14:00:49 #startmeeting glance 14:00:50 Meeting started Thu Oct 19 14:00:49 2017 UTC and is due to finish in 60 minutes. The chair is rosmaita. Information about MeetBot at http://wiki.debian.org/MeetBot. 14:00:51 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 14:00:53 The meeting name has been set to 'glance' 14:00:58 o/ 14:01:03 o/ 14:01:06 o/ 14:01:16 * smcginnis is fighting zuul v3 release fires today, might not fully be here 14:01:48 fires during milestone week ... oh no! 14:02:28 let's wait a minute for jokke_ , we need our release czar here today 14:03:51 In India we are celebrating Diwali festival this time, So wish you all a very Happy and Prosperous Diwali 14:04:03 thank you! 14:04:12 thanks abhishekk 14:04:22 a happy and prosperous Diwali to you, too 14:04:33 thank you :) 14:06:03 ok, guess we should get started 14:06:08 #topic updates 14:06:22 the only update i have is that the Q-1 milestone is this week 14:06:42 smcginnis do we need to cut a release today, or can we wait for next week? 14:07:07 rosmaita: Deadline is still today. If you can submit a release request patch that would be good. 14:07:19 ok, will do 14:07:25 It might be a little while before we can actually approve it, but would be good to have ready to go. 14:08:02 i need some eyes on the release note and doc changes for the registry deprecation, need to get that in before we cut the release 14:08:14 #link https://review.openstack.org/#/c/513390/ 14:08:32 rosmaita I'll check it out today 14:08:44 and definitely have some feedback before COB 14:08:51 thanks, should be pretty uncontroversial 14:08:59 abhishekk: need you to take a quick look, too 14:09:01 rosmaita: we should land the uwsgi bug fix for q-1 if possible, it's been sitting for a while... 14:09:19 rosmaita: yes looking 14:09:35 * jokke_ sneaks to the back row bit late 14:10:01 mtreinish: yeah, sorry about that ... meant to check back after it passed tests last night, but watched the yankees game instead 14:10:16 I have given one comment, but we can enhance it later as well 14:10:20 * croelandt joins jokke_ 14:11:37 (in case i disappear ... there's electrical work going on in my neighborhood) 14:11:45 rosmaita: once that reno is in I'll craft the release review 14:11:57 abhishekk: I had considered you're comment, but the mock isn't used in all the tests, just most of them 14:13:05 mtreinish: yeah 14:13:37 i think mtreinish 's change is good, works locally for me 14:13:38 mtreinish: so back to the import questin. The import will only succeed within the wsgi container, there is nothing providing that as a lib on external environments, right? 14:14:05 jokke_: correct, uwsgi dynamically inserts it's python module into the runtime when you run python code through it 14:14:06 Have no issues with this one 14:14:18 I linked the doc in a comment when smcginnis asked the same question 14:14:29 mtreinish: Thanks, that helped. 14:14:43 it's weird... 14:14:44 yes, that was helpful, i had the same question 14:14:50 mtreinish: cool, then I'm just fine with the change. Thanks for doing it and you patience 14:14:54 mtreinish: Only thought was maybe it needed a comment in the code explaining that since a few of us weren't aware, but that's not really a big deal. 14:15:02 Especially post-review. 14:15:59 smcginnis: I can do that in a follow up 14:16:14 mtreinish: Eh, probably not worth worrying about. 14:17:07 ok, i approved the change 14:17:28 jokke_ we'll need to watch it and the docs change before you get the milestone hash 14:17:44 rosmaita: yeah, will follow those over tonight 14:18:13 jokke_ you can handoff to me at the end of your day 14:18:22 don't need you to work late 14:19:10 :) 14:19:18 about deprecation patch, architecture diagram will be modified in S release, right? 14:19:58 abhishekk yes 14:20:13 i didn't modify any of the developer docs, just the admin and install guides 14:20:29 also, i didn't modify the glossary 14:20:45 it defines "image registry service" as another name for Glance 14:20:52 rosmaita: ok, sounds good, another question, do we need to include your file in architecture.rst as well? 14:21:14 * rosmaita looking 14:21:49 doc/source/contributor/architecture.rst 14:22:22 thanks, i wasn't finding it 14:23:01 everything else is fine, and as smcginnis said smart way to use include :D 14:23:24 i figured only one place to make a change if someone didnt' like the language i used 14:23:48 abhishekk do you mean to add the note in the "Registry Layer" section? 14:24:03 IMO, yes 14:24:24 i can do that 14:24:42 thanks 14:24:43 smcginnis did i spell 'deprecation' incorrectly? 14:26:01 rosmaita: Only thing I had in my mind is that you're specifying again "beginning" of the cycle ;) 14:26:32 we never get these things done at the beginning of the cycle :P 14:27:03 i thought last time we said specify beginning so that we don't have to wait until RC-1 14:27:04 I think I was nagging about that on the spec as well 14:27:17 i remember you saying something 14:27:28 but i guess i remembered inaccurately 14:28:11 i think this has us covered so that we can do it anytime in 'S' 14:28:13 ? 14:28:48 rosmaita: yeah, I think that's right 14:29:18 yeah ... let's just merge it rather than bikeshed the wording 14:30:17 as soon as Mr. Zuul is ok with it ;) 14:32:02 ok, just updated the patch with the change abhishekk requested 14:32:53 all right, i will move the bugs scheduled for Q-1 that didn't make it, to Q-2 14:33:27 and i will commit to getting reviews done by monday 14:33:57 i think that's all on the agenda for today 14:34:13 and it is 14:34:22 #topic open discussion 14:34:24 Okay cool, I'll have some feedback on the docs soon 14:35:03 McClymontS has to be real soon, we need to merge that change before we can cut Q-1 14:35:23 putting down devstack and heading over there lol 14:35:23 but, if you see something i missed, there's always a follow-on patch 14:35:35 ok, cool 14:35:44 I got internet!!! 14:35:57 i saw that in twitter, i think ... congrats! 14:36:07 :D 14:36:25 rosmaita looks good 14:36:41 speed tests giving an average of 280mbps so I think I'll be grand for a while :D 14:36:49 wow 14:36:52 McClymontS ty 14:36:53 * croelandt wants to move next to jokke_ 14:37:15 croelandt: you're in France, is that right? 14:37:20 rosmaita: yes 14:37:36 just checking 14:37:40 :) 14:37:58 280mbps might a reason to move to another country though 14:38:10 croelandt: they offer gig as well 14:38:23 jokke_: why didn't you get that? 14:38:23 I just thought that I'll be fine wit hthe 300meg line 14:38:26 meh 14:38:33 jokke_: so the "economy" connection is 280mbps! 14:38:39 yeah 14:38:52 wow, i agree with croelandt 14:38:55 Well I got fibre pulled into my kitchen yesterday 14:38:56 jokke_: Red Hat could probably pay for the better one 14:39:00 haha 14:39:06 only drawback: you have to work from your kitchen. 14:39:26 nope ... have been pulling ethernet around the house past 3 weeks 14:39:39 another question I have on include, might be silly 14:39:49 abhishekk: go for it 14:40:05 have gigabit network around the house ... if that uplink seems limiting I can just call them and get it pumped up 14:40:09 note is in folder doc/source/deprecate-registry.inc 14:40:56 did i miss a '..' on one of them? 14:41:23 and we are referring it for example in doc/source/configuration/glance_registry.rst as ../deprecate-registry.inc 14:41:44 so will it try to search it in configuration folder or source folder? 14:41:55 or we should have ../../deprecate-registry.inc their? 14:42:06 source as that's one step up from the configuration 14:42:31 for the same folder, no '..' 14:43:07 ok, got it, I was confused :D 14:43:08 i spot checked the generated files, and i'm pretty sure tox/sphinx/whatever would give a warning if an include file couldn't be found 14:43:16 so i think it's ok 14:43:32 abhishekk: it's good to check, because sphinx isn't always "normal 14:43:34 " 14:43:52 i mean, sometimes it behaves differently from what you'd expect 14:44:56 anything for open discussion? 14:45:08 rosmaita: sounds good 14:45:15 I think I'm good 14:45:38 hm 14:45:46 we discussed the oslo_policy issue last week 14:46:02 https://review.openstack.org/#/c/512020/ <- I sent this patch, if anyone feels like reviewing it :) 14:46:55 thanks croelandt noticed it coming and added to my post release review list 14:46:58 i see that doug's been watching it, that's good 14:47:06 corelandt: will have a look soon 14:47:20 croelandt : i will mark that bug for Q-2 so it doesn't slip through the cracks 14:47:26 nice! 14:47:43 thanks for that 14:47:50 just in commit you mentioned, oslo_poolicy' 14:48:15 sounds like a connection pooling kind of thing! 14:48:16 damn 14:48:20 what a good eye 14:48:21 aha 14:48:59 fixed 14:50:09 thanks 14:50:47 * jokke_ is so happy abhishekk joined our nitpicking team full time :D 14:51:07 :D 14:51:51 ./ 14:52:05 well, if nobody has anything, i can talk about something that came up in the glance channel yesterday 14:52:09 hi nikhil 14:52:15 hi :) 14:52:29 whats up 14:52:34 so, someone has a use case where they want an image to display as 'active' in horizon 14:52:37 rosmaita: shoot, you have ~7min :D 14:52:48 but, the image is just a placeholder to some external storage 14:52:57 that ironic knows how to access 14:53:10 lol 14:53:17 i think they're putting the external uuid as an image property or something 14:53:20 anyway 14:53:39 i told them if they don't care about the size or checksum, to just upload a 0-byte image 14:53:46 but, that seems kind of icky 14:53:56 so i was wondering what other suggestions people have 14:54:38 so basically this means, that image sharing, import etc. will be irrelevant in this case? 14:54:45 IIRC there was something about the 0 size images we discussed about. It had some problem there why it was brought up 14:55:14 I think it was something nova related around snapshots or something 14:55:34 (proposal is for a mannequin image?) 14:55:41 yes, basically 14:55:59 it's a public image furnished by the operator for exclusive use of ironic 14:56:10 hmm lol 14:56:24 they just need it to show up in horizon so that it can be selected when people choose a bare metal flavor 14:56:25 this puts me on the heirarchical image train, basically to generalize it even more for - typed images 14:56:29 (something like that) 14:57:01 shared, public, private, community, mannequin, heirarchical - images (image types) 14:57:34 Thats not a bad idea 14:57:35 i think whether its a mannequin or not would be orthogonal to visibility 14:57:44 rosmaita: I assume those images would be created by admins anyways so they could just make location schema for them 14:58:22 i asked that (i think), but they don't have a uri, just a uuid 14:58:42 i guess they could make up "mannequin://dead-beef-deadbeef" as a scheme 14:58:44 but tbh that sounds like hacky enough for us not to support so I don't think I have any opinion with value how they want to hack it together ;) 14:58:48 but would glance barf on that? 14:59:08 yeah, i didn't want to tell them to mess with image locations if it wasnt' necessary 14:59:55 i think they're ok with the fake image data solution 15:00:00 ok we are out of time 15:00:05 thanks everyone! 15:00:14 #endmeeting