting-agenda 14:03:42 Mostly setting the plans for next few weeks and discussing outstanding issues. 14:03:59 Also, we need to come on the same page for the release work. 14:04:15 #topic Release updates (nikhil) 14:04:33 newton-3 is today 14:05:03 #link https://releases.openstack.org/newton/schedule.html 14:05:25 I've created a link for getting the patches that are waiting to go in for newton-3 14:05:27 #link https://review.openstack.org/#/c/363930/ 14:05:48 looks like there are a few still needing reviews 14:06:05 I will list those from my end and others are encouraged to point out here 14:06:26 #info *must* go in newton-3 reviews are below: 14:06:34 #linl https://review.openstack.org/354809 14:06:53 #link https://review.openstack.org/354809 14:06:57 #link https://review.openstack.org/#/c/229220/ 14:07:09 Thanks to sigmavirus we've managed to merge that one 14:07:23 o/ 14:07:40 missing release notes can come in before rc-final is out 14:08:15 #info nice to have reviews for n-3 14:08:20 #link https://review.openstack.org/324012 14:08:33 if that goes in, we *must* bump up the API version 14:08:39 #link https://review.openstack.org/350809 14:08:57 that's it from my side, anyone else? 14:09:32 I see 3-4 reviews related to config help text improvements 14:09:41 https://review.openstack.org/#/c/354332/ 14:09:46 what's the status and requirement for those? 14:09:56 this was +2 already but needed rebase 14:10:53 ok then, I wil assume no others are being requested and we can ship out n-3 right after the above linked are merged. 14:11:52 https://review.openstack.org/#/c/363838/ 14:11:54 needs workflow 14:12:32 same for: https://review.openstack.org/#/c/363870/ 14:13:15 #info client and store stable/newton branches have been cut. Reviewers are free to merge any outstanding commits. Ocata releases don't start until R-0. 14:13:26 thanks abashmak , hoping that's it? 14:13:57 moving on 14:14:10 #topic Import Refactor (nikhil) 14:14:58 as per our discussion from after the glance's weekly meeting 14:15:01 #link http://eavesdrop.openstack.org/irclogs/%23openstack-glance/%23openstack-glance.2016-08-25.log.html#t2016-08-25T15:00:22 14:15:25 the plan is to work on a feature branch where reviewers are more comfortable to test this and move ahead fast 14:15:50 also this will help us avoid any red herrings from outside the project and we can resolve conflicts once everything is ready 14:15:59 I will get that setup later today 14:16:17 FWIW, I couldn't attend last weeks meeting and I'd have appreciated this to be brought up to the mailing list. I actually think a feature branch is not the right approach here as it'll mess up the priority of this work review wise 14:16:22 ha.. 14:16:40 I was writing :) .. flaper87 had some concerns on this plan and I will try to resolve them 14:16:57 I don't think it sends the right message and I'd really appreciate more eyes on the reviews than fewer, which is what we'd get on a feature branch 14:16:59 nikhil: :P 14:17:22 flaper87: sure, so.. more explanation is required definitely on 'why' this plan 14:17:35 flaper87: so, the intent is to not avoid more /wider review set 14:17:38 Also, given the fact next cycle is going to be shorter, I think this will just add more work 14:17:52 nikhil: I know that's not the intent but I'm afraid that's what's going to happen :( 14:18:39 flaper87: unfortunately, this is the only way forward.. I tried to setup many reviews this cycle & poked people for them but haven't gotten reviews. 14:19:06 for example: the value discovery patch has been ready (refreshed) but no reviews 14:19:16 I have to disagree, really. I think a feature branch for such important work is a mistake 14:19:36 well, one review :) 14:19:37 https://review.openstack.org/#/c/270980/ 14:20:00 I think the lack of reviews might also come from a lack of awareness on this work by other folks and definitely lack of focus (glare?) 14:20:21 flaper87: feature branch or not, I think the intent to just focus on 2 things before the summit 14:20:22 * flaper87 is definitely not trying to find the cause of the current state 14:20:49 flaper87: fwiw, the glance reviews and reviewers have dropped drastically since the split of the project 14:20:54 i guess we could think of master as the import feature branch? 14:20:57 we can keep disagreeing :) 14:21:00 nikhil: yup, noticed that :( 14:21:23 rosmaita: yeah, I am not interested in the details really 14:21:35 rosmaita: I think the request for feature branch was for 2 main reasons 14:21:40 1) avoid merge conflicts 14:22:04 2) have a way to test the entire set of patches cleanly (without extraneous commits from master) 14:22:16 thing is, you'll have to deal with those merge conflicts eventually 14:22:24 anyway, just my $0.02 14:22:26 true 14:22:34 flaper87: so the plan fully is like this: 14:22:42 1. create a feature branch 14:23:04 2. set up a test bed for the entire set of patches to deterministically understand import discovery and behavior 14:23:31 3. get glance's (core and other influencers') consensus 14:23:40 4. push the entire set to master 14:23:56 5. link, poke, push cross project teams to review the set again 14:24:17 There are 2 advantages to this approach (besides the disadvatages you describe): 14:24:33 1) have all glancer's on the same page about the feature 14:25:05 2) while discussing with wider set of interest parties, have the information and behavior well udnerstood before making further claims 14:25:34 That's where we seem to disagree, I don't think a feature branch will help with #1 and/or #2 14:25:56 Having said that, I believe in democratic decision making (hence the dissatisfaction with glare too): 14:26:13 so, I am okay to vote on the plan and we can etherpad this out 14:26:46 vote sounds good, It might not be needed if everyone agreed during the last meeting 14:26:56 I don't mean to act as a blocker but I definitely wanted to let it all out 14:26:57 :P 14:27:06 flaper87: well, it was just rosmaita Jokke_ and me :) 14:27:07 I plan to help with reviews where/when possible 14:27:21 Jokke_: is on PTO today 14:27:31 so it'd be you, rosmaita and me (and whomever wants to vote) 14:27:32 :D 14:27:36 he told us.. hence the confusion :) 14:27:44 gotcha 14:27:55 ok, if you and rosmaita still think feature branch is good 14:28:02 i am on the fence 14:28:05 I will stand by it 14:28:13 rosmaita: sorry? :P 14:28:24 i haven't seen this done in glance before, i don't think? 14:28:31 flaper87: nah, we can ask votes from croelandt , sabari , stuart and othes too (whoever thinks they will have time for reviews) 14:28:37 yup 14:28:40 let me put it this way 14:28:47 I'm don't think I'll code on this 14:28:50 I'll provide reviews 14:29:08 if people that are going to code on this prefer it that way, then I guess I'm ok (which doesn't mean I agree :P) 14:29:20 flaper87: I'm on the fence too 14:29:32 just wanted to point out strong opinions I heard from last sync 14:29:42 otherwise we will have another set of confusion! 14:29:46 nikhil: yeah, thanks for that btw. I appreciate the extra context 14:30:03 flaper87: so, I can assume you acting proxy for both Jokke_ and croelandt 14:30:23 Jokke_: would kill me if I do :P 14:30:40 nah, I just found out about the feature branch today and I didn't have a chance to talk to Jokke_ 14:30:46 I thought croelandt was around 14:30:47 croelandt: YO! 14:30:51 * flaper87 shakes croelandt 14:31:09 gotcha 14:31:09 * flaper87 shakes France 14:31:15 LOL 14:31:16 it ain't working 14:31:37 * flaper87 is hereby entitled to speak for croelandt 14:31:39 :P 14:31:53 jokes apart, process wise I don't think it'll help 14:32:01 and by process I mean reviews and get it into master 14:32:20 Code wise, it might. fewer frequent conflicts (you have conflicts when you rebase) 14:32:36 okay, so looks like one -2 on the feature branch and no +2 for it :D 14:32:46 also, I don't want the feature branch to be an excuse for "I'll do weird random commits because meh, it's a feature branch" 14:33:07 flaper87: oh, totoally a possiblity 14:33:39 flaper87: yep 14:33:40 so... this work is open to coding from others but my preference is to keep the team small and connected (So sync regularly and even if there's nothing to sync just talk non-sense) 14:33:41 huhu 14:33:54 I'm a bit skeptical about the whole thing as well 14:34:17 croelandt: say some more, flaper87 has volunteerd you to do lots of work on this 14:34:27 indeed 14:34:37 * flaper87 won't try to hide the fact he has literally volunteered croelandt 14:34:39 :P 14:35:31 rosmaita: I'm afraid we might end up with more issues than we'll solve, and more complexity in the process 14:35:31 rosmaita: any specific opinions you are looking for? 14:35:34 did we lose contact with France again? 14:35:37 oh, there he is 14:35:42 now I might not be perfectly clear on the details, but eh 14:36:16 okay, let's just discuss the details process later but from the opinions in today's meeting 14:36:24 croelandt: you mean by having the feature branch? 14:36:34 #agreed import refactor work will NOT have a feature branch and work continues in master 14:36:47 rosmaita: I mean, it might help wit hthe conflicts, but you'll end up having to fix them anyway 14:36:53 croelandt: gotcha 14:36:57 the whole review process is already heavy enough 14:37:03 heh 14:37:16 croelandt: that's the thought of the day! 14:37:23 i am in agreement with the #agreed statement 14:37:27 + it will get even more complex for non-pro contributors 14:37:28 * flaper87 gets his wine 14:38:08 should we move on? 14:38:15 yep 14:38:31 thanks all for the valuable input! 14:38:34 moving on.. 14:38:37 thanks for listening 14:38:39 :) 14:38:46 #topic Community Images (nikhil, tsymanczyk) 14:38:48 :) 14:39:08 not sure if tsymanczyk is around today.. 14:39:13 should i talk? 14:39:18 he is! 14:39:24 (I did see you hi5) 14:39:34 tsymanczyk: please go ahead or I'd have proxy you :) 14:40:26 * rosmaita just noticed that flaper87 has wine 14:40:40 rosmaita: I have a wine cellar :P 14:41:00 tsymanczyk: you still writing? 14:41:14 the original patchset that swaps in a visibilty column is still full of broken tests and nightmares. locally i believe i have sqlalchemy/api.py solid. starting looking at simple.api.py and there's weirdness from the getgo i still need to identify 14:41:44 doing this non tdd is not usual for me. the broken tests make everything hard 14:41:51 even thoygh i broke them 14:41:56 :) 14:42:11 ipad typing is slow 14:43:18 tsymanczyk: I think we forgot to sync on the TDD aspect. that non-TDD was only to help get things ready for newton reviews. now that the plan is different, we can go back to TDD :) 14:43:22 that's all i know. questions comments? 14:44:15 oh and i'll becollapsing the three sets into two. there's no real calue and actually downside to seperating the latter two 14:44:41 will be easier to comprehend and review 14:44:48 tsymanczyk: only comment is thanks for continuing to work on this! 14:44:55 ++ 14:45:24 thank you foryour continued support and patience! 14:46:14 * nikhil was imagining rosmaita and tsymanczyk xoxo :D 14:46:36 o_O 14:46:37 * nikhil is surpised no comments from flaper87 on the comment 14:47:05 hey, we are trying to show that the glance community is a caring and supportive one! 14:47:09 * nikhil trying to reduce everyone's stress level 14:47:19 rosmaita: indeed a good idea 14:47:20 group hug everyone! 14:47:29 (is there an emoticon for that?) 14:47:33 :hug: 14:47:40 {} i think 14:47:53 \o-O-o/ 14:47:57 ok, so {{{{{{{{{{{}}}}}}}}}}} 14:48:14 mine was more for a team celebration, less of a hug 14:48:25 (celebration posing for a picture) 14:48:46 aight looks like no more comments, moving on to the next topic 14:49:01 #topic Mission Statement (rosmaita) 14:49:22 i put that on the agenda because earlier in the week it was just sitting there 14:49:33 ah 14:49:35 i thought the TC was waiting for more glance folks to weigh in 14:49:57 but shortly after i put it on the agenda, the roll call votes started rolling in 14:50:08 so now my announcement is thtat the TC has approved the change 14:50:56 #info Mission statement for glance has been updated since the glare split see https://review.openstack.org/#/c/354053/ 14:51:10 thank you! that's what i was looking for 14:51:18 nikhil: lol, missed that, sorry 14:51:18 have too many windows open 14:51:23 that's all from me 14:51:25 rosmaita: ++ 14:51:49 :) 14:52:14 #topic Open Discussion 14:52:25 #link https://bugs.launchpad.net/nova/+bug/1616539 14:52:25 Launchpad bug 1616539 in OpenStack Compute (nova) "architecture not validated in "openstack image create"" [Undecided,In progress] - Assigned to tamil vanan (tamilhce) 14:52:46 who proposed that to the agenda? 14:53:06 though, I remember rosmaita chatting with someone regarding this bug?? 14:53:07 probably tamilhce 14:53:10 i suggested it 14:53:22 ah 14:53:33 so briefly here's the issue 14:53:43 we have "common image properties" in glance 14:53:49 that really aren't special 14:54:00 they were introduced for interoperability 14:54:17 so that all openstack clouds would use the same metadata key for these things 14:54:24 (they are listed in the docs, by the way) 14:54:36 but, it was completely voluntary for people to acutally use them 14:54:43 #link http://docs.openstack.org/developer/glance/common-image-properties.html 14:55:00 and at the time, the PTL said that we didn't want to list values because they were likely to change too fast 14:55:12 whereas people's schemas would most likely only change release-to-release 14:55:40 so the problem tamil is addressing is that there's no control over what values go in there 14:56:00 rosmaita: but we do the schematic validation! 14:56:03 and if you misspell something, and a consumer (like nova) is loooking for one of these, it will choke 14:56:11 ah.. 14:56:13 nikhil: we just make sure it's a stirn 14:56:16 *string 14:56:23 yeah, gotcha 14:56:42 now, in these modern times, we have metadefs 14:56:49 lol 14:57:07 so glance has a catalog of metadata keys (standardized) with descriptions of what values they can have 14:57:42 * nikhil remmebers the chat on attempting to enforce them using metadefs 14:57:51 my opinion is that instead of trying to change the image schema piecemeal, if we want to have value enforcement, we should do it through metadefs 14:58:13 that way it will extend past just one or two "common image properties" 14:58:33 time check, 90 seconds 14:58:46 anyway, i guess comment on tamil's patch? 14:59:12 rosmaita: in short, I think this needs much further discussion for we're overlapping two fundamentally different parts of glance together 14:59:12 i will comment, other interested parties please follow up 14:59:23 nikhil: ++ 14:59:26 spec for sure 14:59:31 rosmaita: also, Ive a concern regarding the persistence of the metadefs 15:00:02 for instance, could someone forget to update the .json files and override such tweaks via the api (w/o exporting) 15:00:07 anyway, out of time 15:00:12 thanks all for joining! 15:00:23 #endmeeting