14:00:40 #startmeeting glance 14:00:42 Meeting started Thu Apr 14 14:00:40 2016 UTC and is due to finish in 60 minutes. The chair is nikhil. Information about MeetBot at http://wiki.debian.org/MeetBot. 14:00:43 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 14:00:46 The meeting name has been set to 'glance' 14:00:48 #topic agenda 14:00:49 specially the ones that has coffee and buns 14:00:57 :) 14:00:59 #link https://etherpad.openstack.org/p/glance-team-meeting-agenda 14:00:59 o/ 14:01:10 o/ 14:01:23 Looks like the agenda filled up pretty recently 14:01:28 o/ 14:01:33 we should need the full hour more or less 14:01:42 #topic Update from other teams 14:01:53 #info Glare updates ( mfedosin ) 14:01:59 hello! 14:02:09 no big changes during last week 14:02:24 we stabilize our code and preparing to a big demo 14:02:41 kairat added tooz support to prevent race conditions 14:02:56 that's great news! 14:03:03 dshakhray implemented filtering and sorting support for artifacts 14:03:35 all desired features are imlemented so I'm going to concentrate on the spec next week 14:03:56 kudos 14:04:19 Thanks mfedosin 14:04:29 welcome :) 14:04:35 moving on 14:04:36 #info Nova v1 v2 14:04:45 mfedosin: that's mostly you 14:04:49 I updated the spec 14:05:06 #link https://review.openstack.org/#/c/301741/4 14:05:20 I think it's ready to merge 14:05:27 :) 14:05:56 also I fixed race conditions in commits and waiting John to remove -2 14:06:13 not race conditions, sorry 14:06:17 merge conflicts 14:06:55 ok 14:06:59 headsup: I read some of the latest comments from matt and sean on ML that indicate we need to sync and decide on glanceclient scope for this spec 14:07:03 #link http://lists.openstack.org/pipermail/openstack-dev/2016-April/092153.html 14:07:05 nikhil: may I ask you about summit session? 14:07:14 #link http://lists.openstack.org/pipermail/openstack-dev/2016-April/092158.html 14:07:16 mfedosin: sure 14:07:24 it will be combined with Nova team, right? 14:07:46 it's officially nova session that is tagged for glance too 14:08:07 when will it be? 14:08:12 do you remember? 14:08:16 so, we can expect turn-out from both teams 14:08:24 all glance tagged sessions are here: 14:08:34 #info glance tagged summit sessions https://www.openstack.org/summit/austin-2016/summit-schedule/global-search?t=glance 14:08:56 #info nova session with glance on image v2 https://www.openstack.org/summit/austin-2016/summit-schedule/events/9083?goback=1 14:09:21 anything else? 14:09:36 no, thanks :) 14:09:57 yw 14:10:14 #info cross-project updates 14:10:24 there was no meeting and nothing raised this week 14:10:39 as usual, do visit the community blog post 14:10:53 the cross project session list is up in schedule 14:11:23 I have requested a few signups for cross project session attendees here: 14:11:25 #link https://etherpad.openstack.org/p/newton-glance-summit-planning 14:11:34 line 345 onwards 14:11:57 basically we should try to cover as many sessions as possible as a team 14:12:24 for someone signed up on two sessions at the same time, please add your name instead if you are free then 14:12:50 I wanted to correct some information that I provided during the sync 14:13:10 tuesday full day design summit is CP but there are still main summit tracks that run parallel to this 14:13:43 so in case you _do_ have signed up for a session, you can re-think about your priority towards main summit session too if that's relevant 14:14:07 another note: all sessions marked as in the etherpad (**) are important for us 14:14:11 that's it 14:14:34 #topic announcements 14:14:51 wanted to get a quick poll on the next two meetings 14:15:16 #startvote have the meeting on april 21 (next week)? yes, no, maybe 14:15:16 Begin voting on: have the meeting on april 21 (next week)? Valid vote options are yes, no, maybe. 14:15:17 Vote using '#vote OPTION'. Only your last vote counts. 14:15:50 I will take no answer as "NO" (as I don't have agenda) 14:15:58 #vote no 14:16:08 counting 1 14:16:12 counting 2 14:16:13 #vote no 14:16:14 #vote maybe 14:16:15 #vote maybe 14:16:18 #vote maybe 14:16:24 #vote maybe 14:16:33 mclaren: maybe we can have import meeting in this timeslot if no reg meeting 14:16:59 #endvote 14:16:59 #vote yes 14:17:00 Voted on "have the meeting on april 21 (next week)?" Results are 14:17:01 maybe (4): avarner, abhishek, mclaren, rosmaita 14:17:02 no (2): hemanthm, nikhil 14:17:09 rosmaita: fine by me 14:17:11 ok, we will see 14:17:25 I will update on wednesday if agenda is updated or not 14:17:59 #info We won't have the weekly meeting during summit ie on April 28th 14:18:14 Thanks for your input 14:18:18 moving on 14:18:34 #topic descriptions of the summit sessions 14:18:57 firstly, find the list of all glance sessions: 14:19:02 #link https://www.openstack.org/summit/austin-2016/summit-schedule/global-search?t=Glance%3A 14:19:32 #info the above list refers to design summit Glance sessions 14:20:02 and that query doesn't show sessions tagged by other teams 14:20:34 also, we need to update the descriptions of the Fishbowl sessions specifically 14:20:58 I took the liberty to copy over data from the summit planning etherpad and modified a little bit (mostly grammar) 14:21:26 I need input from you so that the sessions cover the following items in the description (copying from my email): 14:21:43 what's the problem it's trying to solve 14:21:50 what's currently out there 14:21:55 what are best practices, feedback and experiences from cross-prj or other teams 14:21:59 what is the scope of the current proposal or are you posing questions via. this topic 14:22:03 what has been discussed, what are/were the blockers and any other reference information 14:22:25 We need a good chunk of that info up in the descriptions so that people know what they are about 14:22:44 Although, this rule also applies to work sessions, I am insisting on FB ones first 14:23:03 if I don't get input, I will take the liberty to update them myself as per my perspective 14:23:15 questions? 14:23:30 timeframe for completion? 14:23:55 the cut off time from summit planners is EOW (this week) 14:24:12 so, any time by 7pm ET tomorrow should work 14:24:20 thanks 14:24:34 moving on 14:24:39 #topic Mid-cycle 14:25:18 I've announced this on the ML, just bringing it here so that it's not lost in your inbox. I noticed that a few people missed my ML updates. 14:25:25 #link http://lists.openstack.org/pipermail/openstack-dev/2016-April/092009.html 14:26:10 could we maybe get a list of people who are interested in remote attendance? 14:26:20 The etherpad (wiki) is in the email. Please start planning your travel as soon as possible. I know it's a bit hectic period, close to summit, but after the summit may delay things. 14:26:22 (it suddenly looks like i will be one of those people) 14:27:02 rosmaita: sure, please feel free to create a section for remote attendance or add remote, maybe remote, etc. tags against your RSVP name on the etherpad 14:28:20 #action all: please RSVP your confirm or tentative plans for the mid-cycle on https://etherpad.openstack.org/p/newton-glance-midcycle-meetup . It would be good to get a list of remote attendance. 14:28:51 questions? 14:29:12 moving on 14:29:20 #topic Provide a list of request ids to the caller (abhishekk) 14:29:26 hi, https://blueprints.launchpad.net/python-glanceclient/+spec/return-request-id-to-caller was accepted for mitaka and needs to be changed to newton 14:29:57 and https://review.openstack.org/#/c/261288/ is up for quite long time, needs review 14:30:23 abhishek: have you proposed change for that move in the specs repo? 14:30:35 or someone else? 14:30:58 jokke_: specs was not submitted for this as it is approved in cross-project specs 14:31:31 then as per your suggestion, I have created a specs-lite last time 14:31:44 oh cool 14:32:07 jokke_: please refer, https://launchpad.net/bugs/1525259 14:32:18 abhishek: Error: Could not gather data from Launchpad for bug #1525259 (https://launchpad.net/bugs/1525259). The error has been logged 14:32:18 just something we can track off so if there is spec lite (as in wishlist bug) we're tracking it 14:32:38 abhishek: oh nice, might be a good idea to quickly link that lite spec with the bp (as a URL for specification) 14:32:41 abhishek: as we agreed, there is no need to transfer that to the new lite-spec format for this cycle 14:33:04 nikhil: i will do that 14:33:07 thanks for bringing this up abhishek! 14:33:10 jokke_: ok 14:33:45 thank you, thats it from my side 14:33:45 I will talk about reviews and review day in the later in the meeting (impromptu addition) 14:33:45 * jokke_ really needs to fix the lite-spec documentation 14:34:09 we need to talk about that, prolly during contri meetup last session 14:34:32 anyway, let's give enough time for import :) 14:34:36 moving on 14:34:42 #topic Notes on image import implementation tactics (mclaren) 14:35:03 mclaren: thanks for the link 14:35:05 #link https://etherpad.openstack.org/p/glance-image-import-implementation-tactics 14:35:08 Sure 14:35:21 firstly, how is this different from what we want to discuss at the summit 14:35:27 This is just a bit of a brain dump really 14:35:55 ah I see, thanks 14:36:03 But having thought it through it does look like we can avoid some complex stuff 14:36:09 i think it's a good clarification 14:36:29 eg having to store two image blobs at once 14:37:13 I'd be interested if folks wanted to read through and see if they generally agreed/spotted any holes 14:37:55 (eg drop me a mail) 14:38:21 mclaren: I think all are browsing through it atm 14:38:24 oops, i already started putting comments right on the etherpad 14:38:39 oh yeah, that works too! 14:38:57 mclaren: It's pretty detailed and explicitly written. Only have my love for it atm! 14:39:22 in fact, I was planning to start some tiny discussions around this 14:39:47 may be we go ahead and start defining (formally in a spec) what we want to target for newton 14:40:02 Thanks. If we really can punt on some of the complex stuff it does make getting something that 'works' sooner possible 14:40:39 agreed. 14:41:04 I'll check for feedback on the ether pad later. Thanks, that's all I had. 14:41:43 looks like a excellent plan 14:42:20 in the meanwhile: I will try to gauge interest, internal commitments, preferences, etc. and plan a timeline around this 14:43:17 And then use the plan your poise along with developer/reviewer time to sketch out the remnanats 14:43:21 jokke_: is that you putting comments in purple? 14:44:07 rosmaita: yup 14:44:36 rosmaita: I had a feeling that defcore would be lenient on the timeline and not need 3 cycles? 14:44:58 well, i'm not sure the timeline is really the issue 14:45:17 as far as defcore is concerned, they would like glance to have a single API for end-users to get images into the cloud 14:45:35 the API can be "flexible" as outlined on the spec 14:45:50 but we need to hit the widest use cases first, as stuart says 14:46:04 widest == most popular ? 14:46:26 not that it matters, but I wasn't really thinking about widest use cases, more minimally scoped code. 14:46:28 i guess so, measured in number of clouds, not number of users 14:46:46 I see 14:46:52 thanks rosmaita 14:47:08 mclaren: i'm thinking of the point you made in the pre-summit conference aobut even a physics dept cloud should use image import, not direct upload 14:47:18 even though all users are smart and trusted 14:47:35 that way, it's same interface across all clouds 14:47:42 sure 14:47:57 so to make that happen, i think we need to go with stuart's plan for glance-direct first implementation 14:48:14 (much as i'd rather do the swift-based one) 14:48:20 rosmaita: then we aim for a sociology dept? ;-) 14:48:21 yes, and I doubt if there's any other way 14:48:25 LOL 14:48:42 mclaren: not sure i can handle postmodern virtual machine images 14:48:46 lol 14:48:57 rosmaita: yeah, I doubt if this will be MVP (or considered MVP by some) unless we implement swift based one too 14:49:19 anyway, we've a skeleton to talk through at the summit I guess 14:49:22 it being that main reason for this refactor 14:49:50 how about we start working on a spec for this before the summit so that we know what we are discussing on? 14:50:01 ummm... i think we have a spec 14:50:06 approved, and everything 14:50:08 rosmaita: ++ 14:50:13 well it's a direction 14:50:22 spec actually says glance-direct will be done first 14:50:31 we still don't have implementation details and know what are our constraints 14:50:50 for eg. what are the operator impacts with this import way vs. regular way 14:50:57 not sure a spec is needed -- let's discuss at summit? 14:50:59 (ref. staging area) 14:51:50 sure, I don't mind having this in a etherpad (no spec in that case) just need to know what, when, how for newton 14:51:56 my feeling is let's work on stuart's etherpad (within the constraints of the approved spec) before hte summit, then at the summit we'll be in a good position to assign work items and realistically see what can be deliverd in newton 14:53:06 because, I can't be really tentative in my communications of 1) proposer: we have this feature for newton, will you be able to get that in ? -- ans (nikhil): I don't know, I can't say 2) proposer: when will you know ? nikhil: I don't know 14:54:06 as long as we agree to agree (and not agree to disagree or more) at the summit, it's fine with me 14:54:30 well, we will know more at the summit ... i mean, the api is set, no more arguing about that, we (me & stuart) can work on implementation details between now and hte summit 14:54:59 rosmaita: cool, that sounds good. I'd love to be involved in the implementation details too. 14:55:08 let me move on for now 14:55:20 #topic open discussion 14:55:21 oh, sure, don't mean to exclude anyone ... everyone is welcome to join in the mess! 14:55:53 folks, do you know about an issue in glance v1? 14:55:56 rosmaita: I think it's a fun challenging work item! Can't wait to get my hands on (some or more) of the code for this!!!!! 14:56:05 mfedosin: plenty :) 14:56:15 you can't sort by public images there 14:56:22 mfedosin: well, it's glance v1 for first :P 14:56:26 mfedosin: how do you mean? 14:56:29 This will be my last weekly meeting, due to a change in job responsibilities. I'm no longer working on glance, although I do plan to finish the items assigned to me. 14:56:37 if you provide is_public=true then it list all images 14:56:49 heh, give me one second, there's a link for that 14:56:58 darja told me today, but I haven't checked yet 14:57:03 avarner: Thanks Drew, for letting us know! I'd like to sync offline a bit more on that with you. 14:57:12 avarner: sorry to hear, are you staying withing OpenStack still, and thanks for everything you have done! 14:58:08 mfedosin: http://docs.openstack.org/developer/glance/glanceapi.html Note Use of the is_public parameter is restricted to admin users. For all other users it will be ignored. 14:58:44 jokke_, I'll be working on container stuff in IBM, probably not much Openstack 14:58:45 mclaren: ah, cool! thank you 14:58:50 btw, do people need email sent from me endorsing the utmost necessity for your presence at the mid-cycle to help you talk with the managers? (this is serious question) 14:59:07 avarner: well best of luck then for your future adventures 14:59:14 nikhil: yes for me, don't know if it will have any effect, though 14:59:24 mfedosin: the v1 is_public filter was always kinda broken 14:59:24 jokke_, thanks! 14:59:27 rosmaita: it doesn't hurt to try! 14:59:28 :) 14:59:28 nikhil: won't change a thing 14:59:39 jokke_: oopsie :/ 14:59:51 "Use of the is_public parameter is restricted to admin users" but it doesn't work even for admins now 15:00:19 mfedosin: we've had to disable image-list for admins by policy in our cloud 15:00:26 yeah 15:00:28 too many images 15:00:32 we had to disable v1 15:00:43 yeah -> is_public filtering is tricky indeed 15:00:58 we're out of time 15:01:04 mfedosin: hmm, might be a bug so? 15:01:04 Thanks all for joining! 15:01:07 thanks! 15:01:15 nova uses image_type to distinguish base images from snapshots 15:01:33 will check in a moment 15:01:35 guys, I think SL team is waiting for us to finish 15:01:41 let's sync on -glance 15:01:45 thanks again 15:01:48 #endmeeting