13:59:18 #startmeeting glance 13:59:18 Meeting started Thu Mar 31 13:59:18 2016 UTC and is due to finish in 60 minutes. The chair is nikhil. Information about MeetBot at http://wiki.debian.org/MeetBot. 13:59:19 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 13:59:21 The meeting name has been set to 'glance' 13:59:22 #chair jokke_ 13:59:23 Current chairs: jokke_ nikhil 13:59:37 woops 13:59:39 #topic roll call 13:59:41 o/ 13:59:44 o/ 13:59:46 o/ 14:00:26 jokke_: so you want to/had something to lead meeting today? 14:00:38 o/ 14:00:51 may be I am out of sync there, flavio said he will be out and ask me to conduct this one 14:00:57 nikhil: flaper87 is travelling and just asked if I would run it through ... feel free though 14:01:00 o/ 14:01:08 o/ 14:01:10 jokke_: ah, he's keeping back ups 14:01:14 oh so it was Flavio who was out of synk ;) 14:01:18 0/ 14:01:20 o/ 14:01:21 s/k/c 14:01:22 Mitaka is almost over 14:01:29 nikhil: go for it 14:01:31 we've the new rc2 14:01:48 yes, translations only 14:01:49 but it's i18n so, prolly nothing significant to come out of it 14:01:52 :) 14:02:01 o/ 14:02:02 guess we've a good turn out today 14:02:07 let's get started then 14:02:10 #topic agenda 14:02:16 #link https://etherpad.openstack.org/p/glance-team-meeting-agenda 14:02:35 Just wanted to quickly sync over a few items and start preparing for the summit 14:02:44 we've mostly bootstrapping newton at this point 14:02:52 that's what we will discuss today 14:02:58 o/ 14:03:07 by the looks of Mitaka is wrapped and ready for shipping 14:03:11 #topic Glare updates ( mfedosin ) 14:03:17 jokke_: yeah! 14:03:49 hey! we continue the developing of Glare 14:04:10 a lot of new features were added last week 14:04:37 and now we have functional tests that checks that everything is fine 14:04:51 also there was defcore meeting yesterday 14:04:59 and we were talking Glare 14:05:22 so, afaiu, they have no concerns about Glare 14:05:22 what review strategy should we be following on these? i've been ignoring the WIP patches, maybe that is a mistake 14:05:43 rosmaita: good question 14:05:45 rosmaita: no, you're right 14:05:52 we need to decide on this asap 14:06:01 currently we just develop the base code 14:06:14 and it's easy for us to have it in one big commit 14:06:27 About that. I got couple of pings last night asking wtf is going on. Could we please stop selling Glare as replacement for Glance at least until we have a) stable API and b) some level of track record/testing that it actually is successfully working 14:06:30 called Glare All-in-one 14:07:04 ^^ ref defcore meeting 14:07:29 then we implement new features with [WIP] prefix and merge them in the big commit 14:08:10 jokke_, at the current stage that seems more PoC to me 14:08:10 jokke_: selling it as a glance replacement creates a hype around Glare :) 14:08:17 jokke_: interesting, that was more of a question rather than a sell. I think those defcore folks are not accurate on their assumptions. I'd love to clarify them. 14:08:44 mfedosin: and causes hell of a lot confusion in the community where we have already really bad reputation (warranted or not) 14:09:25 link for the all-in-one patch -> https://review.openstack.org/#/c/292327/ 14:09:44 thanks tjcocozz :) 14:09:45 over Mitaka, I found out the bad reputation is only within small pockets and we need to identify those pockets and clear those confusions 14:09:58 mfedosin: I'm not picking fight at this point and I do understand that some level of hype is needed to get the traction to Glare, but we can't do it this way 14:09:59 +1 to nikhil 14:10:33 It would be useful to have clarity on the relationship between the two. Maybe some documentation in one of the Glare patches? 14:10:37 and I agree with jokke_ that we'd live on realistic practices and avoid confusing hypes 14:11:11 mclaren: there will be a spec 14:11:41 Ok, cool. Whatever gets us on the same page. 14:11:51 and misunderstandings around defcore was the root cause for last huge incident we were sorting through multiple sessions last Summit 14:12:05 i think it would be good to get something about "the future" into the regular docs 14:12:14 rosmaita: ++ 14:12:25 rosmaita: sounds good 14:13:01 #action nikhil: clarify "the future" stuff for Glance 14:13:19 i can put up a patch with a lot of handwaving, and you can correct it 14:13:36 sure 14:13:53 anything else on this topic? 14:14:00 rosmaita: whatever gets us going 14:14:18 moving on 14:14:22 #topic Nova v1, v2 14:14:52 I'd a quick catch up with Nova team on the status of this (earlier in the week) 14:14:55 We've a few action items for Newton 14:15:29 Need to create a newton spec, clarify the individual reviews to ensure they correct the eventual goal 14:15:50 basically Nova team thinks that some reviews need more context on why they exist 14:16:22 Matt has tentatively scheduled a cross session with Glance 14:16:43 We will know if it makes it or not in their final proposal 14:16:48 I'll take the spec 14:16:57 but one thing I really like 14:16:58 good news is that he also publicly put this on the Nova Newton priorities ;) 14:17:00 mfedosin: thanks 14:17:10 they agreed to keep this code in Nova 14:17:16 mfedosin: ++ 14:17:17 without compat layer 14:17:28 yes, for this cycle it will be a small concise change 14:17:29 mfedosin: that sounds great 14:17:33 makes sense 14:17:34 most of what mfedosin has already done 14:17:44 all patches are there 14:17:47 there are a few testing improvements that we may have to do 14:18:00 but testing impr. is a research item 14:18:05 volunteers are welcome 14:18:07 I just need to address rosmaita concern about lowering headers 14:19:01 any questions? 14:19:23 moving on 14:19:36 #topic Cross project updates 14:19:51 The etherpad for session proposals is up 14:19:59 #link https://etherpad.openstack.org/p/newton-cross-project-sessions 14:20:06 quite a few sessions up there already 14:20:34 besides that 14:20:50 there was a meeting yesterday 14:21:28 Most of what was discussed was a follow up on the CP specs so that people are aware of them 14:21:45 there was update from release team on the automating their process 14:22:14 a heads up + call for reviews on quotas and service catalog tng 14:22:51 moreover, you can find a gist of CP updates in our community newsletter 14:23:11 ref that release process update 14:23:22 nothing significant, that the glance team need to look into this or next week 14:23:30 jokke_: please go ahead 14:23:42 that mainly affects non-managed release mode projects, so really does not affect us 14:23:54 we keep requesting the tags as usual :) 14:24:46 totally 14:25:01 those release changes shouldn't affect glance 14:25:14 just that if anyone is thinking of taking up that liaison post, it's quite straight forward ;) 14:25:34 if you work on other projects too and see anomalies in the process, then that update is for you 14:25:43 jokke_: ah ha 14:25:46 good call 14:26:05 #info all: glance team is looking for release liaison 14:26:32 #topic Releases 14:26:41 I'd switched topic before :) 14:26:59 nevertheless, summer is coming 14:27:09 and if you're making early holiday plans 14:27:21 the schedule is conveniently available to you 14:27:24 #link http://releases.openstack.org/newton/schedule.html 14:27:53 #info glance team would love to have cores available a week before and on the week of releases 14:28:14 #info Glance RC2 tagged. Diff RC1..RC2 translations 14:28:40 I'd a question yesterday on backports 14:28:44 #info glance_store & python-glanceclient has not had anything major during the freeze 14:28:59 glance_store has 14:29:00 if you're looking to propose changes to stable branches, it should follow stable guidelines 14:29:01 I think 14:29:25 https://review.openstack.org/#/c/297665/ 14:29:28 #link http://docs.openstack.org/project-team-guide/stable-branches.html 14:30:16 kairat: that looks possible from first glance 14:30:18 kairat: let me refrase, has not merged anything that would need to be released for Mitaka 14:30:32 ah, ok 14:30:57 #undo 14:30:58 Removing item from minutes: 14:31:00 #undo 14:31:01 Removing item from minutes: 14:31:14 for backports, please propose change against master first and then based on reviews we will move for likely backport 14:31:33 #info glance_store & python-glanceclient has not merged anything to be included in Mitaka for release time 14:31:41 #link http://docs.openstack.org/project-team-guide/stable-branches.html 14:32:11 kairat's fix looks important 14:32:33 nikhil: and on top of that: Please make sure that there is bug for it. We will not backport without and doing that after merge to master is pain 14:33:14 how do we handle glance_store fixes coming up to release? 14:33:16 jokke_: okies.. bug first it is! 14:33:24 :) 14:33:30 makes life easier 14:33:34 mclaren: can you elaborate? 14:33:54 sure 14:34:15 eg do we tag RCX-candidate or anything? 14:34:23 ah 14:34:40 mclaren: with both glance_store and python-glanceclient we do not follow any stable release schedule so we can backport and release as needed, new deployments should pick those up within the requirements 14:34:42 we now follow SemVer for clients and non-client libs 14:35:19 It looks like if you upgrade to mitaka you could get hit by bug 297665. So ideally we'd have a release of glance_store to prevent that? 14:35:39 yes 14:35:57 we will have a separate release candidate for mitaka 14:36:03 mclaren: so no RCs, the releases we did that was base for the stable/mitaka brnaching is effectively our stable/mitaka release from those libs. then we just make patch releases as needed from that point onwards 14:36:20 mclaren: this was done ~3-4 weeks ago 14:36:37 I understand that in principle they are separated. 14:37:05 I'm just saying that we should do what we can to avoid people hitting 297665 when they deploy mitaka. 14:37:28 mclaren: yes and for that they will be expected to deploy stable mitaka 14:37:32 for eg. iirc, we'd a 0.10.x released for mitaka 14:37:41 mclaren: the one thing I like having glance_store as separate from glance service is the ease of releasing when we have something we want to have fixed in production 14:37:41 the fixes will go to 0.10.x+1 14:37:55 and for newton it will be 0.11.x series 14:38:05 nikhil: correct 14:38:24 the requirements are then updated to reflect the bug fix in stable/mitaka 14:38:35 jokke_: thanks for confirming! 14:39:02 I don't see why releasing a glance_store fix is easier than releasing a glance stable fix? 14:39:19 it is somewhat easier 14:40:01 I mean, you can debate on this but.. 14:40:07 because you can now control the fix through requirements 14:40:08 mclaren: mainly due to the appearance :) 14:40:19 you've a better visibility on what's fixed and what's not 14:40:49 you can switch the fix back if you need to by a simple requirement contraint change 14:41:04 constraint* 14:41:34 glance_store release documentation will also have the info for it (ideally) 14:42:09 btw, sorry to interrupt, but we probably should move one and take this offline with interested parties 14:42:13 running out of time 14:42:15 we're running out of time, but if you've more concerns we can wait a couple more mins 14:42:50 ok, offline then. moving on for now. 14:43:04 #topic Use policy files to control deletion of deactivated images. (bunting) 14:43:11 Me 14:43:15 #link https://review.openstack.org/#/c/256381/ 14:43:20 bunting: go go go :D 14:44:02 It was suggested to bring this to the attention of the group to use policy files rather than the config options as a potential pattern to use in future patches 14:44:56 sure 14:45:02 More of an info topic then anything that needs discussing 14:45:04 I see the comments on PS5 and they make sense 14:45:31 alrite then 14:45:57 regardless how it might be controlled, I'd like to see spec (at least lite unless this discussion is getting complex) as the proposed is directly against what the original deactivate/activate spec is specifying as behavior 14:46:07 #info please help review Halt deletion if image is deactivated patch https://review.openstack.org/#/c/256381/ 14:46:53 thanks for bring this bunting! 14:46:55 moving on 14:46:57 do we really need a spec for a 45 line change? 14:47:07 heh 14:47:13 40 lines are tests btw 14:47:23 well, let's discuss the approach on the review 14:47:38 if it changes the API behavior we may need a (lite)spec 14:47:50 if it's just policies, I think we'd be fine 14:48:03 we have an APIchange commit tag 14:48:05 but we definitely need the latest documented behavior on deactivation 14:48:14 so we need to update history on the older spec 14:48:21 make sense? 14:48:21 ++ to mclaren 14:48:49 mclaren: I don't think the amount of code is the driving reson for specs, as said we have spec that specifies the current behavior. Just want to have it documented in the same place why we changed it so that in Ocata there is no bug removing these things as they are bug against the spec it should behave 14:49:14 jokke_: very well put 14:49:39 s/spec it/spec how it/ 14:49:59 motto being: let's avoid back and forth discussions and changes and keep the code sensible for operators 14:50:14 moving on 14:50:17 #topic Glance priorities for Newton 14:50:27 #link https://etherpad.openstack.org/p/glance-newton-priorities 14:50:35 I've added a few items 14:50:57 etherpad was created by Flavio (last week I think) 14:51:07 we'd use it at the contri meetup 14:51:13 (at the summit) 14:51:23 this was a info update 14:51:24 moving on 14:51:49 #topic CP summit sessions 14:52:20 we need volunteers to be present at the _different_ CP sessions 14:52:32 I'd like to co-ordinate with the team so that we cover the breadth 14:53:02 #action nikhil: create etherpad to gather info regarding CP sessions interest from glance team 14:53:15 moving on 14:53:18 #topic Glance mid cycle 14:53:36 I got a first confirmation regarding the midcycle to happen in Boston 14:53:46 it will be from June 8-10 14:54:00 why so early? 14:54:01 let me know by Tuesday next week if you've any concerns 14:54:10 otherwise that will be the final plan 14:54:16 it's right after m1 14:54:26 it's overlapping with OpenStack Day Ireland 14:54:29 and we'd have the spec freezer around m2 14:54:37 which is at 10th 14:54:51 jokke_: so would people prefer Mon-Wed instead? 14:55:01 I saw a mail, but was there some discussion on the venue at all? 14:55:12 mclaren: I'm trying to move fast this time 14:55:18 as I want to discuss import work 14:55:26 at the mid cycle too 14:55:37 Boston is almost at the midpoint for 14:55:42 west coast and EU zone 14:55:48 Honestly don't take my comments too heavy in. Just wanted to point that out. I'm unlikely to participate anyways 14:56:01 I'd love to have mclaren sabari rosmaita to the very least at the venue 14:56:10 make sense? 14:56:26 feel free to give me your suggestions 14:56:41 I was going to blow people's minds and suggest a non-US venue for a change (we've hosted a couple of mid-cycles here in Galway for example) 14:56:45 this is WIP right now, nothing final no commitments made yet 14:56:55 * rosmaita mind is blown 14:57:03 * nikhil too 14:57:05 lol, too easy 14:57:29 anyway, wanted to throw that out there rather than put a spanner in the works 14:57:37 absolutely 14:57:42 if it's Galway, I might be more likely preset (in the case I will be let into the building) :P 14:57:44 mclaren: +1 for Galway - you have good beer 14:57:51 +1 14:57:54 ++ 14:57:57 and thanks for proposing (assuming: and willing to sponsor :D) 14:58:21 quick poll: likely to get more attendance in galway vs. boston? 14:58:24 iirc flaper might be here around that time as well ;) 14:58:33 btw, 2 mins 14:58:38 reply to that thread 14:58:51 I want to know who can and can't make it there 14:59:03 I'd probably be able to make both I think 14:59:09 ok, will check with my manager this afternoon 14:59:14 i really want to go to galway 14:59:22 1min! 14:59:27 ;) 14:59:29 Galway! 14:59:33 we have outdoor and indoor activities :-) 14:59:37 we can discuss summit topics next meeting 14:59:57 lets get the Galway option to the mailchain 14:59:59 http://blog.petercmurphy.com/wp-content/uploads/2012/06/DSC_1928.jpg http://dd73muqoroywk.cloudfront.net/images/detailed/5/G6046.jpg?t=1428042089 15:00:00 and on the etherpad .... please record your interest, everyone 15:00:04 and/or ML 15:00:08 I will have another theard around virtual meetup before summit 15:00:17 etherpad for design topics, i mean 15:00:25 jokke_: only if there won't be delay 15:00:34 Thanks all! 15:00:37 otherwise I'm going with what's planned 15:00:41 we're out of time 15:00:42 Thanks all for joining! 15:00:44 #endmeeting