14:00:22 #startmeeting Glance 14:00:22 Meeting started Thu Feb 11 14:00:22 2016 UTC and is due to finish in 60 minutes. The chair is flaper87. Information about MeetBot at http://wiki.debian.org/MeetBot. 14:00:23 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 14:00:25 The meeting name has been set to 'glance' 14:00:26 o/ 14:00:31 o/ 14:00:37 o/ 14:00:39 o/ 14:01:01 hey folks 14:01:11 hope you all had a great week 14:01:22 #topic Agenda 14:01:24 #link https://etherpad.openstack.org/p/glance-team-meeting-agenda 14:01:27 o/ 14:01:30 o/ 14:01:32 You can find our agenda there 14:01:47 #topic Updates Glare 14:01:51 mfedosin: ninag 14:01:54 ops 14:01:55 mfedosin: nikhil 14:01:57 :D 14:02:09 it happened 14:02:12 finally 14:02:28 Glare is officially here 14:02:45 we merged it last night 14:02:55 o/ 14:03:06 nnnnnnice 14:03:08 well done folks 14:03:12 https://review.openstack.org/#/c/255274/ 14:03:13 thanks to everyone for the reviews 14:03:24 Now, let's get the spec together and prepare for Newton 14:03:26 thanks everyone who helped there 14:03:47 yes, we began to work on the spec with kairat and nikhil 14:04:01 cool. Do you have an ETA ? 14:04:13 I present a small draft of api on the latest glare meeting 14:04:21 definitely before summit 14:04:34 I'd suggest for it to be ready by M-3 14:04:42 That gives enough time for discussions to happen before the summit 14:04:53 but actually it must be done before 23th of February 14:04:53 so that summit can be used to polish some other things 14:05:22 coolio 14:05:24 anything else? 14:05:28 yes 14:05:32 one question 14:05:42 I wrote a small doc about Glare 14:05:59 and I wonder if we can merge it in Glance docs 14:06:17 it's like FAQ about Glare 14:06:24 Is that the doc you shared at the virtual mid cycle? 14:06:33 I don't think people would be mad for having more docs 14:06:35 :P 14:06:41 (or some docs) 14:06:45 * flaper87 ducks 14:07:09 I didn't share anything there :) 14:07:15 just a small pic 14:07:20 mfedosin: mmh, you did show us a diagram 14:07:22 it was shared in the artifacts meeting 14:07:23 oh ok 14:07:44 mfedosin: the doc team wants each project to take more responsibility, so adding a glare faq sounds like a great idea 14:07:44 https://docs.google.com/document/d/1D4Hik0-XNqwi43gxtybCs9-fMqh5IAIu8CHB33e46OQ/edit 14:07:46 this one 14:07:47 anyway, I think it's fine to have more docs. Let's review all that on gerrit 14:08:02 #link https://docs.google.com/document/d/1D4Hik0-XNqwi43gxtybCs9-fMqh5IAIu8CHB33e46OQ/edit 14:08:12 Olena will help me with that 14:08:16 I guess the ques from mfedosin and I is 14:08:25 can we propose it against glance source tree? 14:08:48 or would people mind it? 14:08:50 or it should go to OpenStack wiki 14:09:00 I don't think there's an issue with that. Glare is part of Glance's source. As long as it's in rst format, I think it's ok 14:09:10 I know searchlight project has awesome in tree docs 14:09:12 flaper87: wow 14:09:24 cool 14:09:25 thank you for that 14:09:40 coolio 14:09:54 anything else? 14:10:03 nothing from my side 14:10:04 i think stuff in the wiki gets ignored, i vote for putting it in with the glance docs 14:10:29 plus, it's hard to update the wiki 14:10:38 and the session lasts less than my hopes for a better world 14:11:00 and my hopes for a better world don't last long.... like nothing 14:11:03 anyway 14:11:10 * flaper87 should stop saying dumb stuff 14:11:14 moving on 14:11:31 #topic Updates Nova v2 -> v1 (it's not a typo) 14:11:41 oh 14:11:48 it's painful 14:11:51 :( 14:11:54 yeah 14:11:58 but we have good news and bad news 14:12:07 good news are: 14:12:27 * nikhil chuckles on mfedosin's style of updates 14:12:27 I spent three sleepless night, but made it work 14:12:35 hooray! 14:12:41 \o/ 14:12:48 https://review.openstack.org/#/c/259097/ 14:12:53 everything 14:13:08 xen plugin, ceph... 14:13:23 Also I got a lot of comments from Bob Ball 14:13:31 they were really useful 14:13:46 mfedosin: man, you better be in Austin because I'm buying you drinks 14:13:47 I hope he likes the patch 14:14:20 I'll be prepare :) 14:14:35 but also there is bad news: 14:14:45 John -2'ed the patches 14:14:48 mfedosin: more updates ? 14:14:56 or are you just dropping the bad news on me 14:14:57 I saw that 14:14:57 because they are non-priority 14:14:59 ? 14:15:01 :P 14:15:07 oh, my IRC is lagging 14:15:09 :( 14:15:24 I'm going to send a message in ML 14:15:33 "Sorry, we have now hit the Non-Priority Feature Freeze for Mitaka. For more details please see: http://docs.openstack.org/releases/schedules/mitaka.html#m-nova-npff and http://docs.openstack.org/developer/nova/process.html#non-priority-feature-freeze" 14:15:47 because afais too many people want it 14:15:54 so, FWIW, this work is not a priority for Nova but it is for Glance 14:15:59 I'd certainly support trying to get it in 14:15:59 +1 14:16:08 How about we make them priority and makr v1 deprecated now ... like literally within this meeting? 14:16:10 It's been hard and I won't hide I'm supper annoyed by this not going in 14:16:50 Let's have another discussion on the ML and see what can be done 14:16:55 As I said even if it's a priority for Nova, it's a priority for whole OpenStack 14:17:04 mfedosin: I do agree with you 14:17:13 like super fully agree 14:17:19 *not a priority for Nova 14:17:26 ++ 14:17:35 we all work for OS 14:17:51 and it's better to collaborate to make it better 14:18:02 so, let's get that email out but I'd recommend not getting the hopes very high 14:18:19 I know, but let's try to do it 14:18:24 can we qualify the risk to some degree? Is it fairly low risk? 14:18:32 mclaren: what risk? 14:18:52 I guess nova don't want to risk taking new changes -- so can we convince them it's low risk? 14:19:04 at least we can set GLANCE_VER = 1 14:19:11 yeah, for example 14:19:18 and it will be the old workflow 14:19:20 "For things that are very close to merging, it’s possible it might get an exception for one week after the freeze date, given the patches get enough +2s from the core team to get the code merged. But we expect this list to be zero, if everything goes to plan" 14:19:31 It's not just a matter of risks, TBH. 14:19:37 I doubt if John would block this 14:19:46 I recommend talking to him first 14:19:48 The motivations I've heard are risk, design, timing 14:19:50 as no one else is blocking it 14:20:14 I think -2 is merely a follow on his Nova exception policy 14:20:15 nikhil: I've been talking to John. The block comes from an agreement with the rest of the team 14:20:28 ah k 14:20:31 The freeze happened 2 weeks ago, he actually gave extra time for this 14:21:00 oh and the extra week discussed with sean wasn't considered by the rest then? 14:21:19 anyways, just dumping my thoughts here...prolly we don't need to discuss this 14:21:58 i'd suggest mike and flavio coordinate on what goes out to the ML 14:22:00 Anyway, let's wait for mfedosin's email and see what the result of the discussion is 14:22:10 I am sure the rest of the OS wants v2 to become default in M 14:22:30 I don't want us to just waste time pointing fingers on the ML but rather evaluating how this can go in in Mitaka 14:22:37 real default (whatever that means) 14:22:40 The plan was *NEVER* for this to be perfect in mitaka 14:22:52 This was agreed upon back in Tokyo (or after tokyo) 14:23:13 flaper87: also you have a spec merged 14:23:19 do I? 14:23:22 oh right 14:23:24 I do 14:23:26 :D 14:23:28 anyway 14:23:28 =P 14:23:30 Let's move on 14:23:49 #topic Cross-project updates 14:23:50 nikhil: ? 14:23:53 anything ? 14:23:57 hey 14:24:02 yeah, that's me 14:24:10 there are 2 (semi) updates 14:24:26 1. Glance & subsequent discussion on (nested) quotas 14:24:46 SO, there are about 5 or so projects with quotas implemented already 14:25:01 only cinder has nested quota work done fairly well 14:25:36 Most of the feedback I got was go ahead with simple quota first and if you don't have HMT the move from simple -> nested won't be that tricky given we keep stuff configurable 14:25:50 But I am still evaluating the technical ramifications 14:26:23 The first step I am taking is work on a cross prj oslo.quotas spec with vilobhm 14:26:54 we are hopeful to get some feedback on this; given it's fresh I want to bring that spec up in a next to next week's mtg 14:27:00 I will update more on this topic then 14:27:18 nikhil: are you working on a corss prj spec for simple or nested quotas? 14:27:20 mtg == xprj mtg and following glance one 14:27:45 that sounds awesome. I guess simple quotas for now it is 14:27:53 ninag: it's a complete quota spec. I think not all projects have implemented even simple so it will start with that 14:27:57 ninag: nikhil did you coordinate with flwang ? 14:28:05 ninag: I will share details with you offline 14:28:12 flaper87: yep 14:28:17 nikhil: sounds good..i have some concerns and ideas 14:28:29 ninag: I can def imaging 14:28:34 imagine** 14:28:46 k 14:28:55 * nikhil == typos with cold fingers 14:28:57 the second update? :D 14:29:04 2. Query config 14:29:15 #link https://review.openstack.org/#/c/242852 14:29:20 There was some discussion on this 14:29:43 most to see if this sort of thing is useful and should the x-prjs go with yes or no 14:29:48 mostly* 14:30:19 I can see the use of it in some projects but from glance perspective it seemed rather unnecessary 14:30:35 it's still under discussion so open for questions here 14:30:42 right 14:30:53 I agree it's not super useful for us 14:30:59 I haven't read that spec 14:31:02 but I will catch up 14:31:15 oh he abandoned it yday, late refresh oops 14:31:34 so yeah, mostly noop from us 14:31:47 ok 14:32:04 Anything else? Questions from folks? Things that should be raised in the next meeting? 14:32:28 Next x-prj mtg you mean? 14:32:35 yes 14:32:44 those were open questions for everyone 14:33:03 I guess people are fine with your update :D 14:33:04 w00h000 14:33:07 thanks a bunch nikhil 14:33:12 moving on 14:33:13 yw 14:33:19 #topic Glance virtual mid-cycle summary (flaper87) 14:33:37 We had a virtual summit last week 14:33:42 it was supposed to be recorded 14:33:44 you can blame me 14:33:46 :) 14:33:55 Apparently I don't know how to click a freaking button 14:34:07 (except for those that say: "Here take 1M USD" 14:34:08 ) 14:34:14 Anywya 14:34:20 haha 14:34:21 anyway, even 14:34:25 I think it was good 14:34:46 We talked about a bunch of stuff an we spent a significant amount of time talking about image import refactor 14:34:58 Here's a small summary (I'll write that email I promissed) 14:35:03 (I promise ;) 14:35:05 :P 14:35:07 #link https://review.openstack.org/#/c/255274/ (Move Artifacts to glance-glare [landed]) 14:35:09 That landed! It's awesome 14:35:19 for sure 14:35:23 We won't be doing anything else on Glare in the mitaka time-frame 14:35:35 The client patch has been -2'd until we have a better API 14:35:40 and folks will work on the spec 14:35:47 The fast track thing is going away 14:36:06 and we'll all review glare's patches like a bunch of sis/bros that love each other 14:36:16 :) 14:36:22 :P 14:36:26 ^_^ 14:36:33 Love is in the air 14:36:37 We also talked about quotas 14:36:50 (para pa para pa pa) 14:37:03 There are 2 proposals but nikhil just updated us with the x-prj effort 14:37:15 it's a semi effort 14:37:18 if I could add something here? 14:37:18 for completeness and "logness" here are the links 14:37:18 :) 14:37:22 #link https://review.openstack.org/266667 (Image quota flwang) 14:37:26 #link https://review.openstack.org/275674 (Image quota Nina) 14:37:35 ninag: absolutely 14:38:06 do not ever feel you can't share in this (or any) place in openstack 14:38:10 while i think it is great to have Nikhil work on the cross-project spec...it was tried before at 2 summits..so I think we should look at both in paralell 14:38:27 working on out own specs while working the cross -project one 14:38:47 otherwise we will end up with no having project level quota support again for another release 14:38:59 ninag: that's a good observation. The way x-prj specs normally work is that ppl end up implementing what's written there in every project 14:39:20 I'm still happily opposing that work taking on before we get the v1 deprecated so we can just focus doing it in one place properly 14:39:40 yeah, so there are many dynamics here 14:39:47 1. API changes 14:39:49 I agree we will only do it for v2 APIs..that is a dependency 14:40:03 2. Conformity on nested 14:40:31 3. Simple to nested upgrade/config complexity (or keeping stuff optional) 14:41:17 4. Timing HMT support with nested (last one) 14:41:42 I think there's some work that can definitely be done in parallel but we gotta be super careful we don't diverge from the x-prj effort 14:41:49 I have a lot to say on this (like in the mid cycle) but avoiding due to time crunch 14:42:05 nikhil: let's put it in next week's agenda 14:42:10 k 14:42:12 and we can dedicate a good slot to it 14:42:12 I agree: and we have been talking with the other projects (mc_nair who is fixing the CInder issues is on the same team ) 14:42:27 Sounds good. thanks 14:42:39 ok, moving on. In addition to the above, we also spent Friday's session talking about the image import refactor 14:42:53 Some things to follow up on: 14:42:55 #link https://review.openstack.org/#/c/271021/ (Image import alternatives) 14:43:01 #link https://review.openstack.org/278086 (Image import requirements) 14:43:14 The first is a list of alternatives mclaren proposed 14:43:25 the second is a clarification of what the requirements are 14:43:35 We agreed to push the coding side of this work to Newton 14:43:43 There's simply no time for that in Mitaka 14:43:44 BUT 14:43:54 we have to reach an agreement by M-3 14:44:23 That's the goal for Mitaka, at least 14:44:36 just want to say something about https://review.openstack.org/#/c/278086/ ... 14:44:47 i know the original spec is a monster, so a lot of people's eyes may have glazed over reading it 14:44:49 If there will be a discussion in austin about this, it has to be very specific and related to future works on this area 14:45:02 ... but it would be great if everyone here could take a look at this much much much shorter patch and leave comments if you have an opinion 14:45:11 I don't want us to spend another 8 months planning this work :) 14:45:17 rosmaita: ++ 14:45:31 yes, now is the time to object or point out flaws in what we are proposing 14:45:34 (actually, december was the time, but here's a second chance!) 14:45:36 (but it would be nice if this is the last chance!) 14:45:41 I said I'd research tempurls a bit (for sigmavirus24) 14:45:49 https://etherpad.openstack.org/p/swift-temp-url-for-glance-uploads 14:45:53 ^ some notes 14:45:57 mclaren: thanks for taking the time 14:46:04 I'll go through those 14:46:33 ok, that's it on this topic 14:46:38 (jeeez I'm slow today) 14:46:38 first problem is tempurl doesn't currently support > 5GB (ie the sub-large-object limit) 14:46:57 mclaren: I think it'd be better to have a topic for this in next week's meeting 14:47:02 sure 14:47:05 thoughts ? 14:47:09 awesome 14:47:27 #topic M-3 approaching (flaper87) 14:47:35 M-3 is around the corner 14:47:53 #link http://releases.openstack.org/mitaka/schedule.html#m-3 14:47:55 like, no joke 14:47:58 cycle is almost over 14:48:02 and it's been super fun 14:48:05 I missed the Fri session, but iiuc the spec that was agreed on is pretty close to the solution we agreed in Tokyo so before reading that PS is that proposing a) us starting the planning from scratch b) refining some details from the merged spec or c) something different? 14:48:16 NOW, there are still things that need to be completed 14:48:34 there are specs that have been approved that need to be implemented 14:48:56 I've set a soft deadline for next week's friday. If patches are not up and green by then, I'll move those specs out of Mitaka 14:49:14 The next two weeks should be spent on reviews only (or mostly) 14:49:44 Any questions? 14:49:58 I've went through all the patches available and -1'd requesting updates 14:50:04 I'll do the same next monday 14:50:11 jokke_: (a) not proposing start from scratch, (b) my update patch is to remind us what we agreed on, (c) so that it's clear what to think about in evaluating mclaren 's alternatives patch 14:50:50 I guess there are no questions on this topic 14:50:59 #topic Vulnerability needs help 14:51:06 Perhaps this deserves review day 14:51:24 kairat_: or review week :P 14:51:25 as we have done previously 14:51:28 yep 14:51:32 ++ 14:51:35 So, vulnerability 14:51:46 just wanted to say about 'review night' 14:51:56 this cycle has been quieter, I guess. BUT, there are some security bugs that need some help and reviews 14:52:14 It's not nice to have sec issues opened and it's not nice to not having folks dedicating enough time to it 14:52:31 So, I'd like to improve that situation asap because it's never too late to tackle security issues 14:52:45 For security reasosn the sec team is not super big 14:52:54 But, I believe we could use 1 more person there 14:53:04 rosmaita: thnx 14:53:10 and perhaps someone willing to take a more active leadership on that front 14:53:20 hemanthm: has done some great reviews 14:53:26 but I believe he's swamped right now 14:53:34 or on vacation 14:53:43 no, swamped 14:53:49 I talked to him, he's swamped 14:53:53 (i was talking to him yesterday) 14:54:20 so, if you would like to help with our security team, please contact me 14:54:31 I'll be doing some changes in glance's coresec team 14:54:36 and we need support 14:54:54 who's in coresec now? 14:55:00 ok, I promise to be more active there. PRolly reduce attn in others. 14:55:08 mclaren rosmaita (you?) nikhil hemanthm and I 14:55:13 those are the names I remember 14:55:16 um 14:55:19 mclaren is our sec liaison 14:55:27 i don't think i am 14:55:29 I might be wrong about that list 14:55:30 I think it's 4 and rosmaita isn't there 14:55:35 oh ok 14:55:44 sorry about that 14:55:46 :D 14:55:47 +1 to mfedosin if he wants 14:55:52 we would need one for glare 14:55:57 ++ 14:55:59 i agree 14:56:00 I'll think about it 14:56:06 :) 14:56:10 sorry 14:56:11 mfedosin: not just think, look at your calendar as well 14:56:24 I'm sure you have like a gazillion of other projects you're working on 14:56:27 :P 14:56:31 perhaps kairat_ ? 14:56:34 :D 14:56:36 * flaper87 is enjoying this 14:56:38 hahahahha 14:56:39 ++ 14:56:46 flaper87: it was a polite repulse 14:57:11 I am changing my email filters, though I just see one bug this cycle 14:57:13 ok, ppl, just think about it. Sec is important, your cloud could go down if there's a sec issue #salespitch 14:57:23 i would support kairat_ , he pays close attention to details 14:57:41 anyway, that's it from me 14:57:41 So I may help with that 14:57:45 kairat is responsible for security bugs here 14:57:51 SOLD! 14:57:54 kairat_ it is 14:57:56 :D 14:57:58 w000h000 14:58:01 mfedosin means in Mirantis 14:58:01 :) 14:58:02 Ok 14:58:11 kairat_: now in upstream as well 14:58:24 kairat_: Congrats! 14:58:38 ok, I'll take the rest offline 14:58:43 Ok 14:58:43 #topic Open Discussion 14:58:47 2 mins 14:59:25 seriously people, please take a look at https://review.openstack.org/#/c/278086/, it's a quick read 14:59:31 rosmaita: ++ 14:59:32 Darja made updates for her filter patch 14:59:41 ppl do what uncle Brian says 14:59:44 >.> 14:59:50 mfedosin: saw that 15:00:05 If someone did not notice yet, my affiliation with HP ended at 3rd of Feb ... this means that my mailing list subscription (to my work e-mail) dropped out as well and I haven't set sufficient spam filter on my private one yet to subscribe there 15:00:17 Splease review it https://review.openstack.org/#/c/248359/ 15:00:24 thanks in advance 15:00:25 so if you need something from me, please drop me direct e-mail or ping me here 15:00:27 jokke_: oh, I didn't notice 15:00:37 here meaning irc ... I have client always online 15:00:48 jokke_: thanks for the heads up 15:01:10 and time.. 15:01:24 out of time 15:01:25 thanks all :) 15:01:27 #endmeeting