| *** ddieterly has joined #openstack-searchlight | 00:25 | |
| *** itisha has quit IRC | 00:30 | |
| *** david-lyle_ has joined #openstack-searchlight | 01:23 | |
| *** ddieterly has quit IRC | 01:46 | |
| *** ddieterly has joined #openstack-searchlight | 02:10 | |
| *** Kevin_Zheng has joined #openstack-searchlight | 02:13 | |
| *** david-lyle_ has quit IRC | 02:31 | |
| *** ddieterly has quit IRC | 02:33 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/searchlight: Updated from global requirements https://review.openstack.org/359528 | 03:10 |
|---|---|---|
| *** TravT has quit IRC | 03:13 | |
| *** TravT has joined #openstack-searchlight | 03:14 | |
| *** TravT_ has joined #openstack-searchlight | 03:37 | |
| *** TravT has quit IRC | 03:37 | |
| *** shu-mutou has quit IRC | 05:30 | |
| *** pcaruana has joined #openstack-searchlight | 06:15 | |
| openstackgerrit | Ha Van Tu proposed openstack/searchlight: Config logABug feature for searchlight api-ref https://review.openstack.org/363428 | 06:37 |
| openstackgerrit | Cao Xuan Hoang proposed openstack/python-searchlightclient: TrivialFix: Remove logging import unused https://review.openstack.org/363472 | 08:17 |
| *** yingjun has quit IRC | 09:51 | |
| *** ddieterly has joined #openstack-searchlight | 11:57 | |
| *** Kevin_Zheng has quit IRC | 11:59 | |
| *** ddieterly has quit IRC | 12:02 | |
| *** Kevin_Zheng has joined #openstack-searchlight | 12:04 | |
| *** ddieterly has joined #openstack-searchlight | 12:12 | |
| *** ddieterly has quit IRC | 12:29 | |
| *** openstackgerrit has quit IRC | 12:34 | |
| *** openstackgerrit has joined #openstack-searchlight | 12:34 | |
| *** ddieterly has joined #openstack-searchlight | 13:05 | |
| *** ddieterly has quit IRC | 13:09 | |
| *** matt-borland has joined #openstack-searchlight | 13:36 | |
| *** ddieterly has joined #openstack-searchlight | 13:39 | |
| *** sjmc7 has joined #openstack-searchlight | 13:50 | |
| *** yingjun has joined #openstack-searchlight | 13:50 | |
| *** matt-borland has quit IRC | 14:09 | |
| *** david-lyle_ has joined #openstack-searchlight | 14:37 | |
| *** ddieterly is now known as ddieterly[away] | 14:38 | |
| *** ddieterly[away] is now known as ddieterly | 14:40 | |
| *** david-lyle_ has quit IRC | 14:59 | |
| *** ddieterly is now known as ddieterly[away] | 15:11 | |
| *** ddieterly[away] is now known as ddieterly | 15:11 | |
| *** ddieterly is now known as ddieterly[away] | 15:11 | |
| *** ddieterly[away] is now known as ddieterly | 15:20 | |
| *** pcaruana has quit IRC | 15:22 | |
| *** TravT_ has quit IRC | 15:40 | |
| *** TravT has joined #openstack-searchlight | 15:44 | |
| *** matt-borland has joined #openstack-searchlight | 15:47 | |
| openstackgerrit | Steve McLellan proposed openstack/searchlight: Allow aggregations in search requests https://review.openstack.org/363773 | 16:02 |
| *** ddieterly is now known as ddieterly[away] | 16:28 | |
| *** ddieterly[away] is now known as ddieterly | 16:34 | |
| *** ddieterly is now known as ddieterly[away] | 16:42 | |
| *** ddieterly[away] is now known as ddieterly | 16:49 | |
| openstackgerrit | Steve McLellan proposed openstack/searchlight: Allow aggregations in search requests https://review.openstack.org/363773 | 16:51 |
| *** ddieterly is now known as ddieterly[away] | 16:52 | |
| *** ddieterly[away] is now known as ddieterly | 17:13 | |
| *** itisha has joined #openstack-searchlight | 17:22 | |
| *** ddieterly is now known as ddieterly[away] | 17:22 | |
| *** ddieterly[away] is now known as ddieterly | 17:27 | |
| openstackgerrit | Steve McLellan proposed openstack/searchlight: Add region name to mappings and documents https://review.openstack.org/353681 | 17:29 |
| *** pcaruana has joined #openstack-searchlight | 17:39 | |
| *** ddieterly is now known as ddieterly[away] | 17:43 | |
| *** Kevin_Zheng has quit IRC | 18:06 | |
| *** Kevin_Zheng has joined #openstack-searchlight | 18:11 | |
| *** ddieterly[away] is now known as ddieterly | 18:24 | |
| *** TravT_ has joined #openstack-searchlight | 18:33 | |
| *** TravT has quit IRC | 18:36 | |
| *** ddieterly is now known as ddieterly[away] | 18:41 | |
| *** ddieterly[away] is now known as ddieterly | 18:45 | |
| *** ddieterly is now known as ddieterly[away] | 18:46 | |
| *** david-lyle_ has joined #openstack-searchlight | 18:52 | |
| *** david-lyle_ has quit IRC | 18:52 | |
| *** ddieterly[away] is now known as ddieterly | 18:59 | |
| *** ddieterly is now known as ddieterly[away] | 19:16 | |
| *** ddieterly[away] is now known as ddieterly | 19:20 | |
| *** TravT_ has quit IRC | 19:41 | |
| *** ddieterly has quit IRC | 19:45 | |
| *** TravT has joined #openstack-searchlight | 20:05 | |
| *** TravT_ has joined #openstack-searchlight | 20:05 | |
| *** TravT has quit IRC | 20:10 | |
| *** TravT_ is now known as TravT | 20:20 | |
| TravT | sjmc7: are you okay with adding the example i mentioned here: | 20:29 |
| TravT | https://review.openstack.org/#/c/363773/2/doc/source/searchlightapi.rst | 20:29 |
| TravT | or against it | 20:29 |
| TravT | i can be arm twisted either way | 20:29 |
| TravT | but either way want to get this patch on in | 20:29 |
| sjmc7 | i can add a nested one | 20:30 |
| TravT | okay, cool | 20:30 |
| sjmc7 | i don’t want to go overboard; i don’t like replicating the DSL documentation | 20:30 |
| TravT | that one actually works. | 20:30 |
| TravT | yeah, i'm just thinking that it could help somebody understand it with an example using real data and a real use case | 20:30 |
| sjmc7 | instead of the one i did? ok | 20:31 |
| TravT | in addition | 20:31 |
| TravT | yours is nice and smiple | 20:31 |
| sjmc7 | mmmmm | 20:31 |
| TravT | simple | 20:31 |
| TravT | you could reduce mine to just be images or flavors | 20:31 |
| TravT | not both | 20:31 |
| sjmc7 | ok | 20:31 |
| TravT | just the nested stuff is ugly | 20:32 |
| sjmc7 | those fields really shouldn’t be nested | 20:32 |
| sjmc7 | mauybe we should pop them out now while we have the chance | 20:32 |
| TravT | hmmm... | 20:32 |
| sjmc7 | or duplicate them at least | 20:33 |
| TravT | we did it in case we de-normalized, but also the nova api has it that way | 20:34 |
| TravT | it would make querying easier to duplicate | 20:34 |
| sjmc7 | they never should have been nested; it’s a 1:1 mapping | 20:35 |
| sjmc7 | we could still store it as “flavor”: {“id”: “1”} | 20:35 |
| sjmc7 | so i’d kind of rather not give that example because it highlights a mistake :) | 20:36 |
| TravT | but it is how to get that info | 20:37 |
| TravT | http://developer.openstack.org/api-ref/compute/?expanded=list-servers-detailed-detail | 20:37 |
| TravT | but, i know what you mean | 20:37 |
| sjmc7 | right, but it doesn’t need to be nested | 20:38 |
| TravT | nova will likely decide to change it all in a future microversion anyway | 20:38 |
| sjmc7 | it can still be an object | 20:38 |
| sjmc7 | https://www.elastic.co/guide/en/elasticsearch/reference/current/object.html | 20:38 |
| sjmc7 | “nested” has a specific function to during indexing | 20:38 |
| sjmc7 | for when you’ve got arrays of inner jobjects | 20:39 |
| TravT | ahh, i was thinking differently | 20:39 |
| TravT | see what you are meaning now | 20:39 |
| sjmc7 | i’ve gone all yoda | 20:39 |
| TravT | you mean it is a mistake how we map it | 20:39 |
| sjmc7 | yeah | 20:40 |
| sjmc7 | i bet if you change “nested” to “object” nothing bad will happen | 20:40 |
| TravT | it makes querying quite a bit harder actually | 20:40 |
| sjmc7 | and querying will become magically easier | 20:40 |
| sjmc7 | yes, it does | 20:40 |
| sjmc7 | it’s necessary in some cases | 20:40 |
| sjmc7 | like the port listings | 20:40 |
| TravT | this might even make some things in current UI possible to query | 20:40 |
| TravT | we should file a bug on this and fix all nested mappings | 20:40 |
| TravT | well, all nested mappings that shouldn't be | 20:41 |
| sjmc7 | yeah. usually with this kind of thing you’d map it in two ways; we’re a little constrained by sticking to the APIs. IP addresses i would argue we should’ve pulled out too | 20:41 |
| sjmc7 | i don’t think there are many. the neutron ones were intentional as far as i know | 20:41 |
| TravT | how does double mapping affect scoring? | 20:41 |
| TravT | for a full text query? | 20:41 |
| sjmc7 | it’ll boost those fields | 20:42 |
| TravT | i'd think it would inappropriately boost | 20:42 |
| sjmc7 | but if you’re looking for an IP i’d be ok with it boosting it | 20:42 |
| sjmc7 | also you can negative boost stuff | 20:42 |
| sjmc7 | https://www.elastic.co/guide/en/elasticsearch/reference/current/index-boost.html | 20:42 |
| sjmc7 | hahaha “Why index time boosting is a bad idea" | 20:42 |
| sjmc7 | i think we should file that | 20:43 |
| sjmc7 | i’ve thought about it a couple of times then always forget again | 20:43 |
| TravT | okay, you want to do it? | 20:43 |
| sjmc7 | ok | 20:43 |
| TravT | assign to rc-1 | 20:43 |
| TravT | i'm going to tag milestone 3 today if i can | 20:43 |
| TravT | just 3 reviews looking to complete | 20:44 |
| sjmc7 | ok. let’s leave the aggs documentation then for now | 20:44 |
| sjmc7 | can add stuff later to documentation | 20:44 |
| TravT | okay | 20:44 |
| TravT | I'll change my vote to +2 | 20:45 |
| sjmc7 | i’ll address rick’s comments on the region patch | 20:45 |
| sjmc7 | you want me to mark it RC1? | 20:47 |
| sjmc7 | this bug? | 20:47 |
| TravT | we probably shouldn't rush that one in today. | 20:47 |
| TravT | so, yes | 20:47 |
| sjmc7 | https://bugs.launchpad.net/searchlight/+bug/1619047 | 20:51 |
| openstack | Launchpad bug 1619047 in OpenStack Search (Searchlight) "Unnest nova server flavor and image mapping properties" [High,New] | 20:51 |
| sjmc7 | some more candidates in there | 20:51 |
| TravT | yeah, if you see anything that should be marked, go ahead and do it | 20:52 |
| sjmc7 | RickA-HP: not sure it matters that the types don’t match in the region name | 20:52 |
| sjmc7 | elasticsearch won’t care | 20:52 |
| sjmc7 | there’s a test checking both | 20:52 |
| sjmc7 | did you find a specific problem with it? | 20:53 |
| RickA-HP | sjmc7: If they don't match it will be harder to programatically analyze the results. For example, searchlight-ui. | 20:54 |
| sjmc7 | ? elasticsearch won’t returna list if there’s one value anyway | 20:54 |
| RickA-HP | But searchlight will return a list (with any number of entries) if that is what we are indexing. | 20:56 |
| sjmc7 | will it? | 20:56 |
| sjmc7 | perhaps i am mistaken | 20:56 |
| RickA-HP | WHen I tried it using both the default and an override, this is what I see returned: | 20:56 |
| RickA-HP | "region_name" : "RegionOne" for the default. | 20:56 |
| RickA-HP | "region_name" : ["OverrideTest01"] for the single override value. | 20:57 |
| RickA-HP | This is from Elasticsearch, I meant. | 20:57 |
| sjmc7 | ah, interesting. ok. i don’t really care either way, can make it a list | 20:57 |
| sjmc7 | but i think [None] will cause problems | 20:57 |
| RickA-HP | How about [""] :) | 20:58 |
| sjmc7 | ok, one sec, will make it a list | 20:59 |
| sjmc7 | but this is a problem with elasticsearch in general; it will unexpectedly return lists on you if you’re not careful | 21:01 |
| *** matt-borland has quit IRC | 21:01 | |
| sjmc7 | it seems a bit weird having it named singular if there’s always a list | 21:04 |
| sjmc7 | not really that bothered either way i guess | 21:04 |
| openstackgerrit | Steve McLellan proposed openstack/searchlight: Add region name to mappings and documents https://review.openstack.org/353681 | 21:07 |
| openstackgerrit | Travis Tripp proposed openstack/searchlight-ui: Display correct number of total results https://review.openstack.org/360290 | 22:22 |
| *** TravT has quit IRC | 22:49 | |
| *** yingjun has quit IRC | 23:10 | |
| *** yingjun has joined #openstack-searchlight | 23:11 | |
| *** yingjun has quit IRC | 23:15 | |
| openstackgerrit | Rick Aulino proposed openstack/searchlight: Add "updated_at" field to all mappings. https://review.openstack.org/361562 | 23:18 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!