14:00:38 <jokke_> #startmeeting glance 14:00:39 <openstack> Meeting started Thu Jun 7 14:00:38 2018 UTC and is due to finish in 60 minutes. The chair is jokke_. Information about MeetBot at http://wiki.debian.org/MeetBot. 14:00:40 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 14:00:43 <openstack> The meeting name has been set to 'glance' 14:00:45 <jokke_> #topic roll call 14:00:51 <smcginnis> o/ 14:00:53 <jokke_> o/ 14:00:54 <abhishekk> o/ 14:01:01 <rosmaita> o/ 14:01:30 <jokke_> ok, looks like that's everyone 14:02:13 <jokke_> I don't have any actual updates apart from the specs which for I put it's own topic 14:02:27 <jokke_> #topic release updates 14:02:43 <rosmaita> hi 14:03:12 <rosmaita> erno's change removing the import enable opt is in the gate now 14:03:39 <rosmaita> i have the config regen patch up, built on master, doesn't depend on erno's change but includes it 14:03:48 <rosmaita> #link https://review.openstack.org/#/c/573238/ 14:04:04 <rosmaita> last i looked erno's patch has an hour more in the gate 14:04:33 <rosmaita> so we need reviews on the config regen: jokke_ smcginnis abhishekk ^^ 14:04:41 <abhishekk> acked 14:04:49 <smcginnis> Will do. 14:04:51 <rosmaita> and i will put up rocky-2 milestone for glance once that merges 14:04:56 <rosmaita> thanks! 14:05:21 <abhishekk> Erno is removing one config option right? 14:05:24 <rosmaita> ok, for glance_store, one patch still needs to be approved: https://review.openstack.org/#/c/534745/ 14:05:26 <jokke_> k 14:05:34 <rosmaita> abhishekk yes 14:05:45 <jokke_> abhishekk: correct it removes config option and obviously the related tags from couple 14:05:59 <rosmaita> the config regen also includes a fix i made to oslo.config for the display of list options 14:06:48 <jokke_> rosmaita: so about that store change ... sorry I did not notice it earlier 14:06:48 <abhishekk> got it 14:07:00 <jokke_> it's logging error condition as info 14:07:09 <rosmaita> arrrgh 14:07:19 <jokke_> I left comment in there ... otherwise it looks good 14:08:03 <rosmaita> is warn ok for that? 14:08:16 <jokke_> warn is fine by me 14:08:26 <rosmaita> ok, that bypasses the translation issue 14:08:37 <jokke_> and that means that we don't need to fiddle around with the translation either 14:08:43 <rosmaita> great 14:08:44 <jokke_> indeed 14:08:52 <rosmaita> ok, i wlll get that fixed while we speak 14:08:54 <jokke_> if you spin it around I'll keep eye on it 14:09:35 <rosmaita> great 14:09:52 <rosmaita> ok, i will put up a release patch as soon as taht change merges 14:10:07 <rosmaita> i take it no release notes are necessary? no major changes to glance_store 14:11:27 <jokke_> I'll double check I don't think there was anything 14:11:33 <rosmaita> ok 14:11:39 <rosmaita> that's it for rocky-2, then 14:11:48 <rosmaita> periodic tips jobs update 14:12:04 <rosmaita> #link https://review.openstack.org/#/c/572572/ 14:12:19 <rosmaita> that's the update to the glance docs about the release CPL watching the tips jobs 14:12:32 <rosmaita> and in that spirit, glance_store is all green 14:12:50 <rosmaita> but both glance and glanceclient are showing the keystone job failing for py35 14:13:01 <jokke_> ok, any idea why? 14:13:02 <rosmaita> i haven't had time to look into it 14:13:05 <jokke_> ok 14:13:27 <rosmaita> also, the keystone job isn't just keystone, it's keystone-related libraries 14:13:31 <jokke_> iirc glance was green yesterday 14:13:43 <rosmaita> yes, both were 14:14:19 <jokke_> ok cool, anything else? 14:14:43 <rosmaita> sorry, was just looking again ... for glance, the failure is in the functional-py35 tests, not py35 unit tests 14:14:51 <rosmaita> glanceclient is py35 unit tests 14:14:55 <jokke_> k 14:15:09 <rosmaita> i personally will not have time to look into it 14:15:36 <abhishekk> i will try to have a look tomorrow my time 14:15:46 <rosmaita> oh yeah, one more thing ... we have a bug squash scheduled for tomorrow 14:15:50 <rosmaita> are we still doing that? 14:16:06 <jokke_> I think we should 14:16:22 <rosmaita> i can only be there for my morning 14:16:41 <rosmaita> i was thinking a good use of time would be to go through reviews instead of bugs 14:16:56 <rosmaita> there are a bunch of old reviews hanging around that we should figure out what to do with 14:17:09 <abhishekk> +1 14:17:10 <rosmaita> guess it depends on who shows up 14:17:25 <rosmaita> i will put together a quick list of easy bugs in case any non-cores show up 14:17:40 <jokke_> so say, myself and abhishekk take a look on the bugs at the morn and we'll see if there is bug scrubbing to do when you wake up :P 14:17:55 <rosmaita> ok, that is even better 14:17:56 <smcginnis> I can try to be around tomorrow for it too. 14:18:03 <abhishekk> sounds good 14:18:09 <rosmaita> excellent 14:18:19 <rosmaita> ok, that;'s all for releases 14:18:23 <jokke_> ok 14:18:30 <jokke_> #topic eventlet situation 14:18:39 <rosmaita> ok, that's me 14:18:53 <rosmaita> #link https://review.openstack.org/#/c/571797/ 14:19:06 <rosmaita> i put up a patch that uses eventlet 0.22.1 14:19:15 <rosmaita> that's the next version the u-c will be raised to 14:19:25 <rosmaita> and something is failing, i have not had time to look 14:19:41 <jokke_> k 14:19:44 <rosmaita> maybe i should file a bug and someone can look tomorrow? 14:19:50 <jokke_> I'll have a look into that 14:20:05 <rosmaita> ok, cool, you can file a bug if you want to track it 14:20:06 <smcginnis> A bug would probably be good for tracking purposes at least. 14:20:19 <rosmaita> ++ 14:20:33 <jokke_> lets see what it's about first and I'll see what to do about it 14:20:40 <rosmaita> works for me 14:20:50 <rosmaita> ok, that's it on that topic 14:20:58 <jokke_> ok 14:21:03 <jokke_> #topic spec freeze 14:21:12 <jokke_> so today is the deadline 14:21:45 <jokke_> so the first one I want to put on the table is the hidden images spec 14:22:25 <jokke_> rosmaita: could you please take a look of my comment about your concern and we make the call if we keep that in Rocky or not 14:22:32 <rosmaita> looking now 14:23:29 <rosmaita> ok, i am cool with that 14:24:10 <jokke_> gr8, if you put note of that into the review and I'll get it merged 14:24:36 <jokke_> then we have the policy refactoring one 14:24:41 <jokke_> what do we do about that? 14:25:00 <rosmaita> that one needs work 14:25:13 <rosmaita> wxy had some good comments 14:25:21 <jokke_> we have the feature branch available, so should we just push that spec to S and we can start working on the feature branch when ever having time for it? 14:25:26 <rosmaita> i dont' think we have anyone ready to work on it? 14:25:34 <rosmaita> now, i mean 14:25:41 <jokke_> it doesn't seem like so 14:26:05 <abhishekk> rosmaita, I want to but might be later 14:26:19 <rosmaita> ok, let's push to S ... would be good to get some comments on my discussion with wxy 14:26:21 <rosmaita> see if that's the right direction to go 14:26:28 <rosmaita> if so, i will work out more details 14:26:39 <jokke_> rosmaita: agreed, lets keep that under discussion 14:26:51 <abhishekk> sounds good, If i get time I will push my patches against feature branch 14:26:58 <jokke_> then there is abhishekk's multistore 14:27:22 <abhishekk> Fingers crossed 14:27:29 <smcginnis> :) 14:28:05 <jokke_> now that needs at least review on the last revision, can we agree that as principle we are in favor to do that or do we drop it from this cycle? 14:28:24 <smcginnis> abhishekk: You have the code pretty much ready to go, right? 14:28:29 <abhishekk> yes 14:28:40 <rosmaita> i think maybe spec freeze exception? 14:28:45 <jokke_> I'd like to see that work happening even we are still not convinced that the implementation details in the spec are where we want to be 14:28:49 <abhishekk> I have uploaded glance and glance_store patches and abandoned it for reference purpose 14:29:01 <abhishekk> I have tested it with file, ceph and swift 14:29:11 <rosmaita> nice! 14:30:16 <jokke_> ok, keep an eye on that, I'll fiddle around with it today after the meeting, might be pinging rosmaita and smcginnis to have a look on my approach :P 14:30:34 <abhishekk> thank you 14:30:45 <jokke_> then we have the OSSN 14:31:16 <jokke_> abhishekk: would you please look through it before you call it a night and give your vote there 14:31:31 <abhishekk> jokke_, I will 14:31:37 <jokke_> I think the current proposal is something we can agree on and will get us to bit better place 14:31:46 <rosmaita> abhishekk: it changed a lot, i suggest just reading through the final patch 14:32:04 <rosmaita> ignore all the other stuff 14:32:04 <jokke_> exactly 14:32:09 <abhishekk> rosmaita, ok 14:32:13 <jokke_> that's why I asked ;) 14:32:48 <jokke_> then I added spec lite for the cluster awareness we talked about before/during the summit 14:33:07 <jokke_> https://review.openstack.org/#/c/572654 14:33:14 <smcginnis> And I think it has been very well communicated between the ML and the project update at the summit. 14:34:42 <smcginnis> jokke_: This might be too "lite" for me to grasp what's actually being proposed. 14:35:17 <jokke_> and I think that would be all we are/can include this time around. The web-location looks like something we won't be getting into agreement anytime soon 14:36:10 <rosmaita> actually, there is one more, has all +2s except from me 14:36:20 <rosmaita> https://review.openstack.org/#/c/561111/ 14:37:06 <jokke_> rosmaita: it has acks from all cores so I did not need to chase it ;) 14:37:25 <rosmaita> cool, just wanted to make sure it didn't get lost 14:37:34 <jokke_> rosmaita: nope ;) 14:37:35 <rosmaita> jokke_ i agree with smcginnis that your lite spec is pretty lite 14:37:47 <rosmaita> how much do you envision getting done? 14:38:17 <jokke_> yes so, the point is, there is no anticipated changes to API as of result to this atm 14:38:48 <jokke_> there is plenty of things that we might be able to benefit from if we decide to do it 14:39:20 <jokke_> for example starting from image import not necessarily needing that shared filesystem with all it's quirkyness 14:39:35 <rosmaita> i agree 14:39:57 <jokke_> so that would be the first thing I'd like to experiment with how much work it is to start communicating between the nodes 14:40:50 <jokke_> as in serializing the import call in glance-direct use case and see if we could get the node that actually has the data to process it without the request initially landing to that node 14:40:52 <rosmaita> so this is sort of a speculative spec saying that it's worth taking rocky time to work on this 14:41:40 <rosmaita> even though there's no guarantee something will land in rocky 14:41:48 <jokke_> yes, if it is not horrible to implement, then we have baseline how to work on it and we can start expanding it to stuff like precaching images 14:42:07 <rosmaita> i personally think it is a good idea 14:42:19 <abhishekk> agree 14:42:42 <jokke_> the spec lite is kind of there to get basic agreement that it would be worth for me to start looking into it 14:43:19 <smcginnis> OK, I just had no idea what was actually being proposed from the spec other than that the api service and a message bus were involved. :) 14:43:22 <jokke_> I don't want to dig into the message queues and o the work just to hear that someone from our core team thinks it's absolutely horrific idea and will -2 it ;) 14:44:21 <jokke_> smcginnis: yeah so like the spec is pretty vague, so is the plan for now apart from looking into getting the nodes communicating with eachother via message queues 14:44:47 <smcginnis> jokke_: OK, good enough for now I suppose. 14:45:36 <jokke_> cool ... I got lots of encouraging feedback of the idea from the summit 14:46:20 <jokke_> so if any of you think that specific implementations needs their own specs, I'm fine with that or amend the one that is proposed now with those details 14:47:25 <jokke_> ok, that's all about specs now 14:47:30 <jokke_> #topic open discussion 14:47:35 <abhishekk> I guess basic idea is clear with this lite-specs 14:47:50 <jokke_> there was couple of things listed under open discussion for today 14:48:05 <rosmaita> those are both me 14:48:22 <rosmaita> just some stuff on the ML relevant to glance 14:48:49 <rosmaita> first one, looks like nova may keep the image-download extension 14:49:12 <rosmaita> but glance is called "glacially slow" somewhere in the thread 14:49:26 <rosmaita> the second one erno should probably look at 14:49:37 <rosmaita> some confusions about image locations, i think 14:50:16 <jokke_> yeah, I've been keeping eye on that image download discussion 14:50:19 <rosmaita> that's all, links on the agenda if you are ineterested 14:50:27 <jokke_> I think I have missed the locations discussion 14:50:33 <jokke_> so will have a look 14:50:45 <rosmaita> i could be wrong, i read through really quickly 14:51:16 <rosmaita> one problem with that edge discussion is that it's split between the edge-computing list and the openstack-dev list 14:51:38 <jokke_> yeah 14:52:22 <jokke_> I'm subscribed to the edge list now as well so I will keep an eye what's going on there and will start bringing stuff up when I see something going on 14:52:31 <rosmaita> update: erno's removal of the enable_import opt has merged 14:52:35 <jokke_> this week has been just other things on my plate 14:52:43 <jokke_> \\o \o/ o// o/7 14:52:47 <rosmaita> the config regen is showing 1 hour more in the check gate 14:52:59 <jokke_> btw I also got the image conversion plugin up for review finally 14:53:07 <jokke_> would greatly appreciate feedback :D 14:53:12 <rosmaita> and the glance_store change is showing 1 more hour, it passed all our tests, now it's the tempest stuff 14:53:22 <jokke_> nice 14:53:32 <jokke_> should be fine then 14:54:04 <rosmaita> yeah, will just need +A 14:54:35 <abhishekk> I will test conversion patch tomorrow 14:54:41 <jokke_> https://review.openstack.org/#/c/572515 14:54:47 <jokke_> abhishekk: ty 14:54:59 <rosmaita> yeah, i have not had time to look at it 14:54:59 <abhishekk> is it good to have release note for this? 14:55:27 <jokke_> yeah, and documentation as well if the approach is ok 14:55:40 <abhishekk> Yeah 14:56:47 <abhishekk> Should I send spec freeze exception mail for multi-store? 14:57:04 <jokke_> smcginnis: thanks, missed that you had commented on the rev1 14:57:36 <smcginnis> ;) 14:57:39 <jokke_> abhishekk: don't yet 14:57:49 <abhishekk> jokke_, ok 14:57:55 <jokke_> ok, i think that's all for today 14:57:59 <jokke_> thanks all! 14:58:21 <rosmaita> i forgot to mention that i'm on vacation next week 14:58:22 <abhishekk> thank you all 14:58:30 <jokke_> rosmaita: ok, enjoy! 14:58:36 <abhishekk> rosmaita, happy holidays 14:58:41 <jokke_> #endmeeting