| TravT | if you are here, please add +1 to this: https://review.openstack.org/#/c/212092/ | 00:59 |
|---|---|---|
| TravT | we need it merged so fresh devstack installs work. | 00:59 |
| sjmc7 | it's been merged | 01:32 |
| TravT | oh, good. my bugging them worked. | 01:35 |
| *** lakshmiS has joined #openstack-searchlight | 05:21 | |
| *** lakshmiS has quit IRC | 05:35 | |
| *** GB21 has joined #openstack-searchlight | 07:09 | |
| *** GB21_ has joined #openstack-searchlight | 07:09 | |
| *** GB21_ has quit IRC | 07:11 | |
| *** lakshmiS has joined #openstack-searchlight | 07:21 | |
| *** GB21_ has joined #openstack-searchlight | 07:49 | |
| *** GB21 has quit IRC | 07:53 | |
| *** GB21_ is now known as GB21 | 09:00 | |
| *** GB21 has quit IRC | 09:14 | |
| *** GB21 has joined #openstack-searchlight | 09:24 | |
| *** sigmavirus24_awa is now known as sigmavirus24 | 13:00 | |
| david-lyle | TravT: when you're around ping me re: roles | 13:29 |
| * david-lyle says as he walks afk for a few | 13:29 | |
| david-lyle | ;) | 13:29 |
| *** lakshmiS has quit IRC | 13:30 | |
| TravT | david-lyle: i'm here, but my computer is experiencing some issues this morning. so, you'll probably see my drop in and out of IRC a few times as I restart. | 13:58 |
| TravT | bbiab | 13:58 |
| *** TravT has quit IRC | 13:59 | |
| *** TravT has joined #openstack-searchlight | 14:04 | |
| david-lyle | TravT: so assume I'm illiterate and couldn't read your email, what was you question about roles? | 14:09 |
| david-lyle | you are using openstackclient to verify roles? | 14:09 |
| TravT | hey. | 14:11 |
| TravT | david-lyle | 14:11 |
| TravT | i did a fresh devstack last night | 14:12 |
| TravT | and i was expecting to be able to see via openstack client that the searchlight user has the admin role on the service project | 14:12 |
| TravT | but it doesn't seem to list them. | 14:12 |
| TravT | i'll try a rest request here. | 14:12 |
| david-lyle | restacking with master | 14:31 |
| david-lyle | TravT: there is a tool that helps visualize all these role assignments in a visual manner. You might have core status on it? | 14:32 |
| TravT | sounds interesting. are you saying it has a GUI. maybe even a web interface? | 14:33 |
| david-lyle | something like that | 14:34 |
| david-lyle | :D | 14:34 |
| david-lyle | or when in doubt just mysql | 14:34 |
| TravT | ok, i used this "gui" tool you mentioned | 14:35 |
| TravT | and it is showing searchlight admin in service project. | 14:35 |
| david-lyle | that's correct | 14:35 |
| TravT | so, maybe something weird with openstack client | 14:35 |
| TravT | i'm getting paste deploy errors again. | 14:36 |
| * TravT changed topics | 14:36 | |
| david-lyle | because of bug 968696 admin in service is admin for all as long as your token is scoped to service | 14:37 |
| openstack | bug 968696 in Cinder ""admin"-ness not properly scoped" [Undecided,In progress] https://launchpad.net/bugs/968696 - Assigned to Brent Roskos (broskos) | 14:37 |
| TravT | ahh, the famous bug | 14:38 |
| *** lakshmiS has joined #openstack-searchlight | 14:38 | |
| david-lyle | that's the one | 14:39 |
| * david-lyle back in a minute, but here | 15:00 | |
| TravT | Courtesy Searchlight review session reminder in #openstack-searchlight: lakshmiS, nikhil_k, rosmaita, sigmavirus24, TravT, krykowski, david-lyle, wokuma, sjmc7 | 15:01 |
| * sigmavirus24 waves | 15:02 | |
| lakshmiS | o/ | 15:02 |
| sigmavirus24 | Wayne's not here it seems | 15:02 |
| TravT | no, i think that's okay though | 15:02 |
| TravT | I think others will trickle in | 15:04 |
| TravT | but let me share out the reviews link | 15:04 |
| TravT | Reviews link: http://tinyurl.com/searchlight-reviews | 15:04 |
| TravT | These four are all high / critical / essential | 15:05 |
| TravT | Add 'sort' parameter, Fix RBAC filters, Enable configurable index names, Add nova instances plugin. | 15:05 |
| TravT | I've already gone through the first 3, but maybe lakshmiS / sigmavirus24 could look at them? | 15:06 |
| TravT | I'm in the middle of testing the latest nova patchset | 15:06 |
| lakshmiS | Have been busy with functional test issues but I already tested the two of them so will go ahead and +2 those two | 15:07 |
| TravT | lakshmiS, the elasticsearch devstack package was broken, but it should be fixed now | 15:09 |
| TravT | FYI | 15:09 |
| TravT | https://review.openstack.org/#/c/212092/ | 15:09 |
| TravT | ceilometer team found it and submitted fix, but it wasn't getting love | 15:09 |
| TravT | so i started bugging cores on devstack | 15:09 |
| sigmavirus24 | sorry, not sorry that I -1'd something with a +2 | 15:10 |
| lakshmiS | TravT: thx for the link | 15:10 |
| TravT | sigmavirus24, which one? | 15:11 |
| david-lyle | sort | 15:12 |
| lakshmiS | https://review.openstack.org/#/c/206268/ | 15:12 |
| TravT | don't be sorry about -1. that's why we have multiple reviewers. | 15:12 |
| sigmavirus24 | sorry, jumping between a video call and reviews | 15:13 |
| TravT | no worries... | 15:13 |
| sigmavirus24 | I don't know if we want to let ES handle that kind of junk | 15:13 |
| sigmavirus24 | But it seems like we would do well by our users to error early since we're already type checking | 15:13 |
| nikhil_k | a bit late, sorry. was on a urgent call | 15:15 |
| sjmc7 | hi, apologies for being late | 15:15 |
| TravT | hey sjmc7 see sigmavirus24's comment on https://review.openstack.org/#/c/206268/ | 15:15 |
| openstackgerrit | Merged openstack/searchlight: Enable configurable index names https://review.openstack.org/207672 | 15:15 |
| TravT | i think it makes sense | 15:15 |
| sjmc7 | yeah, i will take a look | 15:17 |
| sjmc7 | there will be things we can't catch, and we should also look at how we handle errors from e-s | 15:17 |
| rosmaita | TravT: sorry i'm late ... anything simple for me to look at? | 15:20 |
| rosmaita | or complicated, if you don't need a quick response | 15:21 |
| TravT | hi rosmaita: | 15:21 |
| TravT | Here's all the outstanding reviews: | 15:21 |
| TravT | http://tinyurl.com/searchlight-reviews | 15:21 |
| rosmaita | ty | 15:21 |
| TravT | i do have a question on nova for you | 15:22 |
| rosmaita | sure | 15:22 |
| TravT | In this file: | 15:22 |
| TravT | https://review.openstack.org/#/c/198852/15/searchlight/elasticsearch/plugins/nova/servers.py | 15:22 |
| TravT | line 31 | 15:22 |
| TravT | for sjmc7 too | 15:22 |
| nikhil_k | TravT: thanks for the dashboard link, I reviewed a couple easy ones that had jenkins +1 | 15:22 |
| TravT | is hostId really not sensitive? | 15:23 |
| TravT | or should we be filtering it as well. | 15:23 |
| nikhil_k | sjmc7: or TravT : does anyone want me to rebase https://review.openstack.org/#/c/212563/ ? | 15:23 |
| sjmc7 | TravT i looked into it... | 15:23 |
| nikhil_k | or someone already taking care of that? | 15:23 |
| sjmc7 | it looks like hostId is a hash of project+host name | 15:23 |
| sjmc7 | and is judged not sensitive. it's also probably not very useful, so i'm fine redacting it | 15:23 |
| rosmaita | sjmc7: +1 -- it's obscured | 15:23 |
| TravT | ok. | 15:24 |
| rosmaita | actually, it is useful | 15:24 |
| rosmaita | you can tell if you're on the same host or not | 15:24 |
| TravT | sjmc7, maybe you could make your inline comment have that info | 15:24 |
| sjmc7 | ok | 15:24 |
| TravT | nikhil_k: up to sjmc7, but I'd be fine with you rebasing that one | 15:24 |
| rosmaita | but that's just you, you can't use someone else's hostId to tell if you're on the same host as them to conduct a denial of resources attack | 15:24 |
| sjmc7 | rebase away, nikhil_k | 15:25 |
| rosmaita | but you can tell if your instances are same host or not, which sometimes you care about | 15:25 |
| nikhil_k | sjmc7: aight! | 15:25 |
| TravT | sjmc7: would you actually mind making the nova patch depend on that RBAC patch? | 15:26 |
| sjmc7 | nope, i'll do that | 15:27 |
| TravT | sjmc7: lakshmiS: | 15:27 |
| TravT | https://review.openstack.org/#/c/198852/15/searchlight/elasticsearch/plugins/nova/servers_notification_handler.py | 15:27 |
| TravT | line 97 | 15:27 |
| lakshmiS | TravT: why the dependency | 15:27 |
| TravT | lakshmiS: because i test all these different combinations, including making instances in different projects | 15:28 |
| TravT | and it would be nice to have that rbac one in front of it. | 15:28 |
| lakshmiS | but the RBAC patch only does glance | 15:29 |
| TravT | oh, let me see. | 15:29 |
| TravT | you are right | 15:29 |
| TravT | nevermind: sjmc7 | 15:30 |
| TravT | so, back to | 15:30 |
| TravT | https://review.openstack.org/#/c/198852/15/searchlight/elasticsearch/plugins/nova/servers_notification_handler.py | 15:30 |
| TravT | line 97 | 15:30 |
| TravT | in glance images, we moved from doc as upsert | 15:30 |
| TravT | but i think i found a different error that perhaps one of you could triage | 15:31 |
| TravT | https://bugs.launchpad.net/searchlight/+bug/1486781 | 15:31 |
| openstack | Launchpad bug 1486781 in OpenStack Search (Searchlight) "Glance images not searchable after update notification" [Undecided,New] | 15:31 |
| openstackgerrit | Steve McLellan proposed openstack/searchlight: Add 'sort' parameter https://review.openstack.org/206268 | 15:31 |
| TravT | sigmavirus24: ^ | 15:31 |
| sjmc7 | TravT - the nova patch implements RBAC in the same way as that patch already, so that one doesn't need a change | 15:32 |
| TravT | sjmc7: i didn't get all the way through nova testing. | 15:33 |
| TravT | but since you have to update it again, i'm going to hold off. | 15:33 |
| TravT | i did all the testing on previous patches and it worked. | 15:33 |
| TravT | and you addressed most my comments. several of them we agreed should be addressed as future patch sets. | 15:34 |
| TravT | but i will re-test once others look through it and you address my comments / their comments. | 15:34 |
| sjmc7 | ok, taking a look at the nova one now | 15:34 |
| TravT | i'll revalidate the sort parameter now with you latest change. | 15:35 |
| TravT | s/you/your | 15:35 |
| TravT | sjmc7: as a process, would you mind putting a reply of done to comments you address in previous patches (e.g. the sigmarvirus24 comment). It is helpful to see when going back through reviews. I added some on your nova patch yesterday. | 15:37 |
| sjmc7 | yeah, i try to. i will be more diligent | 15:38 |
| sjmc7 | so TravT - we're not doing doc_as_upsert at the moment, pending that other bug? | 15:40 |
| openstackgerrit | Nikhil Komawar proposed openstack/searchlight: Allow indexing of single type or index https://review.openstack.org/210759 | 15:40 |
| TravT | sjmc7: i'm not sure what our best solution is. | 15:41 |
| sjmc7 | ok. i'll keep the nova patch consistent with what's there | 15:41 |
| nikhil_k | TravT: sjmc7 lakshmiS : fixing merge on this too https://review.openstack.org/#/c/212563/ :P | 15:41 |
| sjmc7 | i haven't had a chance to look at it | 15:41 |
| lakshmiS | as per document both update and index does replace the existing doc | 15:41 |
| TravT | nikhil_k, cool. Thanks! | 15:41 |
| TravT | lakshmiS: index wasn't seeming to make the doc searchable. | 15:42 |
| TravT | or I had a bad configuration. | 15:42 |
| TravT | i can retest as well, but if you could try it out. | 15:42 |
| lakshmiS | TravT: what i was trying to say is the new document using "update" and "index" results in exactly same new verion of doc. but anyway let me test it | 15:42 |
| sjmc7 | no, that's not necessarily true | 15:43 |
| sjmc7 | update will leave properties around even if they no longer exist | 15:43 |
| sigmavirus24 | TravT: sjmc7 left another -1 | 15:43 |
| TravT | is there another complementary doc_as_upsert. | 15:43 |
| TravT | to remove properties? | 15:43 |
| sjmc7 | index | 15:44 |
| lakshmiS | sjmc7: that's right but the process of deleting the old doc, and getting the content inserted in the new doc should be same | 15:44 |
| TravT | that's what it is using | 15:44 |
| TravT | but at least in my tests on glance, search wasn't working immediately | 15:44 |
| TravT | there is a refresh option | 15:44 |
| TravT | but that didn't seem to change things, and from docs isn't necessarily recommended. | 15:45 |
| TravT | thanks sigmavirus24: | 15:45 |
| TravT | maybe you can give the nova patch a once over? | 15:45 |
| sjmc7 | no, refresh must be used with care | 15:45 |
| TravT | yeah, perf impact | 15:45 |
| TravT | david-lyle: rosmaita: maybe you can look at this one. Not too big. | 15:48 |
| TravT | https://review.openstack.org/#/c/200225 | 15:48 |
| david-lyle | TravT: looking | 15:54 |
| sjmc7 | nikhil_k, you said you're rebase 212563? | 15:56 |
| nikhil_k | sjmc7: yes, doing that now | 15:57 |
| sjmc7 | ok. i'll hold off the nova one since it's dependent | 15:57 |
| sjmc7 | this is fun, it feels like those group projects in college you left until the night before it was due | 15:57 |
| TravT | sjmc7: funny enough, i think this style of collaboration is how i've done most things in openstack. | 15:59 |
| TravT | not the procrastination part, but the everybody pitching in | 15:59 |
| sjmc7 | well.. :) | 16:00 |
| sjmc7 | partly the procrastination :) | 16:01 |
| TravT | reminds me, i did exchange email with thierry on project planning with liberty. have added that to thursday meeting agenda. | 16:02 |
| TravT | lakshmiS: i'm seeing the index searchability thing still on latest devstack. | 16:06 |
| TravT | have you been able to reproduce? | 16:06 |
| lakshmiS | i will have to switch to my linux boot after this meeting to test. going through docs, there seems to be an auto refresh of 1 sec by default | 16:07 |
| sjmc7 | yeah, that sounds familiar | 16:08 |
| TravT | well, it is much longer than 1 second. i could screenshare to show you what i'm doing and seeing. | 16:08 |
| sjmc7 | i remember when writing tests in the past we put delays of 1-2 seconds | 16:08 |
| sjmc7 | that sounds like a different issue | 16:08 |
| sjmc7 | the only time i've noticed delays is under VERY high i/o or when doing artificial tests (PUT/GET immediately following one another) | 16:09 |
| sjmc7 | incidentally, GET by id should always work | 16:09 |
| sjmc7 | so you might want to test that | 16:09 |
| lakshmiS | ok let me test it. will be right back | 16:09 |
| *** lakshmiS has quit IRC | 16:09 | |
| TravT | it'll return the result if you don't filter at all | 16:09 |
| sjmc7 | ah, ok. then it's definitely something else :) | 16:10 |
| TravT | it is when you filter (by name for example) that it doesn't return | 16:10 |
| openstackgerrit | Merged openstack/searchlight: Clear cached os-client auth on unauthorized https://review.openstack.org/200225 | 16:10 |
| TravT | but then if I do a fresh searchlight-manage index sync | 16:10 |
| sjmc7 | might be tokenized vs untokenized | 16:10 |
| TravT | same query works | 16:10 |
| sjmc7 | hmm. that's a bit concerning | 16:10 |
| sjmc7 | i'll try and take a look later | 16:10 |
| TravT | so, let's eliminate Travis mistakes and have you look over my shoulder later. | 16:11 |
| sjmc7 | i doubt it's a refresh thing | 16:11 |
| sjmc7 | :) | 16:11 |
| sjmc7 | ok | 16:11 |
| TravT | i'll try a few other ways as well. | 16:11 |
| openstackgerrit | Nikhil Komawar proposed openstack/searchlight: Fix RBAC filters https://review.openstack.org/212563 | 16:12 |
| nikhil_k | TravT: this need very minor changes. can we merge it and follow up with another one or can someone make the changes you know? https://review.openstack.org/#/c/197909/ | 16:17 |
| nikhil_k | <tests are also passing? | 16:18 |
| nikhil_k | > | 16:18 |
| TravT | hmm... let me see. | 16:18 |
| TravT | i haven't actually tried it out. | 16:18 |
| TravT | actually line 63 in this file | 16:19 |
| TravT | https://review.openstack.org/#/c/197909/2/contrib/vagrant/Vagrantfile | 16:19 |
| TravT | makes me think it wouldn't work | 16:19 |
| TravT | nikhil_k, maybe you could just co-author it and make the fixes? | 16:20 |
| sjmc7 | sigmavirus24's got his approvin' stick out now! | 16:25 |
| sjmc7 | i'll rebase and push up the nova patch once 212563 lands | 16:25 |
| *** lakshmiS has joined #openstack-searchlight | 16:27 | |
| david-lyle | so I've been looking at https://review.openstack.org/#/c/200225/4 which had me look at https://github.com/openstack/searchlight/blob/master/searchlight/elasticsearch/plugins/openstack_clients.py | 16:28 |
| sjmc7 | before you continue, i did mention i wanted people to look at this a while ago, but continue | 16:29 |
| david-lyle | my question is: is the intent of https://github.com/openstack/searchlight/blob/master/searchlight/elasticsearch/plugins/openstack_clients.py for integrating with python-openstackclient | 16:29 |
| david-lyle | ? | 16:29 |
| sjmc7 | oh. no | 16:29 |
| sjmc7 | it's to provide APIs to the plugins | 16:29 |
| sjmc7 | the reason it's not IN each plugin | 16:29 |
| nikhil_k | TravT: hmm, lemme think about it. May be sometime today.. | 16:30 |
| sjmc7 | is that at the time, nova needed glance | 16:30 |
| david-lyle | def register_cli_opts() ? | 16:30 |
| nikhil_k | I will ping here if/once I start working on it | 16:30 |
| sjmc7 | it now doesn't, so could be self contained if we never have interdependencies | 16:30 |
| TravT | nikhil_k, i just went back through the vagrant patch and saw a few other things. left more comments. | 16:30 |
| lakshmiS | TravT: new image and update image are available in search results immediately | 16:30 |
| nikhil_k | awesome, thanks TravT | 16:30 |
| TravT | lakshmiS: are you filtering? | 16:30 |
| sjmc7 | oh, that... that was to register the credentials settings david-lyle | 16:30 |
| TravT | lakhsmiS: I see them in match all, but not in filters. | 16:31 |
| lakshmiS | no. just listing all results. any specific filter your tried? | 16:31 |
| TravT | let me give example. | 16:31 |
| TravT | i created an image with a name of listener-image | 16:31 |
| TravT | notification received | 16:32 |
| david-lyle | sjmc7: confusing collection of functions to me, but ok, just trying to wrap my head around the intent | 16:32 |
| TravT | and is in unfiltered results | 16:32 |
| TravT | If I do a query like this: | 16:32 |
| TravT | { | 16:32 |
| TravT | "index": "glance", | 16:32 |
| TravT | "type": "image", | 16:32 |
| TravT | "query": { | 16:32 |
| TravT | "query_string": { | 16:32 |
| sjmc7 | it evolved a bit, i'm sure it could do with improvement | 16:32 |
| TravT | "fuzziness": "AUTO", | 16:32 |
| TravT | "query": "name:listener-image" | 16:32 |
| TravT | } | 16:32 |
| TravT | } | 16:32 |
| TravT | } | 16:32 |
| TravT | no results | 16:32 |
| TravT | if i re-run searchlight-manage index sync | 16:32 |
| sjmc7 | david-lyle - there may well be a better way to do it | 16:32 |
| lakshmiS | ok let me try that | 16:32 |
| TravT | and re-execute above, i get the result | 16:32 |
| david-lyle | sjmc7: I'll keep walking through it, just for my own edification | 16:34 |
| TravT | nikhil_k: in my comments on vagrant patch, i put in what the glance, nova, neutron conf should be in devstack. | 16:34 |
| TravT | i think a separate patch could be to just add a note to the glance rst about the glance conf setting for devstack | 16:35 |
| TravT | e.g., just put a little devstack section and say: | 16:35 |
| TravT | add the following to devstack local.conf for notifcations to work: | 16:36 |
| TravT | [[post-config|$GLANCE_API_CONF]] | 16:36 |
| TravT | [DEFAULT] | 16:36 |
| TravT | notification_topics = notifications, searchlight_indexer | 16:36 |
| TravT | without that, people have to manually restart g-api after devstack is done. | 16:36 |
| sjmc7 | sure, david-lyle. call out anything looks fishy | 16:36 |
| sjmc7 | i agree with that TravT, it's a pain | 16:37 |
| lakshmiS | TravT: that query works for me | 16:38 |
| TravT | okay... that's very odd. | 16:38 |
| TravT | is your elasticsearch installed from the devstack package? | 16:38 |
| lakshmiS | nope | 16:38 |
| TravT | maybe bug or config diff? | 16:39 |
| lakshmiS | time to compare the conf files | 16:39 |
| TravT | would you be able to do a hangout with me, so I could show you. Maybe I'm making a mistake. | 16:40 |
| lakshmiS | sure | 16:40 |
| TravT | https://plus.google.com/hangouts/_/gwnwvxgab57pompc63selrfvzua | 16:40 |
| TravT | but just a minute. | 16:41 |
| TravT | i have a meeting in a few minutes that I want to try to get out of | 16:41 |
| lakshmiS | ok | 16:41 |
| ekarlso | hey guys | 16:53 |
| ekarlso | you about ? | 16:53 |
| nikhil_k | TravT: oops, comp restart... thanks for the details, makes sense! | 16:53 |
| ekarlso | https://review.openstack.org/#/c/199099/ < the comment there regarding records | 16:54 |
| ekarlso | anyone wanna help me work that out ? | 16:54 |
| openstackgerrit | Steve McLellan proposed openstack/searchlight: Add nova instances plugin https://review.openstack.org/198852 | 16:56 |
| sjmc7 | ekarlso sure. the comment from graham? | 16:56 |
| mugsie | ekarlso: sorry, meant to ping you on that | 16:57 |
| sjmc7 | speak of the devil | 16:57 |
| mugsie | o/ | 16:58 |
| ekarlso | the records thing.. | 16:58 |
| ekarlso | mugsie: .. | 16:58 |
| mugsie | just definie it as an object | 16:58 |
| mugsie | (i wouold like to have searchlight take the structiured data at some point) | 16:59 |
| ekarlso | mugsie: meaning ? | 16:59 |
| sjmc7 | https://www.elastic.co/guide/en/elasticsearch/reference/current/mapping-nested-type.html | 17:00 |
| sjmc7 | you can leave properties blank if you don't know what they are | 17:00 |
| ekarlso | kk | 17:01 |
| mugsie | ekarlso: you also need to rename that file | 17:02 |
| sjmc7 | if you know you're going to have arrays of objects, i'd strongly consider nested, although because they become separate documents there are some querying weirdnesses | 17:02 |
| sjmc7 | if you ever want to be able to do record.a = 1 AND record.b = 2, nested is the only way to do it | 17:03 |
| ekarlso | what u think mugsie ? | 17:03 |
| mugsie | nested | 17:03 |
| TravT | lakshmiS: here's the local.conf we're talking about in hangout | 17:06 |
| TravT | http://paste.openstack.org/show/427232/ | 17:06 |
| lakshmiS | thx | 17:07 |
| TravT | probably you'll want to comment out a couple things like designate | 17:07 |
| TravT | and ceilometer | 17:07 |
| TravT | just for speed | 17:07 |
| TravT | up to you | 17:07 |
| TravT | for designate, you have to install the plugin first | 17:07 |
| TravT | so, probably comment that out | 17:07 |
| openstackgerrit | Merged openstack/searchlight: Fix RBAC filters https://review.openstack.org/212563 | 17:19 |
| mugsie | sjmc7: https://www.elastic.co/guide/en/elasticsearch/reference/current/mapping-array-type.html | 17:22 |
| mugsie | i thought you said there was no array typer | 17:22 |
| *** lakshmiS has quit IRC | 17:23 | |
| sjmc7 | there isn't :) | 17:25 |
| sjmc7 | it's saying e-s knows how to map lists you send it | 17:26 |
| sjmc7 | you can't define explicit arrays/lists | 17:26 |
| sjmc7 | the simplest thing is to not worry about this until you need to. you could decide not to explicitly map records at all and let E-S handle it | 17:27 |
| sjmc7 | the only slight concern i have with what we spoke about is that if you have the possibility of two records with the same fields but different types, you may get problems | 17:28 |
| sjmc7 | for instance, 'record': {'a': 1} and later 'record': {'a': 'something else'} | 17:28 |
| *** _sigmavirus24 has joined #openstack-searchlight | 18:17 | |
| *** sigmavirus24 has quit IRC | 18:17 | |
| *** _sigmavirus24 is now known as sigmavirus24 | 18:20 | |
| *** sigmavirus24 has joined #openstack-searchlight | 18:20 | |
| openstackgerrit | Steve McLellan proposed openstack/searchlight: More efficient result filtering https://review.openstack.org/207682 | 18:24 |
| *** GB21 has quit IRC | 19:14 | |
| *** rosmaita is now known as brianr-afk | 19:34 | |
| *** TravT has quit IRC | 19:58 | |
| *** TravT has joined #openstack-searchlight | 20:01 | |
| openstackgerrit | Steve McLellan proposed openstack/searchlight: Add 'sort' parameter https://review.openstack.org/206268 | 20:30 |
| *** brianr-afk is now known as rosmaita | 21:42 | |
| ekarlso | you guys wanna test out a patch ? | 22:22 |
| *** bitblt has joined #openstack-searchlight | 23:28 | |
| *** bitblt has quit IRC | 23:28 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!