14:04:03 #startmeeting glance 14:04:04 Meeting started Thu Jan 16 14:04:03 2014 UTC and is due to finish in 60 minutes. The chair is markwash. Information about MeetBot at http://wiki.debian.org/MeetBot. 14:04:05 good morning! 14:04:05 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 14:04:07 The meeting name has been set to 'glance' 14:04:37 good morning 14:05:39 #link https://etherpad.openstack.org/p/glance-team-meeting-agenda 14:05:44 bit light today 14:06:03 it's ok if we end early 14:06:13 indeed 14:06:25 though it would have been better to start later! 14:06:44 heh perhaps 14:07:01 did you get a chance to think about exposing 'owner' in v2? 14:07:02 #topic expose 'owner' in v2 (rosmaita) 14:07:32 when I thought about it explicitly I couldn't really think of a reason why we should hide it 14:07:41 ok 14:07:43 and just mulling on it didn't yield any other reasons for concern 14:08:01 me neither, plus i think it will be needed anyway when we kill off v1 14:08:02 so it seems fine to move ahead with it 14:08:28 this seems like a simple BP, "expose 'owner' in image response" 14:08:40 and probably simple code as well 14:08:45 but i'm not sure if we were thinking about a new process? 14:08:59 i agree, i could possibly code it myself :) 14:09:22 I don't quite see how a new process would be necessary 14:09:32 me neither, thought I'd check though 14:09:48 well I think we're in good shape on this one, no one else objects? 14:10:53 sounds good to me 14:11:03 cool 14:11:15 so we just need to change api controller ? or ned to change domain also? 14:11:24 s/ned/need 14:11:54 I think owner is already an attribute of the image object 14:12:11 yes, it's already in domain object 14:12:18 #topic deprecated options in glanceclient (rosmaita) 14:12:30 (hope switch wasn't too soon) 14:12:33 this is a minor thing 14:12:49 just from a user perspective 14:13:22 ever since i've been working on glance, theCLI help gives you a long list of stuff, mostly DEPRECATED 14:13:30 ok got it (seems my network today has bigger delay, sorry) 14:13:36 was wondering if we could clean that up a bit 14:13:42 (since it's a light week) 14:13:58 but, i'm not real clear on the client contract and versioning 14:14:11 rosmaita: I think its a good point, I was hoping we could do that in the major version bump 14:14:22 so that would be with Icehouse? 14:14:47 but it seemed like we never got anywhere with folks about how we can *do* the major version bump 14:14:59 yeah, that was my recollection, too 14:15:09 maybe we can discuss at mini-summit? 14:15:15 yes 14:15:35 iccha had mentioned that maybe we should have a "stability & cleanup" session 14:16:10 i think she's worried that with all the sexy new metadata catalog stuff, normal stability issues might get lost 14:16:47 that's a fine idea as well 14:17:31 I have added both of those ideas as potential topics on the mini summit agenda 14:17:37 i'll add to the agenda ... maybe sign up everyone as interested by default? :P 14:18:02 network delay! 14:18:23 rosmaita: I'd like to work with some folks to move to the next level on the mini summit agenda, from ideas to more like plans 14:18:52 are there some folks on your end that could help me out? I imagine ashwini would be interested, but perhaps others as well? 14:19:32 i think everyone will be interested in helping out 14:20:04 markwash: interesting, which part ? :P 14:20:13 maybe we can ask people to "adopt" a session and map it out with more details 14:21:09 yes, I think we need to announce a deadline for finalizing the topics perhaps, and if a topic doesn't have a leader/moderator by that time we figure we can hash it out on the ML or in informal time? 14:21:44 rosmaita: will you discuss the image marketplace on the mini summit? 14:21:57 yeah, i was thinking it would be cool if each session got a ML announcment so people who can't attend can reply to the thread and have their views represented 14:22:09 oh neat idea 14:22:11 flwang: hi , yes i was thinking of doing that one 14:22:17 before or after the summit? 14:22:25 before, if we have time 14:22:28 markwash: before 14:22:41 sounds good 14:22:42 markwash: after confirm the topic list? 14:22:42 but it is running out fast! 14:23:17 we probably have about enough if we set up a timetable pretty immediately 14:23:43 rosmaita: I love the image marketplace idea, since both AWS and Nebula have it 14:23:44 how about if we get people to pick their topics by monday and send out email introductions by tuesday? 14:24:00 that seems reasonable 14:24:27 *OpenNebula I think 14:24:32 i think you can ask anyone on the titan team to pick up an orphan topic 14:25:01 does 12 topics total seem okay? about 6/day? 14:25:07 would 5/day be too few? 14:25:47 (I just counted and we have 14 suggestions on the list so far, not all of them adopted of course) 14:25:56 markwash: yep :) thanks for the correcting 14:26:13 rosmaita: okay thanks, we may need to do that 14:26:19 that's a tough question 14:26:46 6 topics would be 6 hours in-session + lunch + coffee == 8 hours 14:26:57 that would leave time for disucssion after dinner 14:27:12 to follow up on that day's stuff 14:27:22 except that wouldn't work on 2nd day 14:27:31 okay, so on that schedule 6 doesn't seem too skimpy, and even 5 might be okay 14:27:42 so perhaps if its an odd number or low, we should frontload the schedule 14:27:42 i think so 14:27:50 +1 to frontloading 14:28:29 i definitely don't think it's too light 14:28:39 yeah, i think our open discussion time will be better if its after the formal-ish sessions 14:28:49 +1 14:29:19 okay cool, glad it doesn't seem too light 14:29:34 I'll be happy then if we do a solid 5/5 or 6/4 14:29:44 #topic open discussion 14:29:50 (since we're on it anyway :-) ) 14:30:29 anyone particularly concerned about their i-2 blueprints? I think we might end up deferring a lot 14:30:48 but we have made good progress lately and have room in i-3 so I'm not very worried 14:31:05 markwash: do you think it's a good time to think about giving LocationRepo to glance v2 domain? 14:31:23 zhiyan: yes, I think its about time 14:31:51 zhiyan: with image location status, it will be nicer to have some domain code to handle the possible status values 14:32:19 and I realize from looking at the db code we really need to allow the clear distinction in the http api between add, replace, and remove to be expressed at the db api level 14:32:25 markwash: or do it in J? 14:32:41 zhiyan: I was trying to work on it last night but couldn't quite get my brain to thaw :-) 14:32:56 I'll revisit it today/tomorrow/this weekend 14:33:29 shouldn't take long once it seems a bit more clear, which usually only takes a few days of mulling 14:33:30 zhiyan: i have not been following closely, we are adding 'status' to locations, but not 'checksum' I hope? 14:33:46 markwash: ah, cool 14:33:52 rosmaita: just status, and its basically an internal field 14:34:05 rosmaita: yes, not yet 14:34:09 markwash: I think I may have to defer the image-recover 14:34:13 yes 14:34:21 rosmaita: the idea is that we can track deletion and "pending_deletion" on the locations themselves, which makes scrubber and other things cleaner 14:34:34 cool 14:34:59 flwang: noted, thanks for the heads up 14:35:27 maybe this can be part of the "stability" session, but before we branch out too much with the metadata catalog stuff, i was wondering if we need to record a glance "philosophy" about what it's core principles are 14:35:42 markwash: since it's depending on zhiyan's locations status bp 14:35:45 AFAIK, right now, it's just folklore 14:35:49 markwash: i believei location_repo will help more, when we try to add multi-hypervisor deployment support right? (like checksum, size, format ...) 14:36:06 yeah, that's a good thought, something I think iccha had mentioned was maybe a glossary 14:36:29 rosmaita: I sense you are a little concerned about the "metadata" repo stuff. . I somehow missed out on the december conversation that was had around that 14:36:43 zhiyan: yes I think there are a lot of avenues there 14:37:08 markwash: well, slightly concerned 14:37:25 though we end up just having to figure out how we make it clear what those mean on a location vs on the image, but I think its do-able 14:37:44 rosmaita: I confess, I think I'm a bit against the idea of storing metadata generically 14:37:49 i think it's a great idea, and will enhance searching and allow relationships and stuff, but i'd still like glance to be glance 14:37:58 so i am kind of mixed 14:38:14 it took a long while to get here, but I think we get a lot of benefit from knowing what our metadata means 14:39:15 okay, I think we're about tapped out (or I've netsplit :-) ) 14:40:00 any other parting thoughts? I will email ashwini to set up a time for more mini summit planning discussion 14:40:34 nothing from me 14:40:59 i'm clear 14:41:18 zhiyan, flwang, rosmaita thanks! 14:41:25 zhiyan, flwang keep up the great reviews! 14:41:30 markwash: nothing from me too. btw, you can ping me for anything around location_repo (actually i'm thinking if i can do that) 14:41:38 markwash: and coding !!! 14:41:52 heh yes and coding! 14:41:52 :P 14:41:54 :-) 14:42:19 zhiyan: sure thing 14:42:47 bye! 14:42:54 #endmeeting