| ekarlso | TravT: anything I can help with ? | 00:00 |
|---|---|---|
| TravT | ahh, thanks for looking at that. i didn't get a chance yet today | 00:00 |
| TravT | i'm trying read sjmc7's aggregation code... | 00:01 |
| ekarlso | what's that for ? | 00:01 |
| ekarlso | :p | 00:01 |
| TravT | facets... basically displaying doc counts in UIs and such | 00:01 |
| *** sigmavirus24 is now known as sigmavirus24_awa | 00:03 | |
| TravT | ekarlso: we've got a number of BP or bugs with nobody assigned. | 00:04 |
| TravT | this one seems like one you probably would be able to fix quickly: https://bugs.launchpad.net/searchlight/+bug/1469336 | 00:05 |
| openstack | Launchpad bug 1469336 in OpenStack Search (Searchlight) "Errors are returned as html" [Medium,New] - Assigned to Ian Cordasco (icordasc) | 00:05 |
| TravT | FYI, i've got to run for a few moments... brb | 00:06 |
| *** itisha has quit IRC | 00:11 | |
| *** daddyjoseph97 has quit IRC | 00:20 | |
| *** yingjun has joined #openstack-searchlight | 00:36 | |
| *** yingjun has quit IRC | 01:45 | |
| *** yingjun has joined #openstack-searchlight | 01:45 | |
| *** yingjun has quit IRC | 01:50 | |
| *** yingjun has joined #openstack-searchlight | 02:00 | |
| *** bpokorny_ has joined #openstack-searchlight | 02:02 | |
| *** bpokorny has quit IRC | 02:06 | |
| *** bpokorny_ has quit IRC | 02:07 | |
| openstackgerrit | Merged openstack/searchlight: Feature/Bug workflow corrections https://review.openstack.org/245939 | 02:59 |
| *** yingjun has quit IRC | 04:04 | |
| *** yingjun has joined #openstack-searchlight | 04:04 | |
| *** yingjun has quit IRC | 04:09 | |
| *** TravT_ has joined #openstack-searchlight | 04:52 | |
| *** sigmavirus24_awa has quit IRC | 05:00 | |
| *** TravT has quit IRC | 05:00 | |
| *** notmyname has quit IRC | 05:00 | |
| *** sigmavirus24_awa has joined #openstack-searchlight | 05:01 | |
| *** sigmavirus24_awa has quit IRC | 05:02 | |
| *** sigmavirus24_awa has joined #openstack-searchlight | 05:02 | |
| *** sigmavirus24_awa is now known as Guest87553 | 05:02 | |
| *** notmyname has joined #openstack-searchlight | 05:03 | |
| openstackgerrit | Merged openstack/searchlight-specs: Setup Initial Spec Repo Contents https://review.openstack.org/245214 | 05:10 |
| *** sballe has quit IRC | 05:25 | |
| *** sballe has joined #openstack-searchlight | 05:30 | |
| *** lakshmiS has joined #openstack-searchlight | 05:39 | |
| *** bpokorny has joined #openstack-searchlight | 05:43 | |
| *** yingjun has joined #openstack-searchlight | 06:09 | |
| *** bpokorny has quit IRC | 06:09 | |
| openstackgerrit | Li Yingjun proposed openstack/searchlight: Init searchlight client base https://review.openstack.org/245660 | 06:30 |
| *** yingjun_ has joined #openstack-searchlight | 06:33 | |
| *** yingjun_ has quit IRC | 06:34 | |
| *** yingjun_ has joined #openstack-searchlight | 06:35 | |
| *** yingjun has quit IRC | 06:35 | |
| *** dhellmann has quit IRC | 07:01 | |
| *** dhellmann has joined #openstack-searchlight | 07:02 | |
| openstackgerrit | yuntongjin proposed openstack/searchlight-specs: This BP is implementing a notification-forwarding for SearchLight index https://review.openstack.org/246220 | 07:05 |
| *** lakshmiS has quit IRC | 07:24 | |
| *** lakshmiS has joined #openstack-searchlight | 07:25 | |
| *** yingjun_ has quit IRC | 08:10 | |
| *** yingjun has joined #openstack-searchlight | 08:10 | |
| *** lakshmiS_ has joined #openstack-searchlight | 08:26 | |
| *** lakshmiS has quit IRC | 08:30 | |
| *** yingjun has quit IRC | 08:40 | |
| *** lakshmiS_ has quit IRC | 08:48 | |
| *** yingjun has joined #openstack-searchlight | 08:59 | |
| *** yingjun has quit IRC | 09:16 | |
| *** ekarlso has quit IRC | 09:22 | |
| *** openstackgerrit has quit IRC | 09:31 | |
| *** openstackgerrit has joined #openstack-searchlight | 09:32 | |
| *** gb21_ has quit IRC | 09:33 | |
| *** ekarlso has joined #openstack-searchlight | 09:37 | |
| *** ekarlso has quit IRC | 09:38 | |
| *** ekarlso has joined #openstack-searchlight | 09:38 | |
| *** GB21 has joined #openstack-searchlight | 09:45 | |
| *** ekarlso has quit IRC | 09:57 | |
| *** openstackgerrit has quit IRC | 10:31 | |
| *** openstackgerrit has joined #openstack-searchlight | 10:32 | |
| *** ekarlso has joined #openstack-searchlight | 10:36 | |
| *** ekarlso has quit IRC | 10:39 | |
| *** yingjun has joined #openstack-searchlight | 11:19 | |
| *** lakshmiS has joined #openstack-searchlight | 11:49 | |
| *** lakshmiS has quit IRC | 11:54 | |
| *** GB21 has quit IRC | 12:26 | |
| *** GB21 has joined #openstack-searchlight | 12:36 | |
| *** GB21 has quit IRC | 12:45 | |
| *** yingjun has quit IRC | 13:00 | |
| *** yingjun has joined #openstack-searchlight | 13:04 | |
| *** rosmaita_ is now known as rosmaita | 14:01 | |
| *** ekarlso has joined #openstack-searchlight | 14:01 | |
| *** yingjun has quit IRC | 14:11 | |
| *** yingjun has joined #openstack-searchlight | 14:11 | |
| *** yingjun has quit IRC | 14:15 | |
| *** nikhil_k is now known as nikhil | 14:39 | |
| *** ekarlso has quit IRC | 14:40 | |
| *** lakshmiS has joined #openstack-searchlight | 14:54 | |
| TravT_ | o/ | 15:02 |
| *** TravT_ is now known as TravT | 15:02 | |
| david-lyle | o/ | 15:02 |
| sjmc7 | mornin | 15:02 |
| lakshmiS | o/ | 15:02 |
| TravT | david-lyle: lakshmiS: sjmc7: rosmaita: nikhil: we were going to go through BP priorities this morning. Still up for it? | 15:02 |
| TravT | i can practically feel your enthusiasm! :P | 15:04 |
| david-lyle | that's why I'm here :) | 15:04 |
| sjmc7 | my enthusiasm hasn’t had any coffee yet | 15:04 |
| lakshmiS | :) | 15:05 |
| TravT | sjmc7: i have begun the intake on that myself | 15:05 |
| * TravT still looking to install an IV drip line | 15:05 | |
| TravT | i thought maybe we should just touch base on bug priorities first. | 15:05 |
| TravT | https://bugs.launchpad.net/searchlight | 15:05 |
| TravT | sjmc7 and I have been talking about plugin cleanups | 15:06 |
| TravT | he filed a related bug | 15:06 |
| TravT | https://bugs.launchpad.net/searchlight/+bug/1515028 | 15:06 |
| openstack | Launchpad bug 1515028 in OpenStack Search (Searchlight) "Separation between plugin and notification handler isn't clean" [Critical,In progress] - Assigned to Steve McLellan (sjmc7) | 15:06 |
| TravT | which I think is top priority | 15:06 |
| TravT | if we're going to make plugin improvements, i think it should be done sooner than later | 15:06 |
| rosmaita | o/ | 15:06 |
| sjmc7 | there’s a patch up for it | 15:06 |
| sjmc7 | which is ready for feedback, but i’m expecting to have to do some more jiggling on it | 15:07 |
| TravT | i littered it with comments last night sjmc7 | 15:07 |
| sjmc7 | splendid. i will jiggle away | 15:07 |
| TravT | rosmaita: o/ | 15:08 |
| TravT | so, i guess take a look at the bug list first and see if there are any priority questions | 15:08 |
| TravT | one of them doesn't have a priority | 15:08 |
| sjmc7 | we have bugs??! | 15:08 |
| TravT | sjmc7: shocking, right? | 15:09 |
| rosmaita | i agree that 1515028 is a priority, and like sjmc7's analysis | 15:09 |
| TravT | so this one doesn't have a priority. | 15:10 |
| TravT | https://bugs.launchpad.net/searchlight/+bug/1515015 | 15:10 |
| openstack | Launchpad bug 1515015 in OpenStack Search (Searchlight) "Rename default config file to searchlight.conf" [Undecided,In progress] - Assigned to Liyingjun (liyingjun) | 15:10 |
| TravT | i'm tempted to make it a high... just from the standpoint that things like that should change early in the cycle. | 15:10 |
| sjmc7 | yeah.. it’s more of an annoyance than anything, but i agree it’d be better to change it earlier than later | 15:11 |
| lakshmiS | yeah makes sense. | 15:11 |
| sjmc7 | i think there was a patch for that one too, which i didn’t get round to looking at yesterday | 15:11 |
| lakshmiS | but lets check quickly if there is any naming convention for multiple configs | 15:11 |
| TravT | yeah, need to also discuss the relationship with this patch: https://bugs.launchpad.net/searchlight/+bug/1493975 | 15:12 |
| openstack | Launchpad bug 1493975 in OpenStack Search (Searchlight) "Need separate config sections for api and listener" [Medium,In progress] - Assigned to Itisha Dewan (ishadewan07) | 15:12 |
| TravT | so should we still have separate config files or not? | 15:14 |
| sjmc7 | i don’t think we need them right now | 15:14 |
| lakshmiS | i think if its just log entries separate sections are enough but depends on how much more we customize | 15:14 |
| *** Guest87553 is now known as sigmavirus24 | 15:15 | |
| sjmc7 | even then, it’s best to set log-dir | 15:15 |
| TravT | so then does this patch close that defect? | 15:16 |
| TravT | https://review.openstack.org/#/c/232200/ | 15:16 |
| sjmc7 | right now, i don’t think it’s required at all | 15:17 |
| *** akanksha_ has joined #openstack-searchlight | 15:17 | |
| sjmc7 | the usual way to do this is not to specify individual files | 15:17 |
| sjmc7 | but to specify a directory | 15:17 |
| sjmc7 | oslo.logging figures out the name | 15:17 |
| *** itisha has joined #openstack-searchlight | 15:18 | |
| TravT | well, i think the guidance you left in itisha 's patchset and on the bug is a somewhat vague. | 15:19 |
| TravT | if i read the bug, it indicates log_dir | 15:20 |
| sjmc7 | ok, will take another look | 15:20 |
| TravT | if i read the patch set, it seems that the whole bug should be superceded by a patch for separate config files | 15:20 |
| TravT | ok, any other questions on priorities for bugs or do we agree with the current priorities on them? | 15:21 |
| sjmc7 | nope, lgtm | 15:21 |
| TravT | ok | 15:21 |
| lakshmiS | not many bugs. that good | 15:21 |
| TravT | features... | 15:21 |
| TravT | https://blueprints.launchpad.net/searchlight | 15:21 |
| TravT | the feature workflow went through, the base spec repo is set up now | 15:22 |
| * david-lyle is very jealous of the bug # | 15:22 | |
| TravT | david-lyle: lol | 15:22 |
| TravT | any of the blueprints with a priority are generally what I thought the priority should be | 15:23 |
| sjmc7 | so many! | 15:23 |
| TravT | but feel free to disagree / discuss | 15:23 |
| sjmc7 | i filed https://blueprints.launchpad.net/searchlight/+spec/denormalize-nova-image-flavor | 15:24 |
| sjmc7 | but i think it’s just me that’s obsessed by it | 15:24 |
| lakshmiS | https://blueprints.launchpad.net/searchlight/+spec/discuss-notifications-with-service-teams this may needed to split to make it easier for someone to pickup | 15:25 |
| TravT | lakshmiS: yeah, it is pretty vague. | 15:26 |
| sjmc7 | we kind of had the discussion, at least with the nova team, at the summit | 15:26 |
| TravT | i did go to the nova session at the summit to talk about it. | 15:26 |
| sjmc7 | so maybe we can close it | 15:27 |
| TravT | and sjmc7 / briancline did go to the swift team meeting | 15:27 |
| TravT | and lakshmiS talked with ceilometer | 15:27 |
| sjmc7 | let’s close it then | 15:27 |
| TravT | i think maybe we should just close | 15:27 |
| sjmc7 | i think i roughly agree with the priorities there. there are some ‘undefined’ ones | 15:28 |
| lakshmiS | this needs to prioritized - https://blueprints.launchpad.net/searchlight/+spec/index-perf-tests | 15:28 |
| lakshmiS | which will drive other priorities or more bugs | 15:29 |
| lakshmiS | maybe this a duplicate of it https://blueprints.launchpad.net/searchlight/+spec/query-perf-tests | 15:30 |
| TravT | lakshmiS: do you mean you want the priority bumped higher? | 15:31 |
| sjmc7 | my feeling is indexing performance is more of a concern than query performance | 15:32 |
| lakshmiS | yeah close one of them and bump it | 15:32 |
| sjmc7 | i don’t think they’re duplicates though | 15:32 |
| TravT | well, i think they are different. | 15:32 |
| sjmc7 | different ends of the pipe | 15:32 |
| TravT | indexing is the initial indexing and notification indexing | 15:32 |
| lakshmiS | so query performance is about the parser? | 15:33 |
| lakshmiS | oh ok | 15:33 |
| TravT | lakshmiS: probably query perf is more about the searchlight process hop, but maybe also about how you configure things in a large scale env (multiple indexes, # shards, etc) | 15:34 |
| TravT | but not sure | 15:34 |
| TravT | i think the indexing perf tests are probably a bigger deal. | 15:35 |
| lakshmiS | which i guess we are not there yet | 15:35 |
| TravT | well, i bumped indexing perf test to medium. | 15:35 |
| TravT | should it be higher? | 15:36 |
| sjmc7 | i think high is ok | 15:36 |
| sjmc7 | er, medium | 15:36 |
| sjmc7 | that was freudian | 15:36 |
| lakshmiS | i think its important but we dont have a use case so its ok | 15:36 |
| * nikhil sneaks in | 15:37 | |
| TravT | nikhil: we're just talking about the perf tests | 15:37 |
| lakshmiS | right timing | 15:37 |
| sjmc7 | sit at the front of the class, nikhil | 15:37 |
| nikhil | aha | 15:37 |
| nikhil | I used to | 15:38 |
| TravT | *nikhil sneaks out | 15:38 |
| nikhil | then decided that kids aren't so cool there :P | 15:38 |
| TravT | ;P | 15:38 |
| nikhil | are we discussing anything in particular? | 15:38 |
| TravT | priorities on blueprints | 15:38 |
| TravT | i think we're now going to go through those with "undefined" | 15:39 |
| nikhil | TravT: hmm, may be I shouldn't have scrolled back up and read your sneaks out comment to actually do it #D | 15:39 |
| TravT | we were just talking perf tests | 15:39 |
| TravT | which you still have assigned as owner | 15:39 |
| nikhil | I was curious about anything in particular for perf tests | 15:39 |
| TravT | we still want them | 15:39 |
| TravT | :D | 15:40 |
| nikhil | ok, and pythonic ? | 15:40 |
| nikhil | in gate? | 15:40 |
| *** bpokorny has joined #openstack-searchlight | 15:40 | |
| nikhil | (non-voting?) | 15:40 |
| *** bpokorny has quit IRC | 15:40 | |
| TravT | well, i'd be quite happy to get the functional tests in gate and voting | 15:40 |
| TravT | https://blueprints.launchpad.net/searchlight/+spec/zuul-functional-tests | 15:40 |
| nikhil | (plugin specific or congregated ?) | 15:40 |
| *** bpokorny has joined #openstack-searchlight | 15:41 | |
| nikhil | that one seems rather straightforward | 15:42 |
| TravT | nikhil: do you want to own it? | 15:42 |
| nikhil | sure, I can do that today or tomorrow | 15:42 |
| TravT | assigned to you. | 15:43 |
| TravT | ok... on to the undefineds | 15:43 |
| TravT | https://blueprints.launchpad.net/searchlight/+spec/notification-forwarding | 15:43 |
| TravT | the intel team interested in this already did a POC | 15:44 |
| TravT | we just need to review the spec and ensure it goes in a direction that makes sense | 15:45 |
| nikhil | wow, nice interest in notifications! prolly first time I have seen | 15:45 |
| sjmc7 | there’s a lot of interest :) | 15:45 |
| nikhil | TravT: should we copy over the contents to a spec? | 15:45 |
| TravT | nikhil, they did that last night. | 15:46 |
| TravT | https://review.openstack.org/#/c/246220/ | 15:46 |
| nikhil | oh, I didn't see the spec url in the BP. | 15:46 |
| nikhil | let me find it | 15:46 |
| TravT | here's where this makes a lot of sense to me | 15:46 |
| TravT | 2 reasons. | 15:46 |
| * david-lyle needs to step away | 15:47 | |
| TravT | 1) in solutions where searchlight is being used for a searchable cache, if you will, you wouldn't want to receive notifications until the cache is updated | 15:47 |
| nikhil | huh | 15:48 |
| TravT | 2) searchlight supplements the data, so you wouldn't want to repeat that work elsewhere for end user facing notifications. | 15:48 |
| TravT | nikhil: here's an example | 15:48 |
| TravT | let's say horizon is hooked up to searchlight and providing list queries from searchlight | 15:49 |
| TravT | and they hook up horizon to receive updates from notifications | 15:49 |
| TravT | you want to avoid a race condition | 15:49 |
| TravT | where a user gets a notice that say an instance is created | 15:49 |
| TravT | but they go to the list instances page which is populated from searchlight and the instance doesn't show up. | 15:50 |
| TravT | ideally that takes less than a second | 15:50 |
| TravT | but, you'd not want to notify the UI of something that isn't visible in search results yet | 15:50 |
| nikhil | hmm, makes sense on the UI consistency. | 15:51 |
| TravT | anyway, i think since we have gotten a lot of interest in this and there are developers working on it, we could give it an initial priority of medium. We do need to work through the spec though... | 15:52 |
| TravT | thoughts? | 15:52 |
| nikhil | I guess we are saying this is pluggable for horizon/dashboards and not necessarily the flow one expects from outbound notifications? | 15:52 |
| nikhil | I mean what if notifications are exposed to end users | 15:52 |
| *** lakshmiS_ has joined #openstack-searchlight | 15:53 | |
| nikhil | do we expect the operators to populate SL cache first in that case? | 15:53 |
| TravT | there is another use case that the spec authors bring up. | 15:53 |
| TravT | they want to leverage zaqar features for subscription, but they want the notifications to have all the data that searchlight has (which many notifications on the internal bus don't) | 15:54 |
| TravT | so they thought plugging searchlight in then having it forward along to zaqar would allow for this to be accomplished | 15:55 |
| nikhil | hmm | 15:55 |
| TravT | i spoke a bit about this with flavio at the summit and he seemed to think it was reasonable | 15:56 |
| nikhil | I guess I have a couple of pointers on that (may be I should comment on the spec too) | 15:56 |
| *** lakshmiS has quit IRC | 15:56 | |
| TravT | well, we did just set up the spec repo for that. :P | 15:57 |
| sjmc7 | the system works!@ | 15:57 |
| nikhil | 1. are we establishing a contract for the convergence on the internal bus, zaqar and SL notifications (what if someone has secure/sensitive information) | 15:57 |
| TravT | nikhil: i think that should be part of the notification forwarding spec. a configurable filter. | 15:58 |
| nikhil | 2. there are systems that have notification buses not so similar to what is described in the BP | 15:58 |
| TravT | 2, i was thinking a driver like interface... but | 15:59 |
| nikhil | for example, there could be internal service -> auditors -> subscriptions tools -> zaqar | 15:59 |
| TravT | let's guide that on the spec | 15:59 |
| nikhil | yeah | 15:59 |
| TravT | as for priority, do we leave unprioritized or give it an initial priority? | 16:00 |
| nikhil | s/could be/are existing/g | 16:00 |
| lakshmiS_ | it could be high if horizon use case will be implemented for this release | 16:01 |
| * TravT just looking at rest of list | 16:02 | |
| sjmc7 | we have a number of things i think are more important like adding resources and improving stability | 16:03 |
| TravT | I think i'd start it at medium | 16:03 |
| TravT | conditional on spec approval | 16:03 |
| TravT | which really means we have to provide feedback on the specs out there (i have one, steve has one, and this one0 | 16:04 |
| nikhil | :D | 16:04 |
| TravT | any objection to starting it at medium? | 16:04 |
| TravT | going once | 16:05 |
| TravT | going twice | 16:05 |
| TravT | done | 16:05 |
| sjmc7 | you have an alternate career as auctioneer | 16:05 |
| TravT | i'd prefer one as a lottery winner | 16:06 |
| TravT | i sense we are nearing overload for this meeting... but there are a few more to touch on | 16:07 |
| TravT | defer to another time or get through the next couple? | 16:07 |
| lakshmiS_ | i guess just few more | 16:08 |
| TravT | https://blueprints.launchpad.net/searchlight/+spec/denormalize-nova-image-flavor | 16:08 |
| TravT | which has an inter-relationship with https://blueprints.launchpad.net/searchlight/+spec/nova-flavor-plugin | 16:08 |
| sjmc7 | i don’t know how important that is (the first one) | 16:09 |
| sjmc7 | it just annoys me | 16:09 |
| TravT | basically, this says that we can include flavor information (more than just ID) when getting servers | 16:09 |
| TravT | and can search by more than just ID | 16:09 |
| TravT | definitely VERY nice to have | 16:09 |
| sjmc7 | the problem is when they change | 16:09 |
| TravT | but to date we just go in circles on how to handle it | 16:10 |
| sjmc7 | although i would imagine flavors change so seldom we could take the hit | 16:10 |
| sjmc7 | yeah. if it’s just me that it annoys, then we can keep this one on the backburner | 16:10 |
| TravT | it annoys me too... which is why you and I have had so many discussions on it. | 16:10 |
| * nikhil wonders why people are in a bitter mood today. checks twitter, news .. | 16:10 | |
| * TravT wonders why nikhil wonders why people are in a bitter mood | 16:11 | |
| nikhil | there you go again | 16:11 |
| lakshmiS_ | i think we should differ meeting to other time ;) | 16:11 |
| * nikhil bbiab | 16:11 | |
| TravT | i think perhaps nikhil wouldn't think we were in bitter mood if over video | 16:12 |
| TravT | hard to see laughing | 16:12 |
| nikhil | guess what I managed to stumble upon at the same time as your comment TravT https://twitter.com/notmyname/status/666402860028956672 | 16:13 |
| TravT | lol | 16:13 |
| nikhil | irc == open office | 16:13 |
| lakshmiS_ | oh thats who notmyname is.. | 16:14 |
| TravT | everybody wave "hi!" | 16:14 |
| TravT | :D | 16:15 |
| TravT | so, back to the flavor bp | 16:15 |
| TravT | sjmc7: i do want to see this in | 16:15 |
| TravT | just hard to say it is high | 16:15 |
| TravT | but i'd be fine with medium | 16:15 |
| *** GB21 has joined #openstack-searchlight | 16:15 | |
| TravT | everybody is right on cue today | 16:16 |
| sjmc7 | more coffee! | 16:16 |
| sjmc7 | yeah, it’s not crucial, it’s a nice to have | 16:16 |
| TravT | it is one of those things that really puts the polish on search giving you something you simply can't get with the regular APIs (I think) | 16:17 |
| TravT | ok, so https://blueprints.launchpad.net/searchlight/+spec/index-management-api | 16:18 |
| *** bpokorny_ has joined #openstack-searchlight | 16:19 | |
| *** bpokorny has quit IRC | 16:19 | |
| sjmc7 | mmmm… is this common? | 16:19 |
| sjmc7 | a number of services have command line tools for initial setup kind of things | 16:19 |
| TravT | index management api? | 16:19 |
| TravT | well, here's what I'm thinking | 16:20 |
| TravT | actually, let me start with saying I would put this at low and not a mitaka iteam | 16:20 |
| TravT | item | 16:20 |
| sjmc7 | ok. i see the value, but i think it’s a pretty hefty chunk of work | 16:21 |
| TravT | how about I leave undefined, but target for next release | 16:21 |
| TravT | ? | 16:21 |
| lakshmiS_ | that makes more sense | 16:21 |
| sjmc7 | i think we can certainly discuss it this cycle | 16:22 |
| lakshmiS_ | and we can change it back if we get a strong use case from UI/users | 16:22 |
| TravT | Ok. | 16:22 |
| TravT | ok, the only other BP worth discussing today is: https://blueprints.launchpad.net/searchlight/+spec/keystone-projects-plugin | 16:23 |
| TravT | david-lyle: ^ | 16:23 |
| sjmc7 | i can’t see keystone ever sending notifications | 16:24 |
| sjmc7 | which makes it hard to index projects | 16:24 |
| david-lyle | sjmc7: yeah notifications will be hard | 16:24 |
| david-lyle | it would have to be a periodic resync | 16:24 |
| TravT | we have same problem with nova flavors, i think. | 16:24 |
| david-lyle | if at all | 16:24 |
| TravT | and probably other entities. | 16:24 |
| sjmc7 | periodic resync wouldn’t be hard to do; if that’s valuable we could certainly do it | 16:25 |
| TravT | yeah, so probably a BP on that concept should be registered | 16:25 |
| TravT | as dependent on https://blueprints.launchpad.net/searchlight/+spec/zero-downtime-reindexing | 16:25 |
| nikhil | would we be indexing tokens too!? My guess, no. | 16:26 |
| lakshmiS_ | that would be fun | 16:26 |
| nikhil | ha | 16:26 |
| TravT | indexing tokens... | 16:26 |
| david-lyle | no | 16:26 |
| sjmc7 | hahaha, no. | 16:26 |
| david-lyle | no tokens | 16:26 |
| TravT | of course we will! | 16:26 |
| nikhil | it could be valuable for bookkeeping admin only | 16:26 |
| TravT | just say yes! | 16:27 |
| nikhil | but that may not be quite complaint | 16:27 |
| david-lyle | no | 16:27 |
| TravT | david-lyle is no fun. | 16:27 |
| nikhil | I am just kidding with a serious face | 16:27 |
| TravT | ;-P | 16:27 |
| nikhil | irc is fun! :D | 16:27 |
| sjmc7 | i can imagine the conversation with our security guys | 16:27 |
| TravT | lol | 16:27 |
| TravT | we could also forward the tokens out to zaqar | 16:27 |
| * nikhil successfully lightens the mood even more evil :D | 16:27 | |
| sjmc7 | oo! | 16:28 |
| david-lyle | so with fernet, there are no tokens to index, so success? | 16:28 |
| nikhil | TravT: oh yeah :P | 16:28 |
| sjmc7 | fernet is magic! | 16:28 |
| TravT | maybe we could index the fernet keys and forward notifications out when they change | 16:28 |
| david-lyle | where are we? | 16:29 |
| * TravT looks behind him and sees the security police running towards him | 16:29 | |
| TravT | #openstack-searchlight | 16:29 |
| TravT | umm, actually we are through the list | 16:30 |
| TravT | of setting initial priority | 16:30 |
| TravT | we didn't really talk about what series everything should be targeted at | 16:30 |
| TravT | but let's touch base on that later | 16:30 |
| david-lyle | serious note on the keystone projects plugin | 16:30 |
| lakshmiS_ | yeah | 16:31 |
| david-lyle | it's blocked until that functionality actually exists in keystone | 16:31 |
| david-lyle | in v3 | 16:31 |
| TravT | or until we do periodic re-sync | 16:31 |
| david-lyle | no, there's no pagination in keystone on projects | 16:31 |
| david-lyle | can't get a list yet | 16:31 |
| david-lyle | only first API_RESULT_LIMIT | 16:32 |
| david-lyle | so ~500 | 16:32 |
| TravT | i guess we could develop an algorithm to fish out all the results. | 16:32 |
| david-lyle | then you're dead in the water | 16:32 |
| david-lyle | no, we need the pagination of projects, then we could fish other items out | 16:32 |
| sjmc7 | who has more than 500 projects? | 16:32 |
| david-lyle | sjmc7: O_o | 16:32 |
| lakshmiS_ | how about changing the limit in config for once | 16:33 |
| david-lyle | you're public cloud is not fully dead yet | 16:33 |
| david-lyle | *your | 16:33 |
| sjmc7 | it’s just sleeping! | 16:33 |
| sjmc7 | it’s pinin’ for the fjords! | 16:33 |
| david-lyle | it's resting comfortably | 16:33 |
| TravT | ok, i just added notes on that to BP | 16:33 |
| TravT | and changed implementation to "Blocked" | 16:33 |
| david-lyle | the breakthrough was keystone would be willing to consider pagination on projects | 16:34 |
| david-lyle | at the summit | 16:34 |
| sjmc7 | that was around the same time they were talking about doing away with domain scoped tokens | 16:34 |
| TravT | so, what was the end result of both those conversations | 16:34 |
| david-lyle | well killing domain scoped tokens is a weekly waffle | 16:34 |
| TravT | they both affect us | 16:34 |
| david-lyle | horizon will go forward with domain scoped token support | 16:35 |
| *** GB21 has quit IRC | 16:35 | |
| david-lyle | as it's part of the v3 API, it's not going away | 16:35 |
| david-lyle | even though it will look different under the covers | 16:35 |
| david-lyle | for project pagination, our usecase was finally understood, and the possibility opened up | 16:36 |
| david-lyle | but would still need a spec and approval from keystone on adding pagination for projects | 16:36 |
| TravT | likelihood for it to happen in Mitaka? | 16:36 |
| david-lyle | depends how active someone is doing the work | 16:36 |
| TravT | Do we need a spec in searchlight to deal with domains and HMT projects? | 16:37 |
| david-lyle | I'd sign up, but that would end in no real action | 16:37 |
| david-lyle | TravT: I don't think so | 16:37 |
| *** GB21 has joined #openstack-searchlight | 16:37 | |
| david-lyle | not sure what you'd add | 16:37 |
| TravT | me neither | 16:37 |
| david-lyle | domains won't have anything other than projects and users | 16:37 |
| david-lyle | nothing from other services | 16:38 |
| david-lyle | well depending on the HMT result I guess | 16:38 |
| david-lyle | with HMT, I think project id's will still be enough | 16:39 |
| TravT | well, that's good to know. | 16:39 |
| david-lyle | not sure we want to try and unwind the permissions nightmare of looking down the tree | 16:39 |
| TravT | at least nothing for mitaka | 16:39 |
| david-lyle | I don't think any service quite has that worked out yet | 16:39 |
| david-lyle | yeah maybe in N we revisit | 16:40 |
| TravT | ok, then i think for today we can stop | 16:40 |
| sjmc7 | whoa, that was a long session | 16:41 |
| TravT | yeah | 16:41 |
| lakshmiS_ | later... | 16:41 |
| TravT | maybe i'll just point our IRC meeting on thursday to this log. | 16:41 |
| TravT | probably for thursday we'll just have open discussion on current spec reviews | 16:42 |
| *** lakshmiS_ has quit IRC | 16:42 | |
| TravT | allright, thanks guys | 16:42 |
| *** yingjun has joined #openstack-searchlight | 17:06 | |
| *** yingjun has quit IRC | 17:10 | |
| *** GB21 has quit IRC | 17:24 | |
| *** bpokorny_ has quit IRC | 18:32 | |
| *** bpokorny has joined #openstack-searchlight | 18:52 | |
| *** bpokorny has quit IRC | 18:59 | |
| *** bpokorny has joined #openstack-searchlight | 19:01 | |
| *** yingjun has joined #openstack-searchlight | 19:08 | |
| *** yingjun has quit IRC | 19:13 | |
| *** bpokorny has quit IRC | 19:19 | |
| *** bpokorny has joined #openstack-searchlight | 19:19 | |
| *** bpokorny has quit IRC | 19:22 | |
| *** bpokorny has joined #openstack-searchlight | 19:22 | |
| *** ekarlso has joined #openstack-searchlight | 19:24 | |
| *** bpokorny has quit IRC | 19:28 | |
| *** bpokorny has joined #openstack-searchlight | 19:29 | |
| openstackgerrit | Steve McLellan proposed openstack/searchlight: Fix doc_count for nested fields https://review.openstack.org/236043 | 20:21 |
| *** yingjun has joined #openstack-searchlight | 21:02 | |
| *** yingjun has quit IRC | 21:07 | |
| openstackgerrit | Steve McLellan proposed openstack/searchlight: Restructure index and notification classes https://review.openstack.org/245981 | 22:08 |
| sjmc7 | i think i missed a conversation about having a python client for SL. if it’s in the searchlight repo, it won’t be possible to package it separately | 22:10 |
| *** akanksha_ has quit IRC | 22:10 | |
| *** akanksha_ has joined #openstack-searchlight | 22:11 | |
| TravT | sjmc7: i do agree that this should be a separate repo. We'll have to create a new project like all the others | 22:29 |
| david-lyle | why not openstackclient? | 22:36 |
| david-lyle | still a separate repo I guess to function as openstackclient plugin | 22:39 |
| david-lyle | http://docs.openstack.org/developer/python-openstackclient/plugins.html | 22:39 |
| TravT | david-lyle: it is being proposed as a plugin to openstack client | 22:40 |
| TravT | https://blueprints.launchpad.net/searchlight/+spec/openstack-client-search-initial-plugin | 22:40 |
| TravT | maybe we need to ask the openstack sdk folks | 22:40 |
| sjmc7 | it’ll be tough to use if it’s in the server code | 22:40 |
| TravT | yeah, might defeat the purpose | 22:41 |
| TravT | sjmc7: maybe you could do some more project config magic | 22:43 |
| TravT | and get a repo going for it | 22:43 |
| sjmc7 | i do love me some infra | 22:43 |
| TravT | leave a comment on the patch... | 22:44 |
| sjmc7 | ok. we’ll have to be prepared to do the maintenance on it | 22:44 |
| TravT | guess that goes with the territory. | 22:45 |
| sjmc7 | yeah. i’ve been doing everythingh with bash aliases :) | 22:46 |
| TravT | i think we need to look at the commands though in the BP... | 22:46 |
| TravT | would like to figure out how to make the commands easy | 22:47 |
| TravT | i saw in the glance room that after 11 months, glance now supports some limited updated_at and created_at comparison operators. | 22:48 |
| TravT | but they are still nowhere near as powerful as ES native. | 22:49 |
| TravT | bbiab | 22:50 |
| david-lyle | TravT: I don't think it should be in the server code repo | 22:55 |
| sjmc7 | no. i’ll get another repo created | 22:55 |
| *** yingjun has joined #openstack-searchlight | 23:04 | |
| *** yingjun has quit IRC | 23:11 | |
| *** sigmavirus24 is now known as sigmavirus24_awa | 23:20 | |
| sjmc7 | TravT - do we have a stable-main group in gerrit? the ACLs need updating | 23:32 |
| sjmc7 | err, stable-maint | 23:32 |
| *** yingjun has joined #openstack-searchlight | 23:32 | |
| *** yingjun_ has joined #openstack-searchlight | 23:33 | |
| *** yingjun has quit IRC | 23:33 | |
| *** yingjun_ has quit IRC | 23:38 | |
| TravT | sjmc7: no we don't | 23:56 |
| sjmc7 | good, that simplifies things :) | 23:57 |
| TravT | at some point, we should have it for stable branch | 23:57 |
| TravT | but i'm not sure we are there yet. | 23:58 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!