| *** sballe has quit IRC | 00:42 | |
| *** openstack has joined #openstack-searchlight | 01:26 | |
| *** lakshmiS has joined #openstack-searchlight | 04:23 | |
| openstackgerrit | Lakshmi N Sampath proposed openstack/searchlight: WIP - Fix oslo messaging listener pool name https://review.openstack.org/202392 | 05:57 |
|---|---|---|
| *** lakshmiS has quit IRC | 05:59 | |
| *** lakshmiS has joined #openstack-searchlight | 06:01 | |
| *** lakshmiS has quit IRC | 07:53 | |
| *** lakshmiS has joined #openstack-searchlight | 11:04 | |
| *** lakshmiS has quit IRC | 12:56 | |
| *** sigmavirus24_awa is now known as sigmavirus24 | 14:02 | |
| *** lakshmiS has joined #openstack-searchlight | 14:04 | |
| openstackgerrit | Merged openstack/searchlight: Remove the i18n module of oslo-incubator https://review.openstack.org/202234 | 14:11 |
| *** sballe has joined #openstack-searchlight | 15:15 | |
| *** ig0r_ has quit IRC | 15:40 | |
| lakshmiS | update on ceilometer notifications: chatted with ceilometer team and they seemed to be opposing using pools since they had previous issues with missing notifications. their suggestion is to configure glance to send to multiple topcis as the stacktach project does. | 16:01 |
| lakshmiS | their PTL said he will send me the link to patchset on how stacktach does it currenlty. | 16:02 |
| sjmc7 | ah, ok. that's fine too | 16:02 |
| lakshmiS | I will change it accordingly to our project | 16:02 |
| sjmc7 | thanks. i guess pick a name for now and we'll make it the default whne we get config options | 16:02 |
| lakshmiS | yeah I set it as "searchlight-listener" in the patch | 16:03 |
| sjmc7 | ok. be good to get that fixed so we don't break ceilometer or vice versa :) | 16:04 |
| lakshmiS | yeah. the right way to do is use pools but oslo.messaging needs to be patched first:) | 16:05 |
| lakshmiS | sjmc7: did you have an agenda for the second speaker session? | 16:08 |
| nikhil_k | sjmc7: good time to discuss the nova instances plugin and related confusion? | 16:12 |
| lakshmiS | nikhil_k: is there data that is not provided by nova api which we will need to index? | 16:15 |
| nikhil_k | lakshmiS: I guess not :) My confusion arose out of the brainstorming session on the need for a cli and scope of searchlight-manage | 16:16 |
| lakshmiS | ah | 16:17 |
| nikhil_k | Also, if we have a ton of data in metadefs, it would nice to index them directly from the DB | 16:17 |
| lakshmiS | or add an API to do bulk output | 16:17 |
| nikhil_k | That was the background+controversy in Glance which confused me. At least one benefit of indexing such data was in my mind | 16:17 |
| nikhil_k | yeah, that would be nice if we can avoid joins on metadefs | 16:18 |
| nikhil_k | in future too :) | 16:18 |
| lakshmiS | yes | 16:18 |
| nikhil_k | lakshmiS: anyways, I am not strongly opinionated on indexing metadefs directly | 16:20 |
| nikhil_k | The thing with nova api is that it's really big | 16:21 |
| nikhil_k | basically it has lots of calls | 16:21 |
| nikhil_k | and they have the newer versioning thing done | 16:22 |
| lakshmiS | yeah the versioning is going to be tricky | 16:22 |
| nikhil_k | so would be nice to get some use cases on this particular bp | 16:22 |
| nikhil_k | my concerns about the oslo.versionedobjects become void if we decide to just use the API | 16:23 |
| nikhil_k | I will have to double check on the notifications though | 16:23 |
| lakshmiS | we currenlty dont have multiple versions of data in index so can we not use the latest(default) version on nova. i need to understand more on oslo.versionedojbects though | 16:23 |
| nikhil_k | I guess extensions are being removed from nova so that complexity on API bit is removed | 16:24 |
| nikhil_k | But John mentioned that some relevant info like images metadata and flavor info, copied over to nova system meta (for consistency sake) won;t be available via the API | 16:27 |
| nikhil_k | we would need to confirm with prod managers and horizon use cases for this | 16:27 |
| lakshmiS | that sucks. | 16:27 |
| * nikhil_k stops brain dump | 16:28 | |
| lakshmiS | or have to add an api patchset to nova which i guess will take couple of months(years?) | 16:28 |
| lakshmiS | :) | 16:28 |
| nikhil_k | lakshmiS: heh, I doubt if instance sys metadata would ever be exposed | 16:29 |
| nikhil_k | because nova uses it for instance states and could-internal transitions | 16:30 |
| nikhil_k | hrmmm, may be admin api can expose it | 16:30 |
| lakshmiS | thats a nice idea | 16:31 |
| lakshmiS | since we do admin specific operations too | 16:31 |
| lakshmiS | lakshmiS goes to sleep nowzzzz.... | 16:38 |
| sjmc7 | :) night | 16:39 |
| *** lakshmiS has quit IRC | 16:45 | |
| ekarlso | sjmc7: yoyo | 21:53 |
| ekarlso | u tharr ? | 21:53 |
| sjmc7 | aye | 21:53 |
| sjmc7 | isn't it midnight for you? | 21:54 |
| ekarlso | so regarding options stuff | 21:54 |
| ekarlso | i'll look at it in a few days :) | 21:54 |
| ekarlso | if that's ok | 21:54 |
| sjmc7 | yeah, sure | 21:54 |
| ekarlso | i'm at the cabin atm so internet coverage may vary :P | 21:54 |
| sjmc7 | the cabin?! | 21:55 |
| ekarlso | like out on a lonely island where it's us the dog and 1 other cabin with people ;) | 21:55 |
| sjmc7 | birch twigs and saunas? or is that finnish? | 21:55 |
| ekarlso | lol yeah that's pretty much finnish | 21:56 |
| ekarlso | sjmc7: let me see if I can dig up a picture for ya | 21:57 |
| ekarlso | sjmc7: http://imgur.com/EIeg8C3 | 22:02 |
| sjmc7 | oo. you need a boat! | 22:03 |
| ekarlso | pretty "primitive" but darn nice place! | 22:03 |
| ekarlso | yap, we got that ;) | 22:03 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!