| *** openstackgerrit has quit IRC | 00:01 | |
| *** openstackgerrit has joined #openstack-searchlight | 00:02 | |
| *** openstack has joined #openstack-searchlight | 00:05 | |
| openstackgerrit | Steve McLellan proposed openstack/searchlight: Correct facet error when no mapping defined https://review.openstack.org/228639 | 00:58 |
|---|---|---|
| sjmc7 | TravT - i will add another patch for recordset facets, but then i'm probably done | 00:59 |
| sjmc7 | anything higher priority? | 00:59 |
| sjmc7 | i can start a little early tomorrow | 00:59 |
| openstackgerrit | Steve McLellan proposed openstack/searchlight: Add facet terms for dns recordsets https://review.openstack.org/229205 | 01:08 |
| sjmc7 | that one is on top of the updated_at one because i added the tests in it | 01:08 |
| sjmc7 | ok, i'm out for now | 01:12 |
| *** sballe has quit IRC | 01:56 | |
| TravT | thanks sjmc7: the designate updates both look good. | 03:13 |
| ekarlso | hey | 05:33 |
| ekarlso | what's up gang ? | 05:33 |
| *** lakshmiS has joined #openstack-searchlight | 06:26 | |
| *** GB21 has joined #openstack-searchlight | 06:48 | |
| *** lakshmiS has quit IRC | 08:32 | |
| *** lakshmiS has joined #openstack-searchlight | 11:37 | |
| *** GB21_ has joined #openstack-searchlight | 11:40 | |
| *** GB21 has quit IRC | 11:44 | |
| *** lakshmiS_ has joined #openstack-searchlight | 11:54 | |
| *** GB21_ has quit IRC | 11:56 | |
| *** lakshmiS has quit IRC | 11:58 | |
| *** lakshmiS_ has quit IRC | 12:17 | |
| *** lakshmiS has joined #openstack-searchlight | 12:43 | |
| openstackgerrit | Merged openstack/searchlight: Add 'id' to metadefs https://review.openstack.org/229110 | 12:48 |
| *** sballe has joined #openstack-searchlight | 12:48 | |
| *** GB21 has joined #openstack-searchlight | 13:11 | |
| *** sigmavirus24_awa is now known as sigmavirus24 | 13:59 | |
| sjmc7 | hi ekarlso | 14:12 |
| openstackgerrit | Lakshmi N Sampath proposed openstack/searchlight: Fix for updating image index on member notifications https://review.openstack.org/228998 | 14:13 |
| *** lakshmiS has quit IRC | 14:54 | |
| TravT | ekarlso: would you have time to take a look at the comments on https://review.openstack.org/#/c/223318/ today? | 14:57 |
| TravT | we are probably going to cut RC1 this week, and hopefully not have another. | 14:58 |
| TravT | sjmc7: are you around? | 15:47 |
| sjmc7 | in body | 15:47 |
| sjmc7 | what's up? | 15:47 |
| TravT | take a look at my comments here: https://review.openstack.org/#/c/228998/3/doc/source/plugins/glance.rst | 15:47 |
| TravT | but, i'm really wondering if we should go ahead and add an "update-only" flag to index sync | 15:48 |
| sjmc7 | the one i took out? :) | 15:48 |
| sjmc7 | yeah, can do. it's not a big change at all | 15:48 |
| TravT | well, similar... | 15:48 |
| TravT | here's what i'm thinking it would do... maybe it was what you had. | 15:48 |
| sjmc7 | i've got two hours of meetings coming up so i can do it during them, though i have some internal fires today too | 15:48 |
| TravT | it would basically refresh everything with latest info. | 15:49 |
| sjmc7 | what i had before was a reindex without deleting anything | 15:50 |
| TravT | i guess that works... it just didn't delete old stuff | 15:50 |
| TravT | to do that, it would have to do some diff | 15:50 |
| sjmc7 | well, it's the same as a reindex except that you assume nothing got deleted | 15:51 |
| TravT | the delete part is definitely nice to have, but it actually has some complications with race conditions i think. | 15:52 |
| TravT | that i'm not sure how to do. | 15:52 |
| sjmc7 | right, that's really all it is - not creating that window where there are no results | 15:52 |
| TravT | yes that, but i was thinking the other end of it. | 15:53 |
| TravT | if you were to "just update" | 15:53 |
| TravT | to know what you can delete you have to have definitive list from source (eg. glance API) and list from current ES index | 15:53 |
| TravT | i mean you could issue a query to delete IDs not in the list... | 15:54 |
| TravT | but the race condition i'm thinking of is there could be notifications coming in while you are doing that | 15:54 |
| sjmc7 | deleting is not something e-s likes doing | 15:55 |
| TravT | e.g. "Glance give me list of all image ids" --> new image created and notified --> "ES delete everything not in list" | 15:55 |
| sjmc7 | it's FAR better to recreate the index | 15:55 |
| TravT | this is actually where aliases probably would help | 15:56 |
| TravT | search requests -> alias | 15:56 |
| TravT | reindex -> notifications point to upcoming index and index is rebuilt -> alias changed to newly created index | 15:57 |
| TravT | no downtime | 15:57 |
| TravT | something like that | 15:58 |
| sjmc7 | yes. if we expect to do a lot of reindexing we can add something to do it | 15:58 |
| TravT | or actually probably... | 15:58 |
| TravT | that obviously is a bigger change we should design into mitake | 15:59 |
| sjmc7 | rather than me cobbling it in today? ;) | 15:59 |
| TravT | but because of this glance notification thing, maybe just adding back the refresh is still desirable. | 15:59 |
| sjmc7 | it feels a bit grubby because we took it out for good reason | 16:00 |
| sjmc7 | but i don't have a good alternative | 16:00 |
| TravT | i'm just trying to think, how could we do this in a way that leads to what I describe above... | 16:01 |
| TravT | --update-only | 16:01 |
| TravT | maybe | 16:01 |
| TravT | and in mitaka underneath the covers whether doing --update-only or not we could still look at the alias thing. | 16:02 |
| TravT | only difference is one clears mappings and the other doesn't? | 16:02 |
| sjmc7 | longer term i think they'd be the same thing | 16:02 |
| sjmc7 | reindex would always do a switcheroo | 16:02 |
| TravT | yeah... | 16:02 |
| sjmc7 | but i can't do that today | 16:02 |
| TravT | no, that takes some thought | 16:02 |
| TravT | i think aliases are our future. | 16:03 |
| TravT | i guess --keep-mapping might be a backwards compatible way | 16:04 |
| TravT | or something like that | 16:04 |
| openstackgerrit | Lakshmi N Sampath proposed openstack/searchlight: Fix for updating image index on member notifications https://review.openstack.org/228998 | 16:05 |
| TravT | sjmc7: just looking at the milestone board, i also see https://bugs.launchpad.net/bugs/1493975 | 16:10 |
| openstack | Launchpad bug 1493975 in OpenStack Search (Searchlight) "Need separate config sections for api and listener" [Medium,New] - Assigned to Steve McLellan (sjmc7) | 16:10 |
| sjmc7 | yeah. i really want to get that in | 16:10 |
| sjmc7 | but it's probably not crucial | 16:10 |
| TravT | from a deployer standpoint, it seems that would be pretty important. | 16:11 |
| sjmc7 | yeah | 16:12 |
| sjmc7 | i had to read up on oslo.config | 16:12 |
| sjmc7 | i don't think it's much work | 16:12 |
| sjmc7 | but it's also wednesday | 16:12 |
| TravT | well, we do have a possible workaround for the glance notification issue... but i really don't know how workable of a solution that is for most deployers. | 16:13 |
| sjmc7 | yeah.. it depends how widely used the members thing is | 16:13 |
| TravT | to carry a random upstream patch... | 16:13 |
| sjmc7 | if it merges soon into mitaka it's an easier sell | 16:13 |
| TravT | an ability to force periodic updates without downtime as a backup to notifications, albeit not perfect, might be more important... | 16:14 |
| sjmc7 | ok, adding that now. --update-only ? | 16:31 |
| TravT | well, here's some possibles | 16:33 |
| sjmc7 | i'm likely to pick the first one, so choose wisely :) | 16:33 |
| ekarlso | looking.. | 16:34 |
| TravT | --update-only --keep-mappings --no-delete | 16:34 |
| TravT | --wtf | 16:34 |
| TravT | i'm partial to that one | 16:34 |
| sjmc7 | :) | 16:34 |
| sjmc7 | no-delete seems clear | 16:34 |
| TravT | yeah, seems reasonable. | 16:35 |
| TravT | although -wtf is nice as well | 16:36 |
| sjmc7 | that does something else! | 16:36 |
| TravT | we could just have -wtf be a floating option. | 16:36 |
| TravT | each release it will be updated to do whatever stupid stuff we need to hack in to deal with things we need it to do | 16:36 |
| * TravT is kidding for those reading along at home... unless you think its a good idea | 16:37 | |
| TravT | ok, new devstack deploy done... now will resume poking and prodding | 16:38 |
| ekarlso | TravT: so, how should the exchange / topic be configured sjmc7 TravT ? | 16:48 |
| TravT | ekarlso: i'm guilty of not fully going through that review... i'll look through it now | 16:50 |
| TravT | i had seen sjmc7 and lakshmiS had a bunch of comments, so was holding off | 16:51 |
| * ekarlso uses a cattleprod on TravT | 16:51 | |
| sjmc7 | ekarlso - my preference would be that exchange is configurable as a default | 16:51 |
| sjmc7 | (searchlight_listener) | 16:52 |
| sjmc7 | but overridable by plugins if you have a weird setup | 16:52 |
| sjmc7 | or rather; overridable per service | 16:52 |
| sjmc7 | the exchange/topic for a given service will always be the same | 16:52 |
| sjmc7 | a good start would be a config section for the various services, then the plugins refer to a service in that group, so eg | 16:53 |
| sjmc7 | [notifications] | 16:53 |
| sjmc7 | nova = searchlight_listener/nova | 16:53 |
| sjmc7 | glance = searchlight_listener/glance | 16:53 |
| sjmc7 | the nova plugin doesn't list topics directly but instead refers to the nova notification config | 16:54 |
| sjmc7 | however, that might bea lot of work at this stage | 16:54 |
| ekarlso | hmmms | 16:54 |
| ekarlso | let me take a stab... | 16:54 |
| ekarlso | wife's gonna kill me thouyg : | 16:54 |
| sjmc7 | yeah, i think that's too much | 16:54 |
| sjmc7 | for now, i think being able to set the exchange at a global level would be a good start | 16:55 |
| ekarlso | sjmc7: uhm, so what if we structure it so | 16:55 |
| ekarlso | plugins define a __service__ and a __resource__ ? | 16:55 |
| ekarlso | or is that stupid | 16:55 |
| sjmc7 | i don't want to do any big restructure now | 16:55 |
| sjmc7 | we're trying to do the release today | 16:55 |
| ekarlso | okok :p | 16:55 |
| sjmc7 | maybe we should leave it and do it properly in mitaka | 16:56 |
| TravT | sjmc7: well, we're voting on releasing in tomorrow's IRC meeting | 16:56 |
| ekarlso | so... | 16:56 |
| ekarlso | I can hammer this tonight... | 16:56 |
| sjmc7 | ok... but there is not really any time for a review cycle | 16:57 |
| ekarlso | meh.. | 16:57 |
| ekarlso | not really that big uf a change though, if you add __resource__ and __service__ to each plugin and change get_docuemnt_type() to return OS::%s::%s __service__ __resource__ | 16:58 |
| ekarlso | it's allmost identical | 16:58 |
| ekarlso | I dunno what youy chieftains think... | 16:58 |
| sjmc7 | i don't think that's what i mean | 16:58 |
| ekarlso | theeen, where else would you put the service name etc? | 16:59 |
| sjmc7 | i don't think i can explain it properly right now | 16:59 |
| sjmc7 | it's not about the service name | 16:59 |
| sjmc7 | it's about the notification parameters | 16:59 |
| ekarlso | hmmmms | 16:59 |
| sjmc7 | what would __resource__ and __service__ do? | 16:59 |
| ekarlso | you'd look for a section notifications:nova f ex | 17:00 |
| ekarlso | but dunno if that's the correct | 17:00 |
| TravT | i'm kinda thinking we should go through all the plugins and put up sample configurations for each plugin. | 17:00 |
| TravT | and make sure it looks like what we want for our current use cases. | 17:01 |
| TravT | just an etherpad or something. | 17:01 |
| TravT | to show desired state. | 17:01 |
| sjmc7 | what we want is to be able to change 'searchlight_listener' and the 'topic' without modifying code, but ideally only once per service | 17:02 |
| TravT | we also have other config options for different resource types that are currently hardcoded. | 17:02 |
| sjmc7 | i think we may have to backburner it for mitaka | 17:03 |
| TravT | which really should only mean a few days. | 17:03 |
| TravT | and then ekarlso doesn't have to feel like he has to kill himself tonight. | 17:04 |
| openstackgerrit | Steve McLellan proposed openstack/searchlight: Add --no-delete to indexing https://review.openstack.org/229526 | 17:04 |
| ekarlso | so wtf to cleanup then.. | 17:04 |
| ekarlso | good gawd : P | 17:04 |
| sjmc7 | :D | 17:04 |
| sjmc7 | TravT ^ there's your --no-delete | 17:05 |
| ekarlso | then there's the q of | 17:05 |
| ekarlso | sjmc7: anyways, since you can't come to a conclusion i'll run off. | 17:05 |
| ekarlso | wife's waiting :p | 17:05 |
| sjmc7 | ekarlso :) ok. thanks for all your hard work | 17:05 |
| ekarlso | ideally though I think you'd configure noticications:nova, notifications:foo | 17:06 |
| ekarlso | or alike | 17:06 |
| TravT | https://etherpad.openstack.org/p/searchlight-plugin-config-options | 17:06 |
| ekarlso | maybe poke into how Ceilometer is doing it ? | 17:06 |
| TravT | yeah | 17:06 |
| TravT | i put up that etherpad, we can try to design out what the config should look like | 17:06 |
| TravT | and then munge code based on it | 17:07 |
| ekarlso | also by the looks you'll be supporting Kafka too going forward... | 17:07 |
| *** lakshmiS has joined #openstack-searchlight | 17:07 | |
| sjmc7 | what what | 17:07 |
| ekarlso | https://blueprints.launchpad.net/oslo.messaging/+spec/adding-kafka-support < thgat | 17:07 |
| TravT | ooh, interesting | 17:07 |
| sjmc7 | no, that's why we use oslo messaging :) | 17:07 |
| ekarlso | sjmc7: haha, wehy ? :p | 17:07 |
| sjmc7 | so we hopefully don't have to care too much | 17:08 |
| TravT | yeah monasca does this funny thing | 17:08 |
| ekarlso | sjmc7: ya, but take a look at how ceilometer does options | 17:08 |
| TravT | they take notifications from rabbit and then shove them into kafka | 17:08 |
| ekarlso | since they alledgedly support rmq / kafka | 17:08 |
| sjmc7 | umm | 17:08 |
| *** nikhil is now known as nikhilkomawar | 17:09 | |
| *** nikhilkomawar is now known as nikhil_k | 17:09 | |
| *** nikhil_k is now known as nikhil | 17:09 | |
| TravT | thanks for the heads up on that ekarlso | 17:09 |
| *** nikhil is now known as nikhil_ | 17:09 | |
| *** nikhil_ is now known as nikhil__ | 17:09 | |
| *** nikhil__ is now known as nikhil_k_ | 17:10 | |
| *** nikhil_k_ is now known as nikhil | 17:10 | |
| ekarlso | the sad thing atm is that you have all resource plugins 1 pr resource that dont give you a easy way to setup options pr service | 17:10 |
| ekarlso | idealy you'd maybe setup options 1 pr service regardless of resources ? | 17:11 |
| TravT | perhaps that makes sense, even though some options are resource type specific. | 17:11 |
| ekarlso | a | 17:12 |
| ekarlso | hmmmms | 17:12 |
| TravT | but IMO i have been thinking that we need to do a pass through what we have in mitaka to see what opportunities we have to centralize more logic / config | 17:12 |
| ekarlso | :P | 17:12 |
| TravT | make plugin dev as easy as possible. | 17:13 |
| TravT | before we get 30 more | 17:13 |
| TravT | ekarlso: i echo what sjmc7 said earlier. i really appreciate your help on everything here. we have a number of things that we wouldn't have had without your work. | 17:15 |
| sjmc7 | yeah. now we have some it's easier to see that | 17:15 |
| nikhil | hey guys, do you have any simple bugs for an intern to work on (something for learning and not too urgent) ? | 17:21 |
| sjmc7 | the day before release? :) | 17:21 |
| sjmc7 | not right now, but i'm sure we can come up with some mitaka stuff in the next few days | 17:22 |
| TravT | well, i do have one | 17:22 |
| TravT | https://bugs.launchpad.net/searchlight/+bug/1479190 | 17:22 |
| openstack | Launchpad bug 1479190 in OpenStack Search (Searchlight) "Start the API Service in tox venv failed" [Medium,New] | 17:22 |
| sjmc7 | ah, that's a good idea | 17:23 |
| nikhil | :) | 17:23 |
| nikhil | cool | 17:23 |
| *** itisha has joined #openstack-searchlight | 17:24 | |
| nikhil | itisha: hi, meet the searchlight team here | 17:27 |
| TravT | hello itisha! | 17:27 |
| itisha | nikhil, TravT: hi :) | 17:28 |
| nikhil | TravT: sjmc7 ekarlso david-lyle rosmaita sigmavirus24 : itisha is the prospective outreachy intern for december cycle | 17:28 |
| TravT | cool! | 17:28 |
| sjmc7 | hi! | 17:28 |
| TravT | welcome to openstack! | 17:28 |
| nikhil | itisha: those are the most active guys. and TravT is our PTL (Project Team Lead) | 17:28 |
| itisha | TravT: thanks! | 17:28 |
| itisha | sjmc7: hi :) | 17:28 |
| TravT | we try to keep things pretty light around here... | 17:29 |
| nikhil | :) | 17:29 |
| TravT | so if you see odd commentary, it probably is intended to be humorous. | 17:29 |
| TravT | unless it is something bad about sjmc7, then we probably mean it | 17:29 |
| * sigmavirus24 waves to itisha | 17:29 | |
| nikhil | itisha: you will find it easier to work here than most other projects so, this is good place to learn. | 17:29 |
| * sigmavirus24 goes back to trying to get his head out of his team's product | 17:30 | |
| lakshmiS | hi itisha | 17:30 |
| nikhil | :P | 17:30 |
| TravT | sigmavirus24: hi! | 17:30 |
| itisha | TravT: ok, cool :P | 17:30 |
| nikhil | ohai lakshmiS is online today so late! | 17:30 |
| sigmavirus24 | hi TravT | 17:30 |
| lakshmiS | yeah getting some reviews done | 17:30 |
| itisha | sigmavirus24, lakshmiS : hi :) | 17:30 |
| nikhil | itisha: lakshmiS is one of the founders too (among active guys) | 17:30 |
| TravT | lakshmiS is a rock star | 17:30 |
| nikhil | ++ | 17:31 |
| itisha | nikhil: ohh ok, cool :) | 17:31 |
| lakshmiS | i guess that was supposed to be humorous too :) | 17:31 |
| TravT | itisha: we're trying to get the first release candidate done this week. | 17:31 |
| TravT | so, pretty heads down the next few days... | 17:31 |
| TravT | but something that actually would be great is you going through the docs and seeing if they work for you and where there is confusion. | 17:32 |
| TravT | having a new person look at docs and getting things setup is great way to help us gain that perspecitve | 17:32 |
| TravT | and you can then propose updates / fixes | 17:32 |
| itisha | TravT: ok, sounds good :) | 17:33 |
| TravT | itisha, i have a doc review up right now as well. | 17:34 |
| TravT | which you should feel free to review and see if it helps you as you go through things | 17:34 |
| TravT | https://review.openstack.org/#/c/228664/ | 17:34 |
| nikhil | itisha: also please feel free to file a bug for docs and work on it as you find it fitting | 17:35 |
| ekarlso | itisha: you're brave to work on openstack :P | 17:35 |
| itisha | ekarlso: thanks, i guess :P | 17:36 |
| *** akanksha_ has joined #openstack-searchlight | 17:36 | |
| nikhil | ekarlso: she is pretty smart | 17:36 |
| nikhil | had her first glance review in just 3 odd days | 17:36 |
| ekarlso | nikhil: well isn't glance a easy one ? :p | 17:37 |
| * TravT thinks ekarlso has never worked on glance | 17:37 | |
| nikhil | :) only if.. | 17:37 |
| nikhil | just look at some of the latest ML threads | 17:37 |
| itisha | nikhil: thanks :) | 17:37 |
| TravT | i think the motto for glance should be "duck and run" | 17:37 |
| ekarlso | :O | 17:38 |
| lakshmiS | ekarlso should look at the onion layer technology in glance | 17:38 |
| nikhil | LOL | 17:38 |
| * nikhil feels lakshmiS' pain | 17:38 | |
| lakshmiS | :) | 17:38 |
| ekarlso | sjmc7: you handling the config review then ? | 17:40 |
| sjmc7 | ekarlso yes, i'm not sure if we'll merge it for the RC | 17:44 |
| *** lakshmiS has quit IRC | 17:46 | |
| ekarlso | sjmc7: rather just fix it up for M1 ? | 17:47 |
| sjmc7 | think so. clarify exactly what we want to do | 17:48 |
| ekarlso | ok | 17:51 |
| *** lakshmiS has joined #openstack-searchlight | 17:58 | |
| *** openstackgerrit has quit IRC | 18:16 | |
| *** openstackgerrit has joined #openstack-searchlight | 18:17 | |
| sjmc7 | TravT - check mappings | 18:20 |
| sjmc7 | with 228639 | 18:20 |
| TravT | sjmc7: i know the mapping is out of date, but i thought applying the patch would mean i don't get a 500 error | 18:20 |
| TravT | but i'm still getting 500 | 18:20 |
| TravT | lakshmiS, just put some comments on https://review.openstack.org/#/c/229058 | 18:24 |
| lakshmiS | i did | 18:24 |
| TravT | I just did i mean | 18:25 |
| TravT | maybe we're slightly talking about different things. | 18:25 |
| TravT | i'm not sure, so take a read and then we can chat | 18:25 |
| lakshmiS | oh let me look at your comment | 18:25 |
| sjmc7 | if the mapping is wrong, you'll get a different error | 18:30 |
| sjmc7 | but that's an exception from elasticzsearch | 18:30 |
| TravT | so what you are saying is that should perhaps be a different bug fix | 18:31 |
| TravT | ? | 18:31 |
| sjmc7 | possibly.. i don't want to start hiding exceptions though | 18:31 |
| sjmc7 | a missing index/mapping is one thing | 18:31 |
| sjmc7 | a wrong one is more serious | 18:32 |
| TravT | no, hiding is bad. i think we could catch and give troubleshooting tips, but i won't consider that in this review since the bug fix is something different | 18:33 |
| lakshmiS | TravT: just looked through your comments | 18:34 |
| lakshmiS | i think we both are talking the same thing here | 18:34 |
| lakshmiS | regarding the scenarios you gave: User in project A creates image projectAimage as private. Admin gets a token scoped to project A. Admin lists images. Admin would see projectAImage. | 18:35 |
| lakshmiS | what i am saying is admin and user both go through the same RBAC filter and the output for both of them would be same when all_projects is false | 18:36 |
| TravT | yes, i think that is correct. | 18:36 |
| TravT | by default you get results for the project as if you are normal... | 18:36 |
| TravT | if we were doing filtering in the rbac queries right now | 18:38 |
| TravT | property filtering that is, then this wouldn't quite work right. | 18:38 |
| TravT | but that isn't being done right now. | 18:38 |
| lakshmiS | so refresh my memory. are users allowed to see each others private images in the same project? | 18:39 |
| TravT | private means private to project | 18:39 |
| lakshmiS | ok. for some reason i was thinking private to user | 18:40 |
| TravT | just a sec... i'm going to review all the queries | 18:40 |
| TravT | recordset is project filter | 18:40 |
| TravT | https://github.com/openstack/searchlight/blob/master/searchlight/elasticsearch/plugins/designate/recordsets.py#L73 | 18:40 |
| TravT | zone is too: https://github.com/openstack/searchlight/blob/master/searchlight/elasticsearch/plugins/designate/zones.py#L77 | 18:41 |
| sjmc7 | they all are | 18:41 |
| TravT | image is too. it is a little harder to maintain | 18:41 |
| TravT | https://github.com/openstack/searchlight/blob/master/searchlight/elasticsearch/plugins/glance/images.py#L84 | 18:41 |
| TravT | but that "owner field" is the project / tenant id | 18:42 |
| TravT | metadefs https://github.com/openstack/searchlight/blob/master/searchlight/elasticsearch/plugins/glance/metadefs.py#L96 | 18:42 |
| lakshmiS | yes | 18:42 |
| TravT | nova is too | 18:43 |
| TravT | https://github.com/openstack/searchlight/blob/master/searchlight/elasticsearch/plugins/nova/servers.py#L134 | 18:43 |
| TravT | i believe we have opportunity in the future to make the get_rbac_filter handle some result filtering as well using _source excludes on nova servers for example | 18:44 |
| TravT | but that isn't how it is done right now | 18:44 |
| TravT | that goes through the https://github.com/openstack/searchlight/blob/master/searchlight/elasticsearch/plugins/nova/servers.py#L137 | 18:44 |
| lakshmiS | yeah but property protection requires post result filtering | 18:45 |
| lakshmiS | but source fields is good candidate | 18:46 |
| *** openstackgerrit has quit IRC | 18:46 | |
| *** openstackgerrit has joined #openstack-searchlight | 18:47 | |
| openstackgerrit | Merged openstack/searchlight: Apply RBAC for admins unless all_projects is given https://review.openstack.org/229058 | 18:51 |
| TravT | btw lakshmiS thanks for putting -1 when you had concerns about the intent of the patch. | 18:51 |
| TravT | always better to discuss if needed | 18:51 |
| lakshmiS | yeah i guess my mind doesnt work well at night ;) | 18:52 |
| sjmc7 | my mind's not working properly at any time of day or night at the moment :) | 18:52 |
| lakshmiS | lol | 18:52 |
| sjmc7 | i appreciate you being around so late, fwiw | 18:52 |
| TravT | i'll let you know if my mind ever works right | 18:53 |
| openstackgerrit | Merged openstack/searchlight: Correct facet error when no mapping defined https://review.openstack.org/228639 | 18:54 |
| *** david-lyle has quit IRC | 18:55 | |
| *** david-lyle has joined #openstack-searchlight | 18:55 | |
| openstackgerrit | Merged openstack/searchlight: Add --no-delete to indexing https://review.openstack.org/229526 | 18:57 |
| openstackgerrit | Merged openstack/searchlight: Default designate updated_at to created_at https://review.openstack.org/229169 | 19:01 |
| openstackgerrit | Merged openstack/searchlight: Add facet terms for dns recordsets https://review.openstack.org/229205 | 19:02 |
| TravT | okay, i'm going to do a complete brand new devstack so i can ensure i have a clean environment for smoke testings all of these changes again now that they've merged. | 19:05 |
| lakshmiS | great. hopefully no more bugs and we will have our first release | 19:07 |
| TravT | yeah, i will spend the afternoon doing queries from horizon plugin / etc | 19:07 |
| sjmc7 | there's still a few hours for me to get some bugs in! | 19:07 |
| TravT | here's a doc bug: https://review.openstack.org/#/c/228664/ | 19:08 |
| lakshmiS | yeah i was just running tox on it | 19:09 |
| TravT | lakshmiS: sjmc7: we'll have to make a go / no go on this one | 19:09 |
| TravT | https://review.openstack.org/#/c/228998/ | 19:09 |
| lakshmiS | you mean whether to suggest using cron job instead of this patch? | 19:11 |
| TravT | i mean whether to merge it or not in liberty... | 19:12 |
| lakshmiS | lets talk about pros and cons then | 19:13 |
| sjmc7 | i've got to go back on the roof for half an hour | 19:13 |
| itisha | Hey everyone, i tried setting up searchlight and then running searchlight-manage index sync, but i got an error: ERROR searchlight.cmd.manage [-] Failed to setup index extension zone: public endpoint for dns service in RegionOne region not found | 19:13 |
| TravT | itisha: do you have designate installed? | 19:14 |
| TravT | sjmc7: lakshmiS: this actually points to a problem... | 19:14 |
| lakshmiS | TravT: yes | 19:14 |
| sjmc7 | yeah :) | 19:14 |
| itisha | TravT: no, i dont | 19:14 |
| lakshmiS | i was thinking about having config to enable/disable plugins | 19:14 |
| sjmc7 | yeah, travis has mentioned this too | 19:15 |
| sjmc7 | and endre | 19:15 |
| TravT | itisha: file a bug | 19:15 |
| sjmc7 | that IS something we could get done today | 19:15 |
| TravT | yes | 19:15 |
| itisha | TravT: ok, I'll file the bug now | 19:16 |
| TravT | and let me get you info on setting up designate | 19:16 |
| lakshmiS | itisha: you can ignore that error for now to continue using glance and nova | 19:16 |
| itisha | although i would like to know what actually causes the problem? | 19:16 |
| TravT | https://review.openstack.org/#/c/228664/1/doc/source/plugins/designate.rst,cm | 19:16 |
| TravT | itisha: here's a script i use to install designate | 19:18 |
| TravT | http://paste.openstack.org/show/474919/ | 19:18 |
| *** GB21 has quit IRC | 19:18 | |
| TravT | and i'll give you the local.conf i use | 19:19 |
| lakshmiS | itisha: the problem is due to not having the API server for the corresponding plugin available on your system. you would get the same error for glance plugin if glance was not available | 19:19 |
| TravT | my section of local.conf for designate: http://paste.openstack.org/show/474920/ | 19:20 |
| itisha | lakshmiS: ohh ok, alright | 19:21 |
| *** david-lyle has quit IRC | 19:22 | |
| sjmc7 | that's actually a pretty serious bug; designate's not a core-core service | 19:22 |
| TravT | yeah, i agree | 19:22 |
| TravT | we need to be able to gracefully handle this in any scenario | 19:22 |
| *** david-lyle has joined #openstack-searchlight | 19:22 | |
| sjmc7 | it seems a bit odd to have to duplicate in setup.cfg and then in config | 19:23 |
| TravT | i mean you could just remove it from setup.cfg | 19:23 |
| sjmc7 | right, but people don't - they pip install stuff | 19:24 |
| lakshmiS | well thats required for stevedore to recognize as entry point in setup.cfg | 19:24 |
| sjmc7 | yeah. just seems odd | 19:24 |
| *** TravT has quit IRC | 19:29 | |
| *** TravT has joined #openstack-searchlight | 19:31 | |
| *** david-lyle has quit IRC | 19:35 | |
| openstackgerrit | Lakshmi N Sampath proposed openstack/searchlight: Fix for updating image index on member notifications https://review.openstack.org/228998 | 19:36 |
| *** lakshmiS has quit IRC | 19:37 | |
| *** lakshmiS has joined #openstack-searchlight | 19:43 | |
| *** ChanServ sets mode: +o TravT | 19:51 | |
| *** david-lyle has joined #openstack-searchlight | 19:52 | |
| *** david-ly_ has joined #openstack-searchlight | 20:07 | |
| *** david-lyle has quit IRC | 20:07 | |
| *** david-ly_ is now known as david-lyle | 20:08 | |
| sjmc7 | TravT - what're we going with for this? a list of enabled plugins in searchlight.conf? | 20:13 |
| TravT | maybe a section for each plugin and then an enabled flag? | 20:14 |
| sjmc7 | ok. right now we don't have a section per plugin | 20:15 |
| sjmc7 | i could cannibalize endre's patch | 20:15 |
| TravT | or should we just finish his patch? | 20:15 |
| TravT | delay release until friday? | 20:15 |
| sjmc7 | :) yeah, maybe | 20:15 |
| sjmc7 | where's that etherpad | 20:16 |
| sjmc7 | i feel bad for coopting his changes :) | 20:16 |
| TravT | i put a link from his review | 20:16 |
| TravT | https://etherpad.openstack.org/p/searchlight-plugin-config-options | 20:16 |
| sjmc7 | third time i'll remember it | 20:17 |
| TravT | FYI: i'm also in the horizon driver's meeting right now. so pardon if i take a moment to respond. | 20:17 |
| sjmc7 | no prob. i've put what i envisaged it looking like in there | 20:18 |
| TravT | lakshmiS ekarlso ^ | 20:18 |
| ekarlso | uhm | 20:21 |
| ekarlso | dont do OS_Nova_.. keep it lower caps... | 20:21 |
| ekarlso | if not the config file will make you want to rake your eyes out | 20:21 |
| sjmc7 | ah yes | 20:23 |
| lakshmiS | what is service_notifications | 20:23 |
| sjmc7 | define how nova, glance, designate are configured | 20:25 |
| lakshmiS | i am still not getting it. how does searchlight_listener/glance relate to config | 20:28 |
| TravT | ok, added a couple config ideas | 20:30 |
| sjmc7 | lakshmiS - the notification exchange/topic names | 20:34 |
| TravT | sjmc7: should the log file for a given type be overridable as well? | 20:35 |
| sjmc7 | not for a plugin i don't think. but api/listener, yes | 20:36 |
| *** lakshmiS has quit IRC | 20:37 | |
| *** lakshmiS has joined #openstack-searchlight | 20:41 | |
| TravT | lakshmiS: you probably should get some sleep! | 20:44 |
| lakshmiS | yeah i was about to logout | 20:45 |
| TravT | we'll go over everything in the IRC meeting tomorrow | 20:46 |
| TravT | probably should get that notification item for glance on their meeting | 20:46 |
| lakshmiS | i will address the comments from flavio before that | 20:48 |
| TravT | ok, cool | 20:48 |
| TravT | have a good night | 20:48 |
| lakshmiS | ttyl | 20:49 |
| *** lakshmiS has quit IRC | 20:58 | |
| *** david-lyle has quit IRC | 21:03 | |
| *** david-ly_ has joined #openstack-searchlight | 21:03 | |
| *** david-ly_ is now known as david-lyle | 21:05 | |
| itisha | hey, now i get an error on running stack.sh: http://paste.openstack.org/show/474950/ | 21:08 |
| sjmc7 | hm... i've not seen that one | 21:17 |
| sjmc7 | i've got to fix something on my building while it's still light but i'll take a look at the configuraiton options after that | 21:18 |
| itisha | ok... | 21:25 |
| *** david-lyle has quit IRC | 21:26 | |
| *** david-lyle has joined #openstack-searchlight | 21:29 | |
| *** david-lyle has quit IRC | 21:48 | |
| *** david-lyle has joined #openstack-searchlight | 21:54 | |
| *** david-lyle has quit IRC | 22:15 | |
| *** david-lyle has joined #openstack-searchlight | 22:36 | |
| *** sigmavirus24 is now known as sigmavirus24_awa | 23:26 | |
| *** akanksha_ has quit IRC | 23:38 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!