| *** openstackgerrit has quit IRC | 00:01 | |
| *** openstackgerrit has joined #openstack-searchlight | 00:02 | |
| *** itisha has quit IRC | 02:51 | |
| *** nikhil has quit IRC | 04:32 | |
| *** nikhil has joined #openstack-searchlight | 04:32 | |
| *** GB21 has joined #openstack-searchlight | 05:02 | |
| *** GB21 has quit IRC | 05:35 | |
| *** GB21 has joined #openstack-searchlight | 05:50 | |
| *** GB21 has quit IRC | 06:21 | |
| *** nikhil has quit IRC | 06:56 | |
| *** nikhil has joined #openstack-searchlight | 06:56 | |
| *** akanksha_ has quit IRC | 08:38 | |
| *** GB21 has joined #openstack-searchlight | 10:09 | |
| *** GB21 has quit IRC | 11:57 | |
| *** GB21 has joined #openstack-searchlight | 12:36 | |
| *** sigmavirus24_awa is now known as sigmavirus24 | 13:50 | |
| *** lakshmiS has joined #openstack-searchlight | 14:05 | |
| *** lakshmiS_ has joined #openstack-searchlight | 14:27 | |
| *** lakshmiS has quit IRC | 14:31 | |
| *** lakshmiS_ has quit IRC | 14:57 | |
| *** lakshmiS has joined #openstack-searchlight | 14:58 | |
| TravT | Courtesy Searchlight meeting reminder in #openstack-meeting-4: lakshmiS, mclaren, nikhil_k, rosmaita, sigmavirus24, TravT, krykowski, david-lyle, wokuma, kragniz, sjmc7, ekarlso, abhijeetm, itisha | 15:00 |
|---|---|---|
| nikhil | thanks! | 15:01 |
| *** david-ly_ has joined #openstack-searchlight | 15:22 | |
| *** david-lyle has quit IRC | 15:23 | |
| *** david-ly_ is now known as david-lyle | 15:24 | |
| nikhil | TravT: hi, is it okay to term searchlight as "official" openstack project? I'm a bit uncertain on that terminology atm.. | 16:11 |
| TravT | yes, it is an official project. | 16:12 |
| nikhil | sjmc7: lakshmiS anyone ^ ? | 16:12 |
| nikhil | thanks! | 16:12 |
| TravT | was voted in by the TC | 16:12 |
| nikhil | TravT: I always get confused when they change some details | 16:14 |
| TravT | yeah, it is a project, and then they have all these tags | 16:14 |
| nikhil | :) | 16:15 |
| *** lakshmiS has quit IRC | 17:23 | |
| *** itisha has joined #openstack-searchlight | 18:00 | |
| itisha | hey guys, I don't think I completely got what sjmc7 was trying to say on the review https://review.openstack.org/232200 | 19:43 |
| sjmc7 | hi itisha. i checked out the patch, but it doesn't seem to have any effect - that code doesn't get executed at all when i start the api | 19:44 |
| sjmc7 | i may have something configured wrong | 19:44 |
| nikhil | sjmc7: weird, I thought the configs were static content | 19:45 |
| nikhil | that you generate using tox -e genconfig (like oslo command) | 19:45 |
| sjmc7 | i think that generates an example config file? | 19:45 |
| nikhil | sjmc7: yes, tox generates an example config file and I have a notion that operators will use that as a reference for creating their own | 19:49 |
| sjmc7 | ok | 19:49 |
| sjmc7 | but the codebase doesn't support the options | 19:49 |
| nikhil | sjmc7: in fact, in glance we keep the same in tree to support such cases. But enforcing dynamic creation is better consistency wise | 19:49 |
| nikhil | sjmc7: I see | 19:49 |
| nikhil | itisha: are you getting it? | 19:50 |
| sjmc7 | most projects no longer keep it in tree because it gets out of date | 19:50 |
| nikhil | I don't want to move ahead of you and ask questions taht you don't get, itisha. | 19:50 |
| nikhil | sjmc7: true | 19:50 |
| itisha | i'm not sure i get what you mean by keeping it in tree? | 19:50 |
| sjmc7 | glance has a sample config file checked in to git | 19:51 |
| itisha | oh ok.... | 19:51 |
| *** akanksha_ has joined #openstack-searchlight | 19:52 | |
| itisha | ok, i think i'm on the same page till now | 19:52 |
| nikhil | itisha: do you know what to do next? | 19:57 |
| nikhil | itisha: if not, check the code and see which options are actually being used and which aren't. You will then need to modify some of the groups of these options so that code uses them. questions? | 20:00 |
| itisha | nikhil: if we create separate config files for api and listener as sjmc7 suggested, that would mean there would be no default at all right (ie no common configs)? | 20:01 |
| nikhil | itisha: sorry, I need to go to a meeting. bbiab | 20:02 |
| sjmc7 | itisha no - it would mean there'd be a common config file, plus role-specific ones | 20:03 |
| sjmc7 | so searchlight-api.conf and searchlight.conf | 20:03 |
| itisha | oh ok.... | 20:03 |
| sjmc7 | i'm only suggesting that because i know oslo.config supports chaining config files like that, and i don't know if it supports overriding options in the way you're trying to do it | 20:03 |
| TravT | hmmm... | 20:05 |
| TravT | itisha, would you be comfortable asking the oslo folks? | 20:05 |
| TravT | #openstack-oslo | 20:05 |
| itisha | TravT, umm, ok, i can check with them.... | 20:06 |
| TravT | maybe just say: we have two services that currently use the same config file. we want to, however, have some settings separated by service such as the log file. what is the best way to do that. can we do that within the same config file or do we need role specific files? | 20:08 |
| itisha | ok, cool, i'll check with them | 20:09 |
| TravT | try calling out these irc names | 20:13 |
| TravT | dims bknudson | 20:13 |
| TravT | if there isn't a response on irc, then it may take a mailing list question | 20:14 |
| itisha | TravT: hmm, ok, i've used the irc names, i'll wait for a while, and if there is no response, i'll then ask it on the mailing list | 20:19 |
| TravT | sure... | 20:19 |
| TravT | this is kind of a crazy time (end of release and summit coming up), so people are often kind of scrambling | 20:20 |
| itisha | ohh, alright, totally understandable :) | 20:20 |
| sjmc7 | borland left me in the lurch a bit, i am not a happy bunny | 20:20 |
| sjmc7 | nikhil - does glance override log_file ? it has multiple components | 20:28 |
| sjmc7 | how does glance-reg and glance-api separate them? | 20:28 |
| nikhil | sjmc7: they have separate file for both services | 20:28 |
| sjmc7 | totally separate? | 20:28 |
| nikhil | yes | 20:29 |
| sjmc7 | or share a common file? | 20:29 |
| sjmc7 | i think nova does shared + specific ones | 20:29 |
| nikhil | as they are often installed on diff nodes | 20:29 |
| sjmc7 | although.. i can't find evidence of that. did i make it up? | 20:29 |
| sjmc7 | that's disappointing | 20:30 |
| nikhil | https://github.com/openstack/glance/blob/master/etc/glance-registry.conf | 20:31 |
| nikhil | https://github.com/openstack/glance/blob/master/etc/glance-api.conf | 20:32 |
| nikhil | sjmc7: are you talking about something else? | 20:32 |
| sjmc7 | yeah.. hmm. maybe log_dir is the recommended way to do this? | 20:32 |
| sjmc7 | yeah, i thought i'd seen chained config files | 20:32 |
| sjmc7 | but now i think i'm going mad | 20:32 |
| nikhil | :) | 20:37 |
| sjmc7 | perhaps log_dir is the right way to solve this particular issue | 20:37 |
| * nikhil back | 20:49 | |
| nikhil | itisha: let me know if you get stuck somewhere | 20:50 |
| nikhil | itisha: I'm just back from a meeting but have to send a few emails now.. so send me a ping | 20:50 |
| itisha | nikhil: ok, sure, i'll do that | 20:52 |
| *** ig0r_ has joined #openstack-searchlight | 20:58 | |
| *** ig0r_ has quit IRC | 21:03 | |
| *** ig0r_ has joined #openstack-searchlight | 21:04 | |
| *** GB21 has quit IRC | 21:10 | |
| *** ig0r_ has quit IRC | 21:10 | |
| sjmc7 | itisha, TravT - the workaround for this i think is to specify log_dir rather than log_file | 21:38 |
| sjmc7 | i think it still makes sense to add [api] and [listener] sections, but they don't have to try to override the oslo logging stuff | 21:38 |
| sjmc7 | looking at other services, it's quite common to have config for various components all in the same file | 21:39 |
| TravT | sjmc7: sounds like good info. what code changes are needed? | 21:41 |
| sjmc7 | none | 21:43 |
| sjmc7 | log_dir is olso.config's answer to this | 21:43 |
| *** sigmavirus24 is now known as sigmavirus24_awa | 21:43 | |
| sjmc7 | it uses the name of whatever's running __main__ to set the filename | 21:43 |
| sjmc7 | documentation change only | 21:43 |
| sjmc7 | then i think making the config alteration can wait until we have a definite need for it (e.g. an option that cannot be common) | 21:44 |
| sjmc7 | so this is my bad, that BP was somewhat of a red herring | 21:45 |
| TravT | on an unrelated note, i just figured out how to handle delay on events for text query typing. | 21:45 |
| sjmc7 | ah, splendid | 21:46 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!