15:00:58 <TravT> #startmeeting openstack search 15:01:00 <openstack> Meeting started Thu Oct 8 15:00:58 2015 UTC and is due to finish in 60 minutes. The chair is TravT. Information about MeetBot at http://wiki.debian.org/MeetBot. 15:01:02 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 15:01:04 <openstack> The meeting name has been set to 'openstack_search' 15:01:11 <nikhil> o/ 15:01:15 <TravT> o/ 15:01:16 <lakshmiS> o/ 15:01:25 <TravT> nikhil, so you'll be making it to this summit? 15:01:38 <nikhil> TravT: :) 15:02:05 <TravT> that's great 15:02:06 <rosmaita> o/ 15:02:19 <nikhil> Though I wonder if they should have kept Austin in Nov and Tokyo in May 15:02:36 <TravT> i have a text from sjmc7 that his internet connection died. 15:02:42 <TravT> he'll join when he can. 15:02:56 <lakshmiS> you mean weather wise 15:02:58 <TravT> nikhil: that's what I thought about Atlanta and Paris as well 15:03:09 <nikhil> true! :0 15:03:10 <nikhil> :) 15:03:18 <nikhil> lakshmiS: yes 15:03:41 <lakshmiS> 64F is not bad in tokyo 15:03:42 <nikhil> texas in may is going to be awesomeeeeee!!!!!!!!!! 15:03:54 <ativelkov> Japan in Nov should be great because of the red leaves festival :) 15:03:58 <TravT> better than texas in July I suspect 15:04:04 <TravT> hi ativelkov 15:04:04 <nikhil> higher 90s, 100s even 15:04:20 <ativelkov> hi TravT, sorry for interrupting :) 15:04:28 <TravT> no worries, just killing time... 15:04:34 <TravT> we're missing a few of our usual suspects 15:04:45 <lakshmiS> ativelkov: long time 15:05:02 <ativelkov> lakshmiS: :) 15:05:08 <TravT> ativelkov: we would like to touch base with you on artifacts v3 and indexing with searchlight 15:05:11 <lakshmiS> heard you have some meeting coming up next week on app catalog 15:05:42 <ativelkov> TravT: sure, we have the integration with the searchlite in our mitaka backlog :) 15:05:48 <TravT> :) 15:06:00 <david-lyle> o/ 15:06:10 <lakshmiS> ativelkov: we could probably see if appcatalog needs searching :) 15:06:10 <ativelkov> lakshmiS: yup, I was asked by the Openstack:Now editorial team to bring some Glance majors :) 15:06:52 <TravT> ativelkov: we have only a couple sessions at the summit for searchlight, but we could also have an informal one to talk about it and artifacts 15:07:02 <ativelkov> TravT: sure 15:07:33 <ativelkov> TravT: I have some Murano-related topics as well :) 15:07:52 <TravT> +1 15:08:09 <TravT> so, let me bring up the session planning etherpad in just a sec. 15:08:19 <TravT> #link https://etherpad.openstack.org/p/search-team-meeting-agenda 15:08:28 <TravT> #topic RC1 released 15:08:48 <TravT> Great job everybody! 15:09:08 <TravT> i put up release notes: https://wiki.openstack.org/wiki/ReleaseNotes/Liberty#OpenStack_Search_.28Searchlight.29 15:09:08 <lakshmiS> yeaaaaah! 15:09:20 <david-lyle> \o/ 15:09:46 <TravT> it is pretty awesome to accomplish what we have in the amount of time we've done it. 15:10:05 <TravT> really a great group of people 15:10:38 <TravT> if you see anything missing from the release notes, please update or let me know and I will 15:10:55 <TravT> we also have a stable/liberty branch 15:10:59 <nikhil> woohoo!!! 15:11:09 <nikhil> TravT: does that mean RC1 is rc-final for liberty? 15:11:15 <TravT> probably. 15:11:33 <TravT> there have been a couple bug fixes in the last few days and i found another, but really nothing critical 15:12:18 <nikhil> gotcha 15:12:29 <TravT> we did do one fix direct on stable/liberty 15:12:30 <TravT> https://review.openstack.org/#/c/231726/ 15:12:34 <nikhil> I was a bit confused on the timeline around rel notes 15:12:34 <sjmc7> hey 15:12:39 <nikhil> now clear 15:12:44 <TravT> hey sjmc7 15:13:10 <TravT> since we decided to lower the release number at the last minute to help set measured expectations 15:13:23 <TravT> allowing a little more wiggle room in mitaka 15:13:47 <TravT> so, i guess nikhil that means we will have a new tag on stable/liberty for final 15:14:01 <nikhil> TravT: I think that's critical 15:14:07 <TravT> yes 15:14:16 <nikhil> possibly early next week? 15:14:24 <TravT> we had one on master https://review.openstack.org/#/c/230086/ 15:14:36 <TravT> which i don't think we should backport 15:14:57 <nikhil> hmm 15:15:05 <nikhil> we don't keep a conf.sample .. 15:15:12 <nikhil> so, I am okay either way 15:15:33 <nikhil> oops, I read it in a wrong tab 15:15:36 <nikhil> nvm 15:15:41 * TravT was feeling confused 15:15:58 <TravT> there is one other bug that if we got figured out we might consider for backport. 15:16:01 <TravT> https://bugs.launchpad.net/searchlight/+bug/1503080 15:16:01 <openstack> Launchpad bug 1503080 in OpenStack Search (Searchlight) "Facet doc count reports more docs than actual" [Undecided,New] 15:17:08 <sjmc7> we've got to figure out a fix for it first :) 15:17:22 <TravT> yeah 15:17:48 <TravT> We do need a deadline for setting rc2 / final. 15:18:17 <TravT> would kind of like it to be tomorrow, but Mon / Tues at the latest 15:18:27 <TravT> I am leaving for japan Wednesday evening. 15:18:37 <TravT> going to vacation for a bit before the summit 15:18:41 <sjmc7> wow, that's a long trip! 15:18:54 <lakshmiS> going around japan other than tokyo? 15:19:14 <TravT> tokyo, hakone, nara, kyoto, and one other place that i can't remember 15:19:40 <lakshmiS> wow. i guess you will speak japanese by the end of it like you did with french 15:19:44 <TravT> lol 15:19:55 <sjmc7> well maybe let's say a deadline of finding any serious bugs by tommorrow evening? 15:19:58 <sjmc7> and doing the final on monday? 15:20:05 <TravT> that sounds reasonable. 15:20:09 <sjmc7> you'll also be sick of fish by then 15:20:16 <TravT> sjmc7: https://review.openstack.org/#/c/232200/1 15:20:24 <TravT> itisha put that up yesterday 15:20:29 <sjmc7> yeah, testing that now 15:20:56 <TravT> we could consider it as well... although not a serious bug. 15:21:52 <sjmc7> it's a nice to have for deployment 15:22:14 <TravT> seems pretty low risk 15:22:44 <lakshmiS> need to add some tests on it 15:23:02 <lakshmiS> if we can figure out one for config 15:23:32 <lakshmiS> manual testing for stuff increasing... 15:23:40 <TravT> yes, we need to decrease that 15:23:53 <sjmc7> maybe functional tests are best for config testing 15:24:10 <lakshmiS> yeah thats possible 15:24:28 <sjmc7> or mock loads of stuff 15:24:45 <TravT> lakshmiS: would you be able to leave a comment with pointer for itisha on how to add that kind of test? 15:24:51 <lakshmiS> sure 15:24:57 <TravT> thx 15:25:36 <TravT> ok, so we'll check in first thing Monday (last thing for lakshmiS) and then will likely push the tag 15:25:56 <TravT> if sjmc7 figures out a fix for the doc_count, let's take a look at it. 15:26:03 <sjmc7> no pressure? 15:26:14 <lakshmiS> :) 15:26:24 <TravT> well, i'm just debating whether or not to remove showing it in the horizon plugin i've got going. 15:26:30 <TravT> it is nice. 15:27:08 <TravT> so, we do have another urgent agenda item. 15:27:18 <TravT> #topic Mitaka session planning 15:27:31 <TravT> #link https://etherpad.openstack.org/p/searchlight-mitaka-summit 15:27:45 <TravT> We have a deadline of October 15th to enter our sessions into sched 15:27:59 <TravT> so, we need to get these figured out in the next day or two 15:28:27 <TravT> take a look at that etherpad: https://etherpad.openstack.org/p/searchlight-mitaka-summit 15:28:27 <sjmc7> ok. i have some ideas that i've not put down yet. i will do so today 15:28:48 <TravT> let's talk through some here. 15:29:31 <TravT> We have 1 fishbowl and 1 workroom slot. 15:29:49 <rosmaita> i was hoping to have some info for performance and scalability, but looks like that won't happen until after the summit 15:30:06 <nikhil> TravT: do we have priorities setup for searchlight or is there a plan? 15:30:21 <TravT> rosmaita: that is one of the ideas I put down for a workroom slot. 15:30:30 <rosmaita> yes, i saw that 15:30:32 <sjmc7> my biggest concern is getting away from hitting the service APIs for updates 15:30:35 <nikhil> I am tying to think from a 3rd person perspective if things look clear 15:31:24 <rosmaita> since just one slot, i think thinking about how to make cross-region searching happen would be a really good use of time 15:31:40 <TravT> nikhil: i have an initial list of ideas for that and started to go through blueprints. 15:31:41 <rosmaita> sjmc7: +1 to getting away from hitting the service APIs 15:31:50 <sjmc7> yeah, HP is also interested in cross-region 15:31:52 <TravT> was also thinking the fishbowl session might help us to set next priorities 15:32:24 <nikhil> TravT: gotcha 15:32:27 <sjmc7> the other one i had was trying to provide some standardization on how string indexing is done 15:32:50 <nikhil> From what I perceive, for a new interested vendor I see searchlight will be focusing on performance, cross region scoping (may be HA). What does it mean for indexing other services and for testing? 15:32:53 <sjmc7> because tokenized vs not tokenized is not very apparent until it goes wrong 15:33:02 <sjmc7> in terms of other services, swift would be a big one 15:33:11 <sjmc7> but i don't think there's any mechanism to do it 15:33:38 <lakshmiS> one takeway would be to hear from other teams on what would they like to be searched or pain points they have today which can also drive our priorities 15:33:55 <nikhil> what do you mean by the service APIs? API for openstack services or some sort of other API for each service that's being called as service API -- may be admin API or something? 15:34:16 <sjmc7> nikhil - anything we have to do over HTTP on updates 15:34:26 <sjmc7> it will severely limit update bandwidth 15:34:31 <nikhil> sjmc7: ++ 15:34:45 <sjmc7> designate's notifications contain all the info we get from the api 15:34:46 <nikhil> I was strongly pro for this approach since the beginning 15:34:47 <sjmc7> which is ideal 15:35:13 <sjmc7> unfortunately, many other services don't provide what we need. one option is to limit ourselves to the notifications and try to improve them, which is my preferred option 15:35:24 <sjmc7> anyway, i will add these to the etherpad 15:35:28 <nikhil> I guess that means we will be giving other projects a stick when notifications aren't good enough 15:35:33 <TravT> Yes, I think everybody is pro the approach, and it means getting changes into other services so that the notifications in other services are generated with the same thing you'd get with a specific version of the API 15:35:51 <nikhil> haha, yeah. typing at the same time sjmc7 :) 15:35:57 <TravT> for example, one of the blueprints i registered with us is to handle nova microversions. 15:36:29 <TravT> because right now, we aren't specifying a version, meaning we are getting the minimum nova version 15:37:23 <TravT> FYI: nova-server-versioning-extensions 15:37:26 <sjmc7> if we don't use the API, we'll get whatever the notification gives us :) 15:38:06 <TravT> which lowers the value of searchlight if the notification doesn't provide enough information 15:38:16 <sjmc7> yep. which is why it needs discussion 15:38:40 <TravT> maybe that should be a primary goal of the fishbowl session, if we can get people from other project into it 15:39:36 <sjmc7> ok. i'll add it there 15:39:57 <lakshmiS> we can never rely on just API or notification. i guess both will be needed based on the plugin 15:40:00 <TravT> that would play into the second idea i have there, i think 15:40:35 <TravT> nikhil: I copied my initial list of priorities onto this etherpad: https://etherpad.openstack.org/p/searchlight-mitaka-priorities-brainstorming 15:40:48 <TravT> anybody can add to it... this would be just ideas for priority themese 15:40:53 <TravT> and we could finalize... 15:41:09 <TravT> or we could do like nova and put a document into our repo with mitaka priorities 15:41:15 <TravT> and we can all gerritt review it 15:42:16 <TravT> ok, i'll just be quiet here for a bit and let people edit on the etherpad, comment, etc. 15:42:47 <TravT> https://etherpad.openstack.org/p/searchlight-mitaka-summit 15:44:47 <nikhil> TravT: ah thanks... I was just saying, not necessarily asking for a review :) BUt it's a bonus!!! 15:45:10 <nikhil> TravT: I have a topic for open discussion if we have time 15:45:29 <TravT> sure, let's give this another minute, i still see some editing happening on the etherpad. 15:45:43 <TravT> and I have to submit these up before our next meeting. 15:50:26 <TravT> sjmc7: how about we combine idea 2 & 3 for the fishbowl 15:50:55 <sjmc7> sure 15:51:09 <TravT> just leave there as is 15:52:01 <TravT> So, we also will be in a horizon fishbowl it looks like 15:52:09 <TravT> with zaqar 15:52:09 <sjmc7> though i think therewe need to decide how we (SL) approach it; do we go with our hybrid approach until notifications are sufficient alone for a service (which reduces update bandwidth) or work with what the notifications have (which reduces what we can index) 15:52:47 <TravT> maybe it could be configurable. 15:52:55 <TravT> graceful degradation 15:53:06 <TravT> that's a good discussion to have 15:53:11 <TravT> we won't solve in the next 7 minutes 15:53:35 <sjmc7> not with that attitude 15:53:42 <TravT> :D 15:53:58 <lakshmiS> when notifications are less we dont have to worry much on calling API 15:54:11 <TravT> We can continue this in the room, but I do want to give nikhil some time 15:54:17 <sjmc7> yep, go for it 15:54:20 <nikhil> THanks TravT 15:54:22 <TravT> #topic open discussion 15:54:42 <nikhil> SO, I am looking for a topic that an intern can possibly work on for 2 months full time 15:54:51 <nikhil> THey can potentially pair on a BP too 15:54:58 <lakshmiS> new plugin :) 15:55:02 <TravT> +1 15:55:11 <nikhil> Currently, I am getting them into Glance 15:55:35 <sjmc7> a new plugin'd be the obvious one though i'd hope that wouldn't be a 2 month effort 15:55:36 <nikhil> but as a trend would be okay to include them here and do people have ideas for their learnings? 15:56:12 <TravT> A plugin integration is actually a good way to learn about how openstack interactions work, i think. 15:56:19 <lakshmiS> tests are one area where a lot could be still added 15:56:22 <nikhil> I was asking for some initial feedback. The outreachy program is the that's most active within openstack, but we can also look into GSoC 15:56:34 <nikhil> I see 15:56:39 <david-lyle> sjmc7: remember a lot of learning time too 15:56:57 <TravT> is this itisha or somebody else? 15:56:59 <sjmc7> i'll hold off cracking the whip 15:57:01 <nikhil> so, I will take it as people think that in general interns are good addition and they can work on searchlight comfortably :) 15:57:18 <TravT> i think so. 15:57:21 <david-lyle> yes 15:57:24 <lakshmiS> +1 15:57:28 <nikhil> TravT: I am just scouting atm. Will ask the coordinators if there is more interest 15:57:36 <TravT> kragniz was like mr extraodirnaire intern 15:57:44 * TravT can't spell 15:57:45 <kragniz> lol 15:57:50 <nikhil> they allow more than one prj per application so, I can help them with Glance and SL 15:58:04 <lakshmiS> kragniz: hello there 15:58:12 <TravT> he's lurking! 15:58:25 <TravT> nikhil, i think it'd be great. 15:58:26 <nikhil> I can come back with more feedback in following weeks once there is some interest 15:58:28 <kragniz> I always lurk! 15:58:31 <nikhil> awesome! 15:58:42 <rosmaita> kragniz: hello 15:58:48 <kragniz> unless my lecture ends, like it just has 15:58:56 * kragniz runs away 15:59:02 <TravT> bye kragniz 15:59:07 <nikhil> thanks TravT. I am hoping to add a topic related to SL to the OPW wiki and will start with the tests stuff as lakshmiS mentioned. 15:59:45 <nikhil> Data + Search if one catchy topic so I expect good interest but time will tell. 15:59:49 <TravT> ok, so let's circle back on final release tag and session planning in the room tomorrow / Monday. 15:59:51 <scarlisle> o/ 16:00:02 <prometheanfire> hi 16:00:14 <TravT> search is actually quite interesting. 16:00:25 <nikhil> :) 16:00:28 <TravT> i guess we better end this meeting so we stop confusing others. 16:00:32 <TravT> #endmeeting