14:00:36 #startmeeting Glance 14:00:36 Meeting started Thu Dec 17 14:00:36 2015 UTC and is due to finish in 60 minutes. The chair is flaper87. Information about MeetBot at http://wiki.debian.org/MeetBot. 14:00:37 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 14:00:37 o/ 14:00:39 The meeting name has been set to 'glance' 14:00:42 o/ 14:00:44 #chair rosmaita 14:00:45 Current chairs: flaper87 rosmaita 14:00:46 o/ 14:00:46 #chair nikhil 14:00:47 Current chairs: flaper87 nikhil rosmaita 14:00:52 o/ 14:00:52 just in case my connection drops 14:00:54 o/ 14:01:01 ack 14:01:04 ok 14:01:17 #topic Agenda 14:01:19 o/ 14:01:19 #link https://etherpad.openstack.org/p/glance-team-meeting-agenda 14:01:37 o/ 14:01:37 As usual, that's the place for our agenda. It's not very long and topics should be easy to go through 14:01:45 o/ 14:01:53 #topic Updates from other teams (flaper87) 14:02:00 Artifacts 14:02:05 nikhil: ativelkov mfedosin 14:02:42 so, we have patches for moving v3 -> v0.1 waiting for review 14:02:46 #link https://review.openstack.org/#/q/topic:bp/move-v3-to-glare 14:02:59 spec about deprecating v3 api is merged 14:03:12 w0000h0000000 14:03:17 Darja made a patch for Glare client 14:03:21 glance patch has -1 from jenkins, but that's because of alembic gate issue 14:03:39 ativelkov: so we need a fix for alembic then too? 14:03:40 I saw Darja's patch and asked to wait for the server-side of the work to land first 14:03:47 o/ 14:03:49 I don't think g-r is going to be blacklisting 0.8.4 14:04:04 flaper87: yeah, it has a depends-on ref there 14:04:05 Has anyone looked at the nova patch to see what we can steal? 14:04:12 I think mfedosin had a fix for that, right? 14:04:14 ativelkov: ++ 14:04:22 mfedosin was looking at it 14:04:24 sigmavirus24: mfedosin was on top of that yday 14:04:27 lol 14:04:28 we'll get to it 14:04:33 Nevermind 14:04:34 lets stay on topic 14:04:42 Sorry 14:04:49 * flaper87 hugs sigmavirus24 14:04:57 ok, anything else? 14:04:59 it doesn't help 14:05:21 Public API spec is also on review, waiting for some comments 14:05:31 https://review.openstack.org/#/c/254710/ 14:05:44 ativelkov: yeah, I'll bring that one up on the next drivers meeting 14:05:52 size and virtual_size still don't pass if you are about model sync 14:06:15 And we wanted to have a voice/video call instead of the next artifact IRC meeting to discuss artifacts roadmap in more details 14:06:34 That'd be on monday, right? 14:06:39 I mean 14:06:41 next monday 14:06:49 * flaper87 knows the meeting happens on mondays 14:07:03 Yup, but that's not final: nikhil promised to create some doodle to pick a slot convinient for all 14:07:27 ack 14:07:42 What about doing it the first week of January? I'll be around next monday but not the one after next 14:07:52 That'd be, I'll be around on 21st but not on 28 14:08:07 actually, I lied and my GF would kill me for saying I'll be around on the 21st 14:08:09 1st week of january is a holiday season in Russia. Unsure about my availability at that time 14:08:10 :P 14:08:15 I am fighting with my own schedule atm, but I hope that this happens on 21st 14:08:31 hmm 14:08:46 what time is the artifacts meeting? 14:08:49 is tomorrow too early? 14:08:49 What about an ad-hoc one on Tuesday? 14:08:57 rosmaita: same as drivers 14:09:04 I'll be around tomorrow 14:09:07 1400UTC 14:09:24 ativelkov: do you think app catalog folks can join tomorrow? 14:09:47 nikhil: don't know. They have their IRC meeting later today, I can ask 14:09:49 ok, lets take this offline and announce it properly on os-dev 14:09:54 sg 14:10:15 I'd suggest direct IRC pings to organize this as a doodle might take more time for ppl to vote on 14:10:43 ok, moving on 14:10:52 Drivers updates: 14:10:59 The deprecation for v3 landed 14:11:15 "removal of experimental api" 14:11:21 There are a couple of other specs that are almost approved: 'in' operator and the image signing enhancements 14:11:26 rosmaita: ++ 14:11:31 * flaper87 should have links handy 14:11:53 sigmavirus24: I added you to the image signing spec. It'd be great to have your vote there 14:12:02 flaper87: will look at it :) 14:12:29 sigmavirus24: awesome, thanks! If you can do that today, that'd be even better. I'd like to merge it this week if possible 14:12:33 and assuming everything is ok 14:12:39 flaper87: roger that 14:12:47 I wonder: the "removal of experimental api" does not say any details "the new home" of that API, i.e. glare v0.1. Do we need a new spec for that, or that one will work? 14:13:05 ativelkov: we can amend that one 14:13:10 imho 14:13:19 no need for a whole new spec, lets just edit the one that landed 14:13:38 ok 14:13:52 ok, am I missing something? any update from other drivers? 14:14:09 (btw, we're working on finalizing the image import process spec, despite all the open questions there seem to be) 14:14:21 what flaper87 said 14:14:37 when is the spec freeze? 14:14:48 ativelkov: on the 28th 14:14:55 actually, LOL, I have that as a sub-topic 14:14:58 T_T 14:15:07 ok, on the week of the 28th, there'll be a spec freeze 14:15:23 We'll be granting exceptions just for very special cases if there are any 14:15:47 so, please, try to update your spec and reach out before the freeze 14:15:52 :) 14:16:02 and lite specs aren't included in the freeze 14:16:04 right? 14:16:09 nope 14:16:21 just adding to the info :) 14:16:35 #info lite specs are not included in the spec freeze 14:16:46 #info Spec Freeze is in the week of Dec 28th 14:16:48 :D 14:16:55 ok, moving on 14:17:00 Nova v1 -> v2 updates 14:17:02 mfedosin: 14:17:23 hey 14:17:34 sorry, I'm working on it right now 14:17:44 I mean in this moment 14:17:52 create and update is almost done 14:18:00 shouldn't you be in a glance meeting? or something like that 14:18:02 :P 14:18:11 mfedosin: wait, remember that Cyril is doing create 14:18:22 mfedosin: what about the xenserver plugin ? 14:18:24 they are the same 14:18:31 create and update 14:18:45 mfedosin: gotcha, I'll try to ping cyril then 14:18:45 they both use the same method _translate_to_glance 14:18:58 so all fixes are there 14:19:30 I'm afraid about copy-from 14:19:41 v2 doesn't support it 14:20:03 mfedosin: mmh, I guess nova exposes it 14:20:20 mfedosin: let's figure that out later. What about the xenserver plugin? 14:20:32 it means that we have to fix tempest tests 14:20:56 I'm going to make xenplugin to use nova.image 14:21:09 I think it's the best solution 14:21:34 when create and update for nova.image is done I'm going to start this 14:21:44 mfedosin: awesome, if possible, make that a priority for the Nova v1->v2 work. Some folks in nova want to see that patch before moving the rest forward 14:21:50 mfedosin: ++ 14:22:08 because it's awful https://github.com/openstack/nova/blob/master/plugins/xenserver/xenapi/etc/xapi.d/plugins/glance#L117 14:22:20 it's HORRIBLE! 14:22:26 * flaper87 suffered because of that 14:22:39 mfedosin: I don't recall if I made it use nova.imges, to be honest 14:22:42 I can lookup that patch 14:22:46 ok, anything else? 14:23:02 nothing from my side 14:23:23 ok 14:23:25 #topic Holidays break (flaper87) 14:23:42 so, I think this will be our last meeting until 2016. 14:23:48 * Steap did the create patch, gotta send it 14:23:51 * Steap == Cyril 14:23:55 Next week is Christmas eve and then new year's eve 14:24:07 mfedosin meet Steap 14:24:24 so, unless someone thinks we should totally meet on either of those days, I'll send an email out notifying this 14:24:54 +1 not meeting 14:24:55 hi Steap! 14:25:05 For communications, discussions, etc, you can use the mailing list 14:25:14 I'll be monitoring it and hopefully other folks will as well 14:25:38 ok, no one screamed, I'll take that as a +1 for skipping the next 2 meetings 14:25:50 #topic glance_store drivers maintainers still required (flaper87) 14:25:59 #link https://review.openstack.org/#/c/258237/ 14:26:11 I started that patch based on what's in https://etherpad.openstack.org/p/glance-store-drivers-status 14:26:14 #link https://etherpad.openstack.org/p/glance-store-drivers-status 14:26:44 There are 2 drivers without maintainers and, as per the last communication (damnit I don't have the link handy), those will be deprecated if no maintainers sign up 14:27:09 that being said, there's one more thing that came up while talking with folks from the swift team 14:27:27 The maintainer of the swift driver (the main contact, at lesat) is someone from the Swift team, which is awesome 14:27:57 However, in order to make this maintenance work better, we should find a way to join efforts. Get feedback from the swift team and provide knowledge from the glanc eside 14:28:09 therefore, we started talking about joining efforts 14:28:11 Are we going to skip 2 meetings? 14:28:35 That is, have a group that includes glance-core+swift-core and give that group permissions on glance_store/_drivers 14:28:50 kairat_: yeah, we can talk about that again in the open discussion. We just switched topics 14:28:58 I wanted to get your thoughts on that 14:29:06 We can do thta for the cinder driver as well 14:29:27 I think we'd all benefit from that and we'd also increase the quality (hopefully) of the driver 14:29:55 we don't need to create a whole new group, we can just give swift-core permissions on glance_store/_drivers 14:30:05 erm, glance_store/_drivers/swift 14:30:20 sounds like a good idea to me 14:30:20 can we give permissions on a certain directory? 14:30:36 kragniz: I believe we can, I need to double check 14:30:40 I thought it was per-repo only 14:30:48 I didn't realise you can give perms on specific directories, interesting 14:30:53 * kragniz might have missed something in the new gerrit 14:30:53 TBH, if we can't, I'd be even ok with giving them access on glance_store based on trust 14:31:03 +1 14:31:30 I don't think any swift core is interested in reviewing all glance_store patches :P 14:31:47 anyway, if there are no objections, I'll proceed with the infra patches to make this happen 14:31:56 (assuming we can give permissions on sub-dirs) 14:32:01 flaper87, sorry, it was a huuge lag because of my WI-FI connection 14:32:04 otherwise, I'll just add the team 14:32:24 kairat: no worries, I thought so. I have some pretty bad lag myself as well 14:32:48 kairat: tl;dr: yes, we're skipping the next 2 meetings, we'll use openstack-dev to communicate as usual 14:32:56 ok, moving on 14:33:06 #topic Reviews are slowing down (flaper87) 14:33:10 this is just a heads up 14:33:27 I noticed reviewes are slowing down a bit, I guess this is because of the coming holidays 14:33:59 not big deal. However, I wanted to know if there's something I can do to help out there. Is the dashboard working for folks? Are you having issues to figure out what should be reviewed? 14:34:38 if there's anything, review-wise, that can be improved, do let me know 14:34:53 ok, that's all I've got on this topic 14:35:00 any comments? thoughts? concerns ? 14:35:33 #topic Reviews & Bugs 14:35:49 I see https://review.openstack.org/#/c/246971/ in the list 14:36:06 but it's getting enough reviews so, I'm confused on what needs to be raised there 14:36:48 fwiw I was surprised by the example response there 14:37:19 mclaren: haven't read it yet, I managed to get internet access 2mins before the meeting 14:37:26 or at least a somewhat stable one 14:37:29 sure, just a heads up 14:37:46 I may have been expecting the wrong thing 14:37:46 I'll read for sure 14:38:00 mclaren: btw, +1 for asking for an example, it helps clarifying the final intent 14:38:05 mclaren: i agree, this should not be a new response 14:38:10 and it seems to have worked in this case 14:38:15 +1 14:38:34 ok, anything else? 14:39:18 #topic Open Discussion 14:39:32 kairat: any issues you wanted to raise re skipping 2 meetings ? 14:39:38 besides the fact that we'll skip 2 meetings 14:39:41 kairat, nope 14:39:41 :P 14:39:51 i have another questions 14:39:53 * flaper87 likes when ppl talk to themselves 14:39:57 :P 14:39:59 shoot 14:40:12 I am trying to deploy glance under apache 14:40:32 AFAIU chunked requests are not part of WSGI spec 14:41:03 It means that glance has troubles when deploying under different WSGI frameworks 14:41:05 #info there won't be a christmas core cleanup... Talked to some of the ppl that would be elegible and they mentioned they'll get back to glance and they are still interested in contributing with reviews 14:41:28 Are we going to stay on eventlet for a long time because of this? 14:41:49 I don't think there are plans of moving away from eventlet anytime soon 14:41:59 that's one of the things we all hate but no one wants to fix 14:42:09 It'd have to be a x-proj effort 14:42:12 and a HUGE one 14:42:52 To some extent, I'd like to see glance becoming a pure wsgi app and not shipping any wsgi container 14:42:57 but that's definitely material for N 14:43:06 flaper87, agree 14:43:20 Just raise awareness 14:43:32 that we need to think about solution for that 14:43:37 as soon as we have a name for N and we're past the spec freeze, It'd be lovely to see a spec proposed for N with this 14:43:41 *hint *hint* 14:43:44 :D 14:43:51 It seems that only glance is using chunked requests 14:43:58 swift? 14:44:00 I know there are services that follow that pattern (Zaqar, hence my bias) 14:44:02 Swift may use it 14:44:05 but I don't know 14:44:22 Usually we install OS with Ceph back-end 14:44:39 right but there are many swift deployments out there 14:44:58 anyway, this needs to be thoroughly thought but I'd love to explore the possibility of becoming a pure wsgi app 14:45:20 not sure what others think, though 14:45:23 :D 14:45:29 ok, anything else? 14:45:33 nope 14:45:56 ok, in that case, I'll take the chance to wish everyone a Merry Christmas and a Happy new year! 14:46:11 See you all in 2016, have fun, misbehave because behaving is boring 14:46:21 you too! 14:46:22 Thanks! 14:46:31 thanks, right-back-atcha! 14:46:32 kragniz: misbehaving? OF COURSE! :D 14:46:38 kairat: you know me already 14:46:40 :P 14:46:46 does anyone know what web browser the new gerrit interface is optimized for? 14:46:47 erm, kragniz ^ 14:47:02 rosmaita: mmh, no idea, I've been using chrome 14:47:04 lately 14:47:20 rosmaita: works fine under chromium for me 14:47:21 i will have to try that, it's pretty bad in firefox 14:47:35 :( 14:47:55 the image import spec may test it to its limit 14:47:55 thanks everyone! Have fun! 14:48:03 mclaren: LOL 14:48:17 #endmeeting