14:00:14 <flaper87> #startmeeting Glance Drivers 14:00:14 <openstack> Meeting started Tue Jan 19 14:00:14 2016 UTC and is due to finish in 60 minutes. The chair is flaper87. Information about MeetBot at http://wiki.debian.org/MeetBot. 14:00:16 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 14:00:19 <openstack> The meeting name has been set to 'glance_drivers' 14:00:26 <rosmaita> o/ 14:00:31 <ativelkov> o/ 14:00:32 <flaper87> o/ 14:00:50 * flaper87 waits a couple of more mins 14:02:11 <jokke_> o/ 14:02:32 <flaper87> Other drivers seem to be out :( 14:02:49 <rosmaita> #link https://etherpad.openstack.org/p/glance-drivers-meeting-agenda 14:02:53 <flaper87> rosmaita: meantime, there are 3 specs pending approval for their exception request 14:03:02 <nikhil> o/ 14:03:04 <flaper87> #link https://review.openstack.org/248681 14:03:10 <flaper87> #link https://review.openstack.org/170564 14:03:19 <flaper87> #link https://review.openstack.org/254710 14:03:24 <flaper87> nikhil: HEY! 14:03:32 <nikhil> hey hey 14:03:35 <flaper87> 2s 14:05:41 <flaper87> back 14:05:42 <flaper87> sorry 14:05:47 <flaper87> so, any thoughts on those specs? 14:05:52 <nikhil> I am okay with the multi-tenant swift store spec 14:06:04 <nikhil> will give another run through before +2 14:06:13 <flaper87> ++ 14:06:26 <mclaren> nice :-) 14:06:34 <flaper87> nikhil: you had some comments on Friday about the Glare's spec. Any chance you can dump those on the spec? 14:07:16 <flaper87> mclaren: hey there :D 14:07:17 <nikhil> if I am blocking the work, I would say please go ahead with the already plan 14:07:34 <nikhil> I wanted to comment yesterday but couldn't so I will try to review it today 14:07:40 <flaper87> I'm ok with mclaren's and mfedosin's secs. Need to go through Glare's one again 14:08:02 <flaper87> but I don't think I have major objections 14:08:10 <flaper87> rosmaita: mclaren thoughts on Glare's spec ? 14:08:15 <nikhil> multi tenant spec doesn't have core reviewers assigned just fyi 14:08:38 <rosmaita> flaper87: i am in favor, but need to read through one last time 14:08:40 <mclaren> haven't looked - sorry :-/ 14:08:53 <nikhil> we can amend that too 14:09:03 <flaper87> mclaren: can you add myself and swift's driver maintainer as core reviewers for your spec ? 14:09:25 <mclaren> absolutely 14:09:27 <flaper87> rosmaita: mclaren it'd be super cool to have those reviews done today, if possible 14:09:46 <flaper87> we need to provide feedback to these folks so they can move forward with the work 14:09:51 * flaper87 captain obvious 14:10:21 <flaper87> coolio 14:10:25 <flaper87> Now, lite specs 14:10:42 <flaper87> #link https://bugs.launchpad.net/python-glanceclient/+bug/1489534 14:10:43 <openstack> Launchpad bug 1489534 in python-glanceclient "Move metadef commands to sub-commands" [Wishlist,In progress] - Assigned to Niall Bunting (niall-bunting) 14:11:22 <flaper87> #link https://bugs.launchpad.net/python-glanceclient/+bug/1531050 14:11:23 <openstack> Launchpad bug 1531050 in python-glanceclient "Support create images with locations in v2" [Wishlist,In progress] - Assigned to wangxiyuan (wangxiyuan) 14:11:29 <flaper87> #link https://bugs.launchpad.net/glance/+bug/1528453 14:11:30 <openstack> Launchpad bug 1528453 in Glance "Provide a ranking mechanism for glance-api to order locations" [Wishlist,New] 14:11:34 <flaper87> #link https://bugs.launchpad.net/glance/+bug/1535437 14:11:35 <openstack> Launchpad bug 1535437 in Glance "Deprecate `show_multiple_locations` in favor of policy" [Wishlist,Triaged] 14:11:48 <flaper87> Those are the ones I've identified for today's meeting 14:12:13 <flaper87> IIRC, we have talked about metadef's one already 14:12:44 <flaper87> in one of Glance's meetings 14:12:49 <flaper87> need to formalize the vote 14:13:09 <flaper87> knock ? 14:13:19 <mclaren> +1 from me 14:13:24 <nikhil> I am ok with our earlier agreement 14:13:25 <rosmaita> i am +1 on making metadefs help a sub-command 14:13:56 <flaper87> ok, I'll comment back on the bug 14:14:15 <nikhil> Is this "Provide a ranking mechanism for glance-api to order locations" limited to adding formal metadata to img loc ? 14:14:29 <flaper87> yes 14:14:31 <flaper87> AFAIU 14:14:43 <mclaren> (a related question is will artifacts commands be exposed by the glance command or just the openstack command) 14:15:29 <nikhil> earlier it was on the lines of just osc 14:15:38 <nikhil> but not aware of latest development on that 14:15:47 <flaper87> nikhil: we can ask for clarifications on the ranking spec 14:16:05 <flaper87> Any comments on the "Support create images with locations in v2" ? 14:16:14 <nikhil> flaper87: I think we should. It seems to put some aspects out of scope so would be good to get more info 14:16:52 <flaper87> nikhil: will comment back on the bug and ask for more info 14:16:59 <nikhil> thanks 14:17:09 <flaper87> also, any comments on the deprecation of `show_multiple_locations` ? 14:17:15 <nikhil> Even the support images needs more info 14:17:52 <nikhil> as the logic is quite different from v1 and can be disabled by ops, I think it would be good to know the user experience 14:19:08 <rosmaita> just for the record, i am +1 on more info for location ranking via metadata spec-lite 14:19:37 <flaper87> rosmaita: roger! We need to clarify what the strategy is but I think we agree on the general idea 14:19:44 <mclaren> I'm not sure I 100% understand 'Support create images with locations in v2' 14:20:04 <flaper87> mclaren: I think it's something like: glance image-create --location blah://blah/blah 14:20:16 <flaper87> rather than glance image-create --file /path/to/file 14:20:27 <flaper87> it's a glanceclient spec lite 14:20:43 <mclaren> ahhh, ok 14:21:08 <rosmaita> so the current workflow is that you image-create and then PATCH to get the location set? 14:21:13 <mclaren> backwards compatabiltiy, yes please :-) 14:21:15 <flaper87> yea 14:22:01 <rosmaita> only problem i see is that location setting tends to be more admin than end-user, but mabye that doesn't matter 14:22:01 <nikhil> but the experience of the user is different, with an explicit PATCH call and with a direct specification of location 14:22:20 <nikhil> with v1, data is copied while v2 will ignore that 14:22:51 <nikhil> so when user doesn't know which API version is being used underneath they are likely to misunderstand the loc where the data is 14:23:05 <flaper87> mmh, TBH, I'm not sure I like this proposal as-is 14:23:10 <flaper87> I'll have to put more thoughts on it 14:23:19 <flaper87> nikhil: right 14:23:39 <flaper87> well, in v1 it was --copy-from to copy the data 14:23:47 <flaper87> all those parameters... meh 14:23:57 <rosmaita> flaper87: let's ask for a revision, where the author describes current workflow and what the differernce will be, and how it will be more like v1 experience 14:24:07 <flaper87> rosmaita: ++ 14:24:17 <mclaren> yes, clarification sounds good 14:24:37 <flaper87> so, +1 on metadef's help, clarification on the other 2 14:25:04 <flaper87> other 2 = location ranking and image-create w/ location 14:25:46 <flaper87> Pls, comment on the show_multiple_locations deprecation 14:25:47 <rosmaita> what did we say aobut deprecating show_multiple_locations in favor of policy? 14:25:53 <flaper87> sigmavirus24_awa: dropped a comment there 14:26:00 <flaper87> rosmaita: nothing yet 14:26:07 <rosmaita> just checking 14:26:07 * flaper87 thinks we should do that 14:26:12 <flaper87> we already have policies for it 14:26:42 <jokke_> I think the reson why we did not implement it was for user experience as user has no way of knowing what drivers glance has enabled 14:26:51 <jokke_> and what the location format for those might be 14:26:52 <rosmaita> sounds ok to me, but did nikhil have a commment about v1? or was i misreading the discussion? 14:27:15 <flaper87> rosmaita: that was re the image-create one, wasn't it? 14:27:20 <flaper87> nikhil: ^ 14:27:22 <jokke_> ^^ ref the locations stuff 14:27:38 <flaper87> lol, this is confusing :P 14:27:44 <nikhil> ignore that 14:27:55 <flaper87> Ok, time is running out. Please, drop comments on the bugs 14:28:05 <flaper87> I'll ask for clarifications on the ones we agreed on 14:28:07 <nikhil> I just checked the code and --location is being used to set the direct location in v1 too 14:28:12 <flaper87> and triage the metadefs one 14:28:33 <nikhil> but I think the allowed values of location in v1 has smaller set 14:28:42 <flaper87> if there are no objections, I'll triage the `show_multiple_locations` one as well 14:28:45 <rosmaita> flaper87: re-read, you are right, nikhil was talking about support location with create 14:28:53 <flaper87> ok, coolio 14:29:04 <flaper87> 2 lite specs triaged, 2 need clarification 14:29:13 <flaper87> Thanks for attending, folks! 14:29:17 <flaper87> Have a great day :) 14:29:25 <rosmaita> you too, bye! 14:29:32 <nikhil> have a nice one 14:29:33 <nikhil> thanks 14:29:36 <flaper87> (oh, PLEASE PLEASE PLEASE, PRETTY PLEASE, remember the specs waiting for approval) 14:29:44 <flaper87> #endmeeting