14:01:14 #startmeeting Glance 14:01:15 Meeting started Thu Oct 8 14:01:14 2015 UTC and is due to finish in 60 minutes. The chair is flaper87. Information about MeetBot at http://wiki.debian.org/MeetBot. 14:01:16 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 14:01:19 Courtesy meeting reminder: ativelkov, cpallares, flaper87, flwang1, hemanthm, ivasilevskaya, jokke_, kragniz, lakshmiS, mclaren, mfedosin, nikhil_k, 14:01:19 The meeting name has been set to 'glance' 14:01:19 o/ 14:01:21 Courtesy meeting reminder: Nikolay_St, Olena, pennerc, rosmaita, sigmavirus24, sabari, TravT, pkoniszewski, krykowski, ajayaa, GB21, bpoulos, harshs, abhishekk, bunting, dshakhray, wxy, dhellmann 14:01:23 o/ 14:01:26 o/ 14:01:32 o/ 14:01:34 o/ 14:01:38 o/ 14:01:45 w000h000! moar ppl, MOAR PPL 14:01:59 #topic Agenda 14:02:02 o/ 14:02:07 we've a short agenda for today 14:02:11 let's get to it 14:02:13 #link https://etherpad.openstack.org/p/glance-team-meeting-agenda 14:02:26 #topic Updates Artifacts 14:02:37 ativelkov: mfedosin nikhil 14:02:40 :) 14:02:43 go crazy 14:02:44 o/ 14:02:47 o/ 14:02:58 So, a quick update 14:03:13 * nikhil stops typing 14:03:14 Thanks to jokke_ and nikhil we had some discusssions on v3 interoperability 14:03:31 And I think we've found a good design solution 14:03:48 great news 14:03:49 There will be a set of public API calls which are plugin-independnent 14:04:00 So, they will be "def-core" compliant 14:04:27 that's super important 14:04:35 is there a link/spec where we can read more about that? 14:04:53 And the plugin-dependent ("dynamic") APIs may be considered plugin, i.e. the ones intended for the particular project (murano, heat, magnum etc) to interact with glance internally )within a cloud) 14:05:09 flaper87: not yet, I am bad in finding time to document things ) 14:05:19 here's a start 14:05:20 Will do that in ML and later that article I am still trying to write 14:05:20 http://eavesdrop.openstack.org/meetings/glance_artifacts_sub_team/2015/glance_artifacts_sub_team.2015-10-05-14.01.html 14:05:23 ativelkov: aren't we all? aren't we? 14:05:25 :D 14:05:30 nikhil: thanks 14:05:37 #link https://etherpad.openstack.org/p/glance-team-meeting-agenda 14:05:40 #undo 14:05:42 Removing item from minutes: 14:05:44 #link http://eavesdrop.openstack.org/meetings/glance_artifacts_sub_team/2015/glance_artifacts_sub_team.2015-10-05-14.01.html 14:06:05 ativelkov: nikhil mfedosin question for you guys: Are you planning to have a session at the summit ? 14:06:16 I didn't see any proposal and it seemed weird 14:06:26 #link https://etherpad.openstack.org/p/mitaka-glance-summit-topics 14:06:28 proposals go there 14:06:31 :D 14:06:44 I would like to have one 14:06:45 Thanks :) I'd like to discuss the v3, yup 14:06:55 I guess that's a question for the session proposals topic. and personally think we should at least have a work session. 14:07:17 ok, please, add it to that etherpad as we're in the phase of setting up the schedule 14:07:19 it would be good to try to get some people from heat and magnum there, too ... want them to get ready to use artifacts 14:07:19 thanks for the reminder flaper87, yeah we'd ideally have something up already! 14:07:32 ++ 14:07:37 I'd like to hav ethe schedule done by the end of next week 14:07:54 rosmaita: sounds like a good idea 14:08:18 nikhil: ativelkov mfedosin could any of you sync with folks from those team to make sure we can pick a slot that works for at least 1 person of each team? 14:08:20 :) 14:08:28 I know, I know. Scheduling things is hard 14:08:43 I can work with someone from heat 14:08:53 nikhil: that's a start, thanks 14:08:57 and will ping you all if I can reach out to magnum in short time 14:09:11 ativelkov: mfedosin you guys surely know someone from Murano 14:09:13 nikhil: i can reach out to adrian otto 14:09:14 :) 14:09:20 flaper87: I am from murano ;) 14:09:32 And I'll ping Heat's PTL 14:09:42 ativelkov: I know, I forgot to tag that message 14:09:43 flaper87: so, we are looking at all the proposals listed by early next week? and final it by thursday or so? 14:10:12 anyone know how magnum plan to use artifacts? 14:10:33 nikhil: yes, that's the rough plan! If there are not enough proposals, we may come up with something or give those slots back 14:10:39 ok, lets move on 14:10:41 mclaren: I don't know for sure, but I remember Adrian Otto saying that he is looking forward to use it 14:10:41 mclaren: not sure, adrian otto was very interested, not sure why 14:11:04 #topic Updates drivers 14:11:32 mclaren: there were a few small discussions earlier in liberty, we need to come up to sync now once we've a direction for the API 14:11:41 On our last meeting, we discussed a couple of topics. One is the inheritable image properties spec in nova 14:11:42 * nikhil shuts up 14:11:47 ok, thanks 14:11:54 Another one is the spec-lite process 14:12:18 For that last one, I'll start a thread on the ML to reach out to more people and get feedback from others 14:12:40 We didn't reach consensus on that process and there's more to be discussed 14:12:57 Did I forget something important from the last meeting? 14:13:17 ok, I'll take that as no 14:13:20 moving on 14:13:24 #link http://eavesdrop.openstack.org/meetings/glance_drivers/2015/glance_drivers.2015-10-06-14.01.log.html#l-24 14:13:48 #topic duplicate exceptions between glance_store and glance bunting 14:14:13 * flaper87 wonders if that's the same topic we discussed last week (or we didn't discuss it?). Perhaps mclaren started it 14:14:16 bunting: floor is yours 14:14:19 When looking at the bug posted on the agenda I noticed that a few exceptions are duplicated between glance and glance_store 14:14:30 wow, that sounds familiar 14:14:31 I was wondering if we should 'merge' these exceptions 14:14:39 mfedosin: you worked on that, right? 14:14:45 yup 14:14:58 These are the exceptions that are currently duplicated: paste.openstack.org/show/475612/ 14:15:10 Are you sure we can merge them? 14:15:19 but finally we decided to duplicate handlers in the code 14:15:32 glance_store is an independent library 14:15:50 sometimes we just need to catch those exceptions in glance and do something or re-raise 14:15:54 #link paste.openstack.org/show/475612/ 14:15:59 mmh, perhaps Glance is the one that needs to be cleaned up 14:16:00 https://review.openstack.org/#/c/200000/ 14:16:12 so we have glance.ImageDataNotFound and glance_store.ImageDataNotFound. Any reason we can't just have glance_store ImageDataNotFound? 14:16:31 mclaren: I don't think there's anything preventing us from doing that 14:16:47 umm, wait a sec 14:16:51 flaper87, it doesn't work, because ImageNotFound can be raised either in glance or in glance_store 14:16:56 unless I misunderstood 14:17:00 so we have to catch them both 14:17:02 can those exceptions be used by cache logic and then re-worked? 14:17:22 mfedosin: ah right, that's the other thing 14:17:33 glance uses webob exceptions, any reason it can't use glance_store exceptions? 14:17:33 mfedosin: so what's the problem glance raising glance_store.ImageNotFound? 14:18:18 mmh, depends on when it is raising that exception 14:18:30 I think Glance raises ImageNotFound when it's not in the DB 14:18:46 perhaps we should spinn off new lib called glance_exceptions :P 14:18:47 in which case I don't think it should raise glance_store's exception 14:19:00 but my memory is bad and I need to read the code again 14:19:00 but not ImageNotFound - just NotFound. We have to fix it in million places 14:19:27 eh 14:19:37 NotFound sounds like a parent exception 14:19:47 common for different assets -- tasks, metadefs etc 14:20:04 members.. 14:20:12 bunting: I think the conclusion of the above is: ppl are not against you refactoring that code but there will likely be comments on the review (unless you find issues that won't allow you to move forward) 14:20:43 flaper87: Thats, fine its just pointed out its odd we have the same exception twice :p 14:20:49 yeah, it's probably a case of we shouldn't start from here :-) (ie with 29 duplicate exception names) 14:20:52 While I think that we can reduce the duplicate exceptions, I don't think we can get rid of all of them 14:21:04 mclaren: yea 14:21:06 nikhil: I agree 14:21:28 let's let bunting dig more into this 14:21:32 I think first question is do we actually raise them all from both? 14:21:39 it sounds like a good clean up 14:21:48 or how much that is still remains of the separation and just copy over 14:21:48 Let me explain 14:21:49 or attempt to 14:22:03 NotFound from here https://github.com/openstack/glance_store/blob/master/glance_store/exceptions.py#L66 14:22:24 is not exactly the same as NotFoundFrom here https://github.com/openstack/glance/blob/master/glance/common/exception.py 14:22:54 it's more identical to https://github.com/openstack/glance/blob/master/glance/common/exception.py#L334 14:23:02 yeah, I think ppl get that. The question is not w.r.t the base exceptions but the ones inheriting from those that are duplicated 14:23:14 so we can't use store's exception in Glance everywhere 14:23:50 sure, if they have different uses that's ok. 14:24:10 IMHo, glance shouldn't be raising glance_store exceptions. re-raising, fine. Raising from scratch, nope. 14:24:19 ++ 14:24:26 agree 14:24:29 and the only one solution we found was duplicate them in the code. it's dirty, but it works 14:24:44 we may be at a point where we may just have to live with some confustion around same names 14:25:03 like here https://github.com/openstack/glance/blob/master/glance/api/v2/images.py#L241 14:25:04 Let's let bunting dig into this if he's willing to 14:25:18 flaper87: I'm willing to have a look 14:25:30 I think the question is: if the glance store didn't exist would we have duplicated any particular exception (case by case)? 14:25:30 bunting: awesome! You just won something 14:25:34 I don't know what but something 14:25:53 whisky popcorn 14:26:02 a year's supply of exceptions 14:26:06 mclaren: we had store in glance and it was the same project so, no, we wouldn't 14:26:06 ++ 14:26:14 anyway, we gotta move on :D 14:26:34 #action bunting to dig into the "Attack of the cloned exceptions" 14:26:47 #topic Summit Schedule (flaper87) 14:27:01 I was there - it hurts :) 14:27:13 As hinted before, we gotta get the schedule done! 14:27:25 The allocations were uploaded to sched.org 14:27:28 #link http://mitakadesignsummit.sched.org/type/glance#.VhZhTc-1RZK 14:27:44 and I've used 2 already for topics that we must discuss for sure :D 14:28:04 1 is a fishbowl for the image import workflow refactor 14:28:19 and the second one (the last work room session) to finilize mitaka priorities 14:28:37 The rest of the slots we will all vote for and allocate rooms accordingly 14:28:46 please, do read the etherpad and add topics 14:28:54 oh the last work session is booked 14:28:54 #link https://etherpad.openstack.org/p/mitaka-glance-summit-topics 14:29:10 I probably meant latest 14:29:29 as in the one on Thursday at 17:20 14:29:46 * flaper87 can't change the topic of work sessions 14:29:47 :( 14:29:48 ack 14:30:07 yeah, it's on purpose to avoid large participation.. 14:30:12 yup 14:30:15 wonder if that's still case for tokyo :/ 14:30:26 as they aren't allowing a ton of people to design area 14:30:31 anyways.. 14:30:37 * nikhil shuts up 14:30:56 if there's something you would like to discuss, please add it to the etherpad 14:31:02 Last week we'll be closing the schedule 14:31:20 #info Glance's summit schedule will be finalized on Thursday 15 14:31:27 #undo 14:31:27 Removing item from minutes: 14:31:37 #info Glance's summit schedule will be finalized on Thursday October, 15th 14:32:03 questions? concerns? otherwise, we can move on 14:32:24 I'll take that as a no 14:32:30 #topic Cleanup old patches http://lists.openstack.org/pipermail/openstack-dev/2015-October/076232.html (flaper87) 14:32:45 I don't mean to go in deep details about this here because that's what the email thread is for 14:33:11 I just want to make sure people in this meeting know there's that discussion on the mailing list and that there's a workflow that is being planned to be followed 14:33:23 (at least once at the beginning of mitaka) 14:33:29 ++ 14:33:51 I half followed the thread 14:34:05 I found cases when people wanted to restore patches and couldn't themselves 14:34:07 that said, the plans to improve the review bandwidth are different from that thread and they involve other activities 14:34:19 Are we still planning on marking patches abandoned? 14:34:19 It would be best to have a POC for those who wish to do so 14:34:35 "points" of contact* 14:34:35 nikhil: is that something we should mention to infra? what cases specifically? 14:34:49 nikhil: ha! that clears POC in that context 14:34:52 :D 14:35:06 flaper87: cases when people were inactive for along time and wanted to restore a patch 14:35:14 let me get one latest example 14:35:44 mclaren: as I mentioned, I'd like to run this at least once for mitaka. We'll see how it goes. The feedback on the thread is great! 14:36:01 but really, lets keep the discussion on the ML 14:36:10 If a patch is still potentially ok (doesn't need a rebase, is passing the tests) will we mark it abandoned? 14:36:14 oops, ok 14:37:06 eh, can't find it 14:37:07 mclaren: I suggested that we should potentially try reviewing them instead of marking abandoned 14:37:09 mclaren: no, it'll be set back to not-in-work-in-progress. We can filter those out from the to-WIP list (think I mentioned this in the thread) 14:37:26 hmm, this one 14:37:29 #link https://review.openstack.org/#/c/206120/ 14:37:30 bunting: and I told you that's the goal :D 14:37:44 * mclaren will go back and read the thread 14:37:49 that's one of the points of this process 14:37:55 flaper87: Yeah, just mentioning as he said it half read it :p 14:37:58 I mean, identify those patches and move them forward 14:38:41 nikhil: thanks, I'll read through and try to understand why restoring it was not possible 14:39:06 perhaps is because of the -2? 14:39:08 * flaper87 shrugs 14:39:21 ok, please, chime in! This is important too. 14:39:35 moving on 14:39:50 #topic image import refactor is ready for discussion: https://review.openstack.org/#/c/232371/(rosmaita) 14:40:06 rosmaita: want to say something? or just a heads up? 14:40:30 just a heads up 14:40:42 That spec will likely require several iterations. It'd be great if ppl interested could read it and comment on it 14:40:48 thanks for that rosmaita 14:41:00 yes, i expect to be going through many patches 14:41:01 Things that will be discussed at the summit on that session depend on this spec 14:41:02 * jokke_ did first read through today 14:41:05 it's ok to comment on it now? 14:41:11 mclaren: def want your help, too 14:41:16 yes 14:41:22 thank rosmaita for takin the time to put it together 14:41:24 and it's a key part of Glance 14:41:27 * mclaren is chomping at the bit 14:41:56 And, of course, thanks rosmaita for taking the time to write this and leading the session. 14:42:04 * rosmaita blushes 14:42:07 you definitely will get whisky popcorn 14:42:11 yay! 14:42:13 +s 14:42:32 or a year's supply of exceptions 14:42:32 ok 14:42:43 rosmaita: you'll have to fight bunting for that 14:42:45 rosmaita: thats my prise ;) 14:42:51 HA! see? 14:42:53 :D 14:43:00 i imagine there's plenty to go around 14:43:20 ok, moving on 14:43:31 #topic Open Discussion 14:43:47 I think we'll have good attendance this year at the summit 14:43:50 that's awesome! 14:43:53 Lunch ? 14:44:02 ++ 14:44:02 I see that rosmaita proposed a cross-project topic here: 14:44:03 Or pop-corn party? 14:44:03 #link https://etherpad.openstack.org/p/mitaka-glance-summit-topics 14:44:12 is that approved yet rosmaita ? 14:44:21 line 83 14:44:38 cross-project discussion will happen next week 14:44:58 oh right, thanks 14:45:10 I have a small topic 14:45:13 Monday 20:00 UTC we'll discuss topics, Tuesday at the TC meeting the votes 14:45:13 TC needs to be formed completely.. 14:45:20 right 14:45:23 ativelkov: shoot 14:45:47 I heard bears are going hunting in Russia now 14:45:52 is that what you want to talk about? 14:45:55 nope ) 14:46:02 awwww :( 14:46:07 Do you know Openstack:Now newletter? 14:46:15 I wonder if we should have that topic irrespective of what TC decides.. in Glance if nowhere else.. 14:46:44 nikhil: oh, mmh, I thought that was implicit, tbh 14:46:51 ut it's good you brought it up 14:46:53 :) 14:46:58 I'd agree with that 14:47:08 we still have slots 14:47:28 Also, if there are free work sessions, I'd like us to use them to discuss the import workflow further 14:47:43 I've the feeling that fishbowl won't be enough 14:47:52 and a further face to face discussion would be useful 14:48:05 I see this page still have liberty etherpads 14:48:08 #link https://wiki.openstack.org/wiki/Design_Summit/Planning 14:48:12 has* 14:48:25 ativelkov: nope :/ 14:48:32 flaper87: totally. +1 on more time.. 14:48:58 Next week there will be some kind of a "panel" discussion (online, google hangouts, I guess) about the "community app catalog". And the organizers are looking for particpants 14:49:00 nikhil: yup, it hasn't been updated yet. Some projects are still figuring the schedules out, I guess that's probably why 14:49:02 * flaper87 shrugs 14:49:08 I see 14:49:40 ativelkov: where's the signup list? 14:50:04 There is no list, but they asked me to bring somebody from glance :) 14:50:05 I remember we wanting to have some discussion with them. 14:50:19 ativelkov: link? 14:50:22 I am willing to sign up. 14:50:32 if it's in a reasonable time, I might attend as well 14:50:56 There is no link (yet), they dodn't announce the time, as they are trying to gree with the particpants 14:51:11 ativelkov: ok 14:51:23 let folks know in -glance when you have that info :D 14:51:35 They are trying to invite the present and fromer Heat's PTLs - so, if flaper87 and nikhil can join, that would be awesome 14:51:49 ativelkov: I'm happy to 14:51:58 just need to make sure it's doable time-wise :D 14:52:01 Also there will be somebody from Murano - and the PTL of the comunity app catalog, of course 14:52:07 ativelkov: I can def join. Please include me in the emails, lists, links :) 14:52:15 sure 14:52:20 nikhil: you missed post-cards 14:52:22 ativelkov: someone from foundation? 14:52:30 donno yet 14:52:32 flaper87: ha ! ye...a 14:52:44 I see 14:52:49 It's mostly a PR initiave, they are trying to bring more attentioan to the app catalog 14:53:11 But at the same time it may be a good chance to discuss the technical details 14:53:16 hmm, most likely another keynote prep :) 14:53:24 yes 14:54:01 So, I'll send an email and give your contacts to organizers 14:54:43 awesome 14:54:45 thanks 14:54:49 anything else? Anyone? 14:54:58 https://review.openstack.org/#/c/224991/ 14:55:00 * nikhil ack 14:55:08 I'll figure out a day for our lunch in tokyo 14:55:28 we did a picnic in Vancouver, I don't think a picnic would be doable in Tokyo 14:55:30 :P 14:55:33 anyway 14:55:38 lucnh sounds cool 14:55:38 TravT: thanks 14:55:47 quick poll ... when people are arriving and leaving? 14:55:57 25th-31st 14:56:13 I'll be in tokyo Sun evening or Mon morning and will have flight back at Sun morning 1st 14:56:14 25th-31st 14:56:22 TravT: most likely affects horizon directly? 14:56:38 Yes, people were asking about it in the room over night 14:56:42 hoping to get a review on it 14:57:09 sounds straightforward, I can take a quick peek 14:57:18 cool, thx 14:57:20 jokke_: sun->sat 14:57:32 26th - 15th :) 14:57:32 the same 14:57:44 ativelkov, not with you =P 14:57:45 flaper87: you going out of japan in the meantime? 14:57:47 So Glance Drinks at Mon evening? 14:57:51 nope 14:58:02 but I'll be back in tokyo on the 25th 14:58:16 gotcha 14:58:39 nikhil: when are you getting in? 14:58:52 * nikhil was doing a timezone check 14:59:04 flaper87: sun evening 14:59:15 AWESOME! 14:59:15 25th ie 14:59:27 okidoki 14:59:30 ran out of time! 14:59:34 thanks folks 14:59:37 #endmeeting