| *** sigmavirus24_awa is now known as sigmavirus24 | 14:23 | |
| *** lakshmiS has joined #openstack-searchlight | 14:57 | |
| sjmc7 | we meeting in here? | 15:00 |
|---|---|---|
| lakshmiS | yes that was the plan | 15:00 |
| TravT | Courtesy prioritization meeting reminder in this room: lakshmiS, nikhil_k, rosmaita, sigmavirus24, TravT, krykowski, david-lyle, wokuma, kragniz, sjmc7 | 15:00 |
| rosmaita | hello | 15:01 |
| TravT | Hello Let's wait until a couple mins for people. | 15:02 |
| david-lyle | o/ | 15:02 |
| TravT | here's a few things to ruminate on while we wait | 15:03 |
| TravT | https://wiki.openstack.org/wiki/Liberty_Release_Schedule | 15:03 |
| TravT | https://wiki.openstack.org/wiki/FeatureFreeze | 15:04 |
| TravT | david-lyle: for FF in the past, how have you handled it on horizon? | 15:05 |
| david-lyle | For Horizon I leave it at the 3rd milestone | 15:06 |
| david-lyle | then FFE (Feature Freeze Exceptions) can be sent to the dev mailing list | 15:07 |
| nikhil_k | o/ | 15:07 |
| david-lyle | Horizon is odd though as we get hit a lot with feature support for late landing features in other projects | 15:07 |
| david-lyle | as TravT is well away ;) | 15:07 |
| nikhil_k | guess so | 15:08 |
| TravT | you mean aware, i think. :) | 15:08 |
| david-lyle | I indeed | 15:08 |
| david-lyle | gah | 15:08 |
| nikhil_k | was having trouble with irc, I though | 15:08 |
| nikhil_k | thought* | 15:08 |
| david-lyle | no typy today | 15:08 |
| nikhil_k | no-grammar day? | 15:08 |
| TravT | So, we're looking at Sept 3rd being our last day without FFE. | 15:08 |
| nikhil_k | TravT: my thought was, if we really needed a FFE? | 15:09 |
| nikhil_k | since this would be our first package | 15:09 |
| sjmc7 | yeah, i had the same thought. perhaps if nothing else to comply with the Process | 15:09 |
| nikhil_k | it can be as solid as we want and when we want except not too far away from the rc3 date | 15:10 |
| sjmc7 | let's see what we've still got left to do though | 15:10 |
| TravT | I *think* we want to follow the common release milestones. | 15:10 |
| sjmc7 | yeah.. we should really spend a lot of time debugging as well towards the release | 15:10 |
| sjmc7 | we've found some really pretty elementary bugs the last couple of weeks | 15:10 |
| TravT | but we will need to have an evaluation meeting right before to make sure it helps us more than hurts us. | 15:10 |
| nikhil_k | that sounds good | 15:11 |
| sigmavirus24 | sorry, I'm lurking | 15:11 |
| sigmavirus24 | on a video call | 15:11 |
| sigmavirus24 | war calling | 15:11 |
| sjmc7 | i thought i felt someone behind me | 15:11 |
| TravT | that sounds fun sigmarvirus24 | 15:11 |
| TravT | well, let's walk the blueprints first I think | 15:12 |
| TravT | https://blueprints.launchpad.net/searchlight | 15:12 |
| TravT | i did some minor cleanup. | 15:12 |
| TravT | Top to bottom | 15:12 |
| TravT | glance-metadef-plugin | 15:12 |
| TravT | https://blueprints.launchpad.net/searchlight/+spec/glance-metadef-plugin | 15:12 |
| sjmc7 | isn't that in? the update mechanism needs some testing | 15:12 |
| sjmc7 | the port frmo DB access was a bit gnarly | 15:13 |
| sjmc7 | but it's indexing them properly | 15:13 |
| TravT | IMO this is basically done, but we haven't written queries against it | 15:13 |
| TravT | So, not entirely validated. | 15:13 |
| TravT | I marked it as Beta available, but we could mark as implemented and handle anything else as bugs. | 15:14 |
| TravT | thoughts? | 15:14 |
| sjmc7 | i'm ok with doing that but i'd be more comfortable if someone with some domain knowledge does some testing | 15:14 |
| TravT | that would be lakshmi and I who should do that | 15:15 |
| sjmc7 | ok. so maybe close this when you've at least given it a once-over | 15:15 |
| TravT | ok. | 15:15 |
| lakshmiS | sure | 15:15 |
| sjmc7 | i'll add a work item to remind us why it's open | 15:15 |
| TravT | priority wise on it, i'd actually say it is medium. | 15:16 |
| TravT | in other words, if i look at the nova blueprint, i'd say it is higher priority | 15:16 |
| nikhil_k | now the tests are fixed, nova one is ready to merge with another +2A | 15:17 |
| sjmc7 | i've done a decent amount of testing on the nova one | 15:17 |
| sjmc7 | i think TravT has too | 15:17 |
| sigmavirus24 | sjmc7: "a once-over" you mean... a glance? | 15:19 |
| sjmc7 | :| | 15:19 |
| sigmavirus24 | I'm sorry and you're welcome | 15:19 |
| *** TravT_ has joined #openstack-searchlight | 15:19 | |
| TravT_ | ugh | 15:20 |
| TravT_ | still there? | 15:20 |
| david-lyle | yup | 15:20 |
| rosmaita | you missed ian's pun | 15:21 |
| david-lyle | or am I | 15:21 |
| TravT_ | so, i will test the nova plugin again today and look over code again. | 15:21 |
| *** TravT has quit IRC | 15:21 | |
| *** TravT_ is now known as TravT | 15:21 | |
| TravT | I don't think there is any argument on nova being High and targeted to liberty | 15:21 |
| * TravT is now paranoid that is disconnected again | 15:22 | |
| nikhil_k | you should be | 15:23 |
| sjmc7 | :) | 15:23 |
| TravT | ok, https://blueprints.launchpad.net/searchlight/+spec/indexing-cli | 15:23 |
| TravT | This one had one sentence in it from our brainstorming. | 15:23 |
| TravT | i don't know if it originally intended to mean a remote CLI (openstack client) or if the searchlight-manage command covers it | 15:24 |
| david-lyle | TravT: I think there are permissions issues with openstackclient | 15:24 |
| david-lyle | you don't want just anyone triggering | 15:24 |
| lakshmiS | policy should allow indexing only for admins if turned on | 15:25 |
| david-lyle | I suppose it could be exposed and error out for non privileged users | 15:25 |
| nikhil_k | may be the intent was that admin would use os-client | 15:25 |
| nikhil_k | yes, the last comment | 15:25 |
| david-lyle | a plugin for osclient then | 15:25 |
| nikhil_k | I think this should be low for liberty though | 15:26 |
| TravT | I propose we open a new blueprint for osclient blueprint and close this one as meaning searchlight-manage | 15:26 |
| nikhil_k | because we don't know yet, how people would want to deploy this | 15:26 |
| david-lyle | depends how badly we are getting out of sync :) | 15:26 |
| nikhil_k | ha | 15:26 |
| nikhil_k | I just meant different clouds, but yeah if a couple big ones are okay then sure! | 15:27 |
| nikhil_k | my note on the whiteboard reminds me that this was a value addition suggestion and not much of desird use case to begin with | 15:27 |
| sjmc7 | i think at this point i'd stick with searclight-manage | 15:27 |
| lakshmiS | there's direct rest api to call if they need to get in sync | 15:27 |
| david-lyle | low is fine with me | 15:28 |
| lakshmiS | or maybe we didnt expose that yet :( | 15:28 |
| nikhil_k | iuc, direct as in REST api w/o sdk | 15:28 |
| TravT | lakshmiS, I didn't think we exposed that yet | 15:28 |
| nikhil_k | iiuc* | 15:28 |
| TravT | the index is per document, i believe | 15:28 |
| lakshmiS | yup | 15:28 |
| lakshmiS | we could do that very easily for admin users | 15:29 |
| nikhil_k | right | 15:29 |
| TravT | so, that blueprint needs to be opened | 15:29 |
| david-lyle | that would be the higher priority item | 15:29 |
| nikhil_k | and possibly redefine the BP | 15:30 |
| TravT | ok, i'd like to close this as implemented for local cli with searchlight-manage. and open a new one for what we just discussed. that one wouldn't be liberty though. too late i think. | 15:30 |
| nikhil_k | +1 | 15:30 |
| TravT | aye or nay? | 15:30 |
| *** sigmavirus24 is now known as sigmavirus24_awa | 15:30 | |
| nikhil_k | lol | 15:31 |
| rosmaita | aye | 15:31 |
| *** openstackgerrit has quit IRC | 15:31 | |
| TravT | done | 15:31 |
| *** openstackgerrit has joined #openstack-searchlight | 15:32 | |
| TravT | https://blueprints.launchpad.net/searchlight/+spec/set-up-functional-tests | 15:32 |
| TravT | Clearly still targetted for liberty | 15:32 |
| TravT | My question is whether this is high or critical? | 15:32 |
| *** sigmavirus24_awa is now known as sigmavirus24 | 15:33 | |
| TravT | would we release without the functional tests? | 15:33 |
| nikhil_k | hmm, I am a little out of sync on the regressions. so, depending on the level of pain , it could go either way | 15:33 |
| lakshmiS | i would say critical since we cannot do regression test properly without it for bugs | 15:33 |
| nikhil_k | glance overall test coverage was 30 odd % at the beginning of L | 15:34 |
| TravT | that's pretty awesome | 15:34 |
| nikhil_k | functional tests even less | 15:34 |
| nikhil_k | so, we could have a few important ones | 15:35 |
| TravT | well, i guess my opinion is we should have something. we don't have to have 100%. | 15:35 |
| nikhil_k | yes | 15:35 |
| lakshmiS | we should atleast have major functionality code paths | 15:35 |
| rosmaita | i agree | 15:35 |
| nikhil_k | echo all over :) | 15:35 |
| david-lyle | definitely | 15:35 |
| TravT | Ok, I changed to essential | 15:36 |
| TravT | BTW, never got input on whether metadefs could be medium or not | 15:36 |
| TravT | we can come back to it after looking at more. | 15:37 |
| TravT | We have a pair of blueprints related to horizon integration | 15:37 |
| david-lyle | certainly not high as now marked | 15:37 |
| TravT | https://blueprints.launchpad.net/searchlight/+spec/horizon-integration-images and https://blueprints.launchpad.net/searchlight/+spec/horizon-integration-instances | 15:37 |
| TravT | david-lyle: there are a few paths we can pursue on this. | 15:38 |
| david-lyle | from my perspective the priorities are in reverse order | 15:38 |
| david-lyle | but I understand the reasoning | 15:38 |
| TravT | I agree on that, but was thinking from a practical point of view | 15:38 |
| TravT | NG images table is in | 15:38 |
| TravT | NG instances isn't | 15:39 |
| sigmavirus24 | nikhil_k: I've been secretly working to make glance have 1200% test coverage | 15:39 |
| sigmavirus24 | just don't tell anyone | 15:39 |
| david-lyle | but as far as pain point instance is 2M x images | 15:39 |
| TravT | sigmavirus24, now wonder you have no time. | 15:39 |
| david-lyle | for the user | 15:39 |
| TravT | so there is another path | 15:39 |
| TravT | the top nav search bar | 15:40 |
| lakshmiS | sigmavirus24: i guess thats what the video call is all about | 15:40 |
| TravT | which steve prototyped a bit of | 15:40 |
| nikhil_k | sigmavirus24: you da man | 15:40 |
| TravT | and we took to our ux people and they got really excited about | 15:40 |
| david-lyle | turn your testing up to ludicrous | 15:40 |
| sigmavirus24 | lakshmiS: sure | 15:40 |
| sigmavirus24 | we'll go with that | 15:40 |
| sigmavirus24 | david-lyle: the testing includes testing the socket module from stdlib so | 15:41 |
| david-lyle | lol | 15:41 |
| sigmavirus24 | also testing the language that elasticsearch is written in =P | 15:41 |
| nikhil_k | you forgot to test the byte codes? | 15:41 |
| nikhil_k | no way! | 15:41 |
| david-lyle | TravT: a pluggable topbar search would be more achievable in Liberty | 15:42 |
| TravT | so, we can switch the priority of the two for searchlight perspective | 15:42 |
| TravT | with pluggable top bar | 15:42 |
| david-lyle | but if it's only images, not sure that warrants being in the topbar | 15:42 |
| TravT | doens't matter so much | 15:42 |
| david-lyle | instances yes | 15:43 |
| david-lyle | but running out of time | 15:43 |
| TravT | the topbar also includes searching for general pages. | 15:43 |
| TravT | yeah | 15:43 |
| sjmc7 | i can get a patch up to do the topbar search early next week if we want to do that, just for people to look at | 15:43 |
| TravT | sjmc7, i don't think it is feasible if you use that other library | 15:43 |
| sjmc7 | ah, right | 15:44 |
| david-lyle | other library? | 15:44 |
| TravT | diana offered to put up an xstatic package for it | 15:44 |
| TravT | selectize | 15:44 |
| sjmc7 | i knew there was something. yeah... one of the javascriptees pointed me at it | 15:44 |
| TravT | it is an angular library | 15:44 |
| david-lyle | more xstatic gah | 15:44 |
| sjmc7 | yeah | 15:44 |
| sjmc7 | so maybe it's not realistic for liberty, but can definitely show something at the summit | 15:44 |
| david-lyle | we really don't like deployers | 15:44 |
| TravT | i'm not convinced it takes bringing in a whole new library | 15:45 |
| david-lyle | we can talk horizon later, | 15:45 |
| TravT | yeah | 15:45 |
| TravT | anyway, it is clear that from horizon perspective instances is higher | 15:45 |
| david-lyle | but instances > images for me | 15:45 |
| TravT | I switched 'em. | 15:46 |
| TravT | but let's talk later. | 15:46 |
| david-lyle | sure | 15:46 |
| TravT | I think I can have a patch on ng-images for support by end of next week. | 15:46 |
| TravT | https://blueprints.launchpad.net/searchlight/+spec/designate-plugin | 15:46 |
| TravT | designate plugin | 15:46 |
| TravT | i think priority of medium is still right | 15:47 |
| sjmc7 | i;'d really like to get this in since endre's put so much work in and the designate team were really enthusiastic | 15:47 |
| TravT | but, i'd really like to get it in... i've just been having to prioritize my time on bugs, nova, etc | 15:47 |
| TravT | sjmc7, do you want to bump it to high? | 15:49 |
| sjmc7 | i think so, yeah | 15:49 |
| TravT | i won't argue... i agree on the effort we've received from that team. | 15:50 |
| sjmc7 | having content is pretty crucial | 15:50 |
| sjmc7 | i'll talk to them about the patch next week (they're all in europe) | 15:50 |
| TravT | there may not be a user for it, but if we get it in, then we could also take advantage from horizon top nav bar | 15:50 |
| david-lyle | I think ducttape_ was interested | 15:51 |
| david-lyle | but may not be ready to fire up searchlight just for that | 15:51 |
| TravT | ducttape_ wants searchlight primarily for instances | 15:51 |
| david-lyle | that too | 15:51 |
| TravT | ok, i'm going to bump it to high, unless we get a nay? | 15:51 |
| TravT | anybody? | 15:51 |
| david-lyle | but apparently the designate plugin to horizon lacks a lot of information | 15:52 |
| TravT | that seems more like horizon plugin issue | 15:52 |
| david-lyle | it may be an API issue | 15:52 |
| david-lyle | can't remember, if API not sure searchlight can fix that | 15:52 |
| * david-lyle stops steering left | 15:53 | |
| TravT | next one targeted for liberty is query-perf-tests | 15:54 |
| TravT | https://blueprints.launchpad.net/searchlight/+spec/query-perf-tests | 15:54 |
| TravT | nikhil_k, will you be doing anything here | 15:54 |
| nikhil_k | yes | 15:54 |
| TravT | or should we bump to next? | 15:54 |
| TravT | ok. | 15:54 |
| nikhil_k | But, I am not sure if these are befitted to be written in python | 15:54 |
| nikhil_k | err, I just didn't say that | 15:55 |
| nikhil_k | tbh, I have some for solr work that I had to do on side | 15:55 |
| TravT | so, leave as liberty and as medium? | 15:56 |
| nikhil_k | may be mark it medium, possibly one or two levels lower than functional ones | 15:56 |
| nikhil_k | yep | 15:56 |
| TravT | https://blueprints.launchpad.net/searchlight/+spec/plugin-config-options | 15:56 |
| TravT | currently medium | 15:56 |
| TravT | nobody is assigned. | 15:57 |
| TravT | oh wait, | 15:57 |
| TravT | endre is | 15:57 |
| TravT | ekarlso^ are you doing any work on plugin config options? | 15:58 |
| TravT | let's leave as is | 15:58 |
| TravT | sjmc7: dont-expose-index-doctype | 15:58 |
| TravT | https://blueprints.launchpad.net/searchlight/+spec/dont-expose-index-doctype | 15:58 |
| TravT | No priority | 15:58 |
| sjmc7 | needs reviews | 15:58 |
| TravT | I think I want to change to high or essential | 15:58 |
| TravT | this is an API impact | 15:58 |
| sjmc7 | i think it's high priority, having done the horizon POC | 15:58 |
| TravT | We are at top of hour, but only a couple more marked as liberty | 15:59 |
| TravT | ok to finish or are people done? | 15:59 |
| sjmc7 | fine to keep going | 15:59 |
| david-lyle | I can stay | 16:00 |
| rosmaita | me too | 16:00 |
| TravT | https://blueprints.launchpad.net/searchlight/+spec/enable-cors | 16:00 |
| TravT | i entered this one | 16:00 |
| TravT | it basically enabled UI's to directly talk to it | 16:00 |
| TravT | not necessary for horizon at the moment. | 16:01 |
| TravT | but will be next release, i think. | 16:01 |
| TravT | i'd like to mark it as low | 16:01 |
| david-lyle | I'd put this at low | 16:01 |
| TravT | k | 16:01 |
| david-lyle | if searchlight and CORS seems like a rather small subset | 16:01 |
| david-lyle | I think optimization comes later :) | 16:01 |
| sjmc7 | agreed :) | 16:02 |
| TravT | o, there is one here that lakshmi has marked as started, but it isn't targeted for liberty | 16:02 |
| TravT | https://blueprints.launchpad.net/searchlight/+spec/named-queries | 16:02 |
| TravT | lakshmiS^ should we keep for liberty or bump to next? | 16:02 |
| lakshmiS | i would have it next week early on. if we can get reviews i would keep it for liberty | 16:03 |
| TravT | ok | 16:03 |
| david-lyle | would be a nice starting point for users, certainly | 16:03 |
| rosmaita | +1 | 16:04 |
| sjmc7 | can we make that a doc thing? | 16:04 |
| david-lyle | sjmc7: just doc? or doc too? | 16:04 |
| TravT | we can doc a number of example queries | 16:04 |
| TravT | should | 16:05 |
| TravT | but let's see what lakshmi has | 16:05 |
| lakshmiS | we should be an ongoing process based on how users use it | 16:05 |
| lakshmiS | s/we/that | 16:05 |
| sjmc7 | doc's easy, api changes hard given timescale, and we don't know what people will want exactly | 16:05 |
| lakshmiS | there are a couple nikhil and rosmaita suggested which i have it done | 16:06 |
| nikhil_k | there was one that came to mind during glance midcycle but now I forgot (hope I mentioned it to someone possiby lakshmiS ) | 16:06 |
| lakshmiS | once we have plugin config blueprint implemented it should be easy to add more queries as needed | 16:07 |
| TravT | well, lakshmi is working on it. we can evaluate further in code reviews. | 16:07 |
| nikhil_k | ++ | 16:07 |
| TravT | i don't even want to ask this. | 16:08 |
| TravT | but https://blueprints.launchpad.net/searchlight | 16:08 |
| nikhil_k | oh, I think it might have to do with signed images | 16:08 |
| TravT | are any that are NOT marked as liberty that should be | 16:08 |
| nikhil_k | but that is still wip | 16:08 |
| TravT | i think our list is already rather optimistic | 16:09 |
| TravT | but if we can get reviews, it is possible. | 16:09 |
| nikhil_k | can't think of anything. but we should try to focus on versioned objs after liberty | 16:09 |
| sjmc7 | a lot of these have patches up already though | 16:09 |
| nikhil_k | and some are small | 16:09 |
| nikhil_k | hopefully | 16:09 |
| lakshmiS | version objs: favorite topic | 16:10 |
| TravT | i can target for next | 16:10 |
| nikhil_k | agree on reviews | 16:10 |
| nikhil_k | if people can ping/email me for reviews, I can definitely take a look | 16:10 |
| nikhil_k | it's hard to be in sync with all reviews and their latest updates | 16:11 |
| lakshmiS | looks like we need review meetings as we get closer :) | 16:11 |
| TravT | here's one right now: https://review.openstack.org/#/c/211047/ | 16:11 |
| TravT | yes, we could do some virtual sprint like meetings | 16:11 |
| TravT | for real time feedback. | 16:11 |
| nikhil_k | lakshmiS: ha yes! review day too :) | 16:11 |
| sjmc7 | i noticed i have about 12 patches up but have only had a couple merged :) | 16:12 |
| nikhil_k | i glanced over it yday but stayed away after looking at big comments :P | 16:12 |
| sjmc7 | one thing we should get better at is WIP-ing, otherwise it gets annoying for reviewers | 16:12 |
| nikhil_k | +1 | 16:13 |
| nikhil_k | in general in os | 16:13 |
| sjmc7 | TravT moans about that constantly :) | 16:13 |
| * TravT tries to look innocent | 16:13 | |
| sjmc7 | but it is polite to mark stuff as "first cut, general comments please" | 16:13 |
| TravT | david-lyle really likes non-wip patches... ;) | 16:13 |
| david-lyle | :/ | 16:13 |
| nikhil_k | my m/c is about to restart... | 16:14 |
| TravT | so, that's it for BPs | 16:14 |
| TravT | i don't think we need to step through bugs today | 16:14 |
| TravT | unless objections | 16:14 |
| TravT | I will put an agenda item for next week to schedule review slots. | 16:15 |
| TravT | if we could maybe have a day or two each week the last few weeks of liberty that people would be available to talk / work on reviews, that could help | 16:16 |
| sjmc7 | bug scrub kind of thing? | 16:16 |
| TravT | bug day would be good too. | 16:17 |
| TravT | I meant just an hour or two. | 16:17 |
| TravT | for live interactions on reviews. | 16:17 |
| lakshmiS | sounds good | 16:17 |
| TravT | okay | 16:17 |
| TravT | will add to agenda for discussion | 16:18 |
| TravT | thanks everybody... of course i'll still be on here. | 16:18 |
| david-lyle | thanks TravT | 16:19 |
| TravT | but we can't basically call this ad-hoc meeting a wrap for now. | 16:19 |
| lakshmiS | thanks guys | 16:19 |
| rosmaita | bye | 16:19 |
| TravT | s/can't/can | 16:21 |
| ekarlso | TravT: yes | 16:32 |
| ekarlso | basically the same pattern that designate has | 16:32 |
| ekarlso | i'll submit it up today. | 16:33 |
| ekarlso | sjmc7: / TravT | 16:33 |
| ekarlso | . | 16:33 |
| sjmc7 | what what what | 16:33 |
| TravT | ekarlso | 16:33 |
| TravT | cool! | 16:33 |
| lakshmiS | that will be awesome to test it out along with named queries | 16:35 |
| *** lakshmiS has quit IRC | 17:06 | |
| *** david-lyle is now known as cbrown | 17:17 | |
| sjmc7 | :D | 17:17 |
| sjmc7 | nobody has any context on that if they're not in the horizon room :) | 17:17 |
| ekarlso | what's that about sjmc7 ? | 17:25 |
| *** TravT_ has joined #openstack-searchlight | 17:50 | |
| *** TravT has quit IRC | 17:50 | |
| *** TravT_ is now known as TravT | 18:44 | |
| *** TravT_ has joined #openstack-searchlight | 19:01 | |
| *** TravT has quit IRC | 19:01 | |
| *** TravT_ is now known as TravT | 21:09 | |
| *** cbrown is now known as david-lyle | 21:20 | |
| *** TravT_ has joined #openstack-searchlight | 22:41 | |
| *** TravT has quit IRC | 22:43 | |
| *** TravT has joined #openstack-searchlight | 23:17 | |
| *** sigmavirus24 is now known as sigmavirus24_awa | 23:20 | |
| *** TravT_ has quit IRC | 23:20 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!