15:00:39 <sjmc7> #startmeeting openstack search
15:00:39 <openstack> Meeting started Thu Dec  8 15:00:39 2016 UTC and is due to finish in 60 minutes.  The chair is sjmc7. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:00:40 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
15:00:44 <openstack> The meeting name has been set to 'openstack_search'
15:00:48 <TravT> o/
15:00:59 <sjmc7> https://etherpad.openstack.org/p/search-team-meeting-agenda
15:01:00 <sjmc7> morning
15:01:05 <Kevin_Zheng> o/
15:01:24 <sjmc7> might be a really quick one today
15:01:32 <lei-zh> o/
15:02:40 <sjmc7> ok, let’s get started; some others may dribble in
15:02:49 <sjmc7> #topic Elasticsearch-5
15:03:20 <sjmc7> I fired up a fresh devstack with es-5 and everything seemed to work fine, with the exception of our devstack plugin
15:03:43 <sjmc7> elastic.co changed the location of downloads. i filed https://bugs.launchpad.net/searchlight/+bug/1648255 to fix that
15:03:43 <openstack> Launchpad bug 1648255 in OpenStack Search (Searchlight) "devstack plugin doesn't support elasticsearch 5.x" [Undecided,New] - Assigned to Ruslan Gustomyasov (rusik)
15:03:43 <yingjun> o/
15:03:52 <sjmc7> But other than that, think everything’s fine with it
15:04:28 <sjmc7> So I may try to put a patch up to use the IP type for ipv6 fields if elasticsearch 5 is being used
15:05:06 <TravT> well, that's super nice that changes don't seem to be required.
15:05:19 <TravT> just to see it work...
15:05:22 <sjmc7> Yeah. I think they made all their breaking changes with v2
15:05:57 <sjmc7> That also means we could set ES 5 to be the CI default for ocata
15:06:27 <TravT> hmm... i wonder if we should or not.
15:07:30 <TravT> i mean should it be whatever is most widely deployed? or most recent?
15:08:02 <sjmc7> dunno. i’d go with what people would install if they were deploying it from fresh, but if there aren’t any compatibility issues it doesn’t matter either way i suppose
15:08:53 <TravT> given all the performance increases in 5, it seems like it would be what we'd recommend.
15:09:00 <sjmc7> yeah
15:09:02 <caowei_> hi
15:09:06 <sjmc7> hello!
15:10:07 <sjmc7> well, i’ll see about making that change and then toss a coin about updating the default / CI version
15:10:13 <TravT> hi caowei_
15:10:45 <sjmc7> #topic python 3.5 test jobs
15:10:48 <TravT> that sounds reasonable...
15:10:54 <Kevin_Zheng> i guess there are not too many deployments right now
15:11:09 <sjmc7> no, probably not
15:11:35 <sjmc7> sorry, i jumped the gun a bit there, over-caffeinated
15:11:43 <Kevin_Zheng> so maybe newer version will be better
15:12:00 <TravT> sort of on that note i see the fuel ccp is adding searchlight now: https://review.openstack.org/#/c/408651/
15:12:28 <Kevin_Zheng> cool
15:12:53 <sjmc7> i’ve still not looked much at the os-ansible patch, i need to do that
15:14:23 <sjmc7> ok.. any more on that?
15:14:57 <TravT> not from me.
15:15:05 <TravT> I'm just +1 on the move to v5
15:15:07 <Kevin_Zheng> Me neither
15:15:13 <TravT> but do we need a v2 test job?
15:15:21 <sjmc7> we can’t have both
15:16:09 <sjmc7> unless we make the py27 and py34 tests run with different versions
15:16:16 <sjmc7> which is an option, although potentially confusing
15:16:42 <robcresswell> fwiw, you probably want to make the call ASAP since O-2 is next week. I'd also vote for increasing it for newer deployers trying it for the first time :)
15:16:55 <sjmc7> the more votes the merrier
15:17:07 <robcresswell> You could do py27es2 etc, like Horizon does with Django.
15:17:18 <robcresswell> (I think)
15:17:25 <sjmc7> yeah… infra just spent 2 weeks unravelling our custom jobs
15:17:32 <robcresswell> Ah
15:17:37 <sjmc7> you’re next!
15:17:46 <robcresswell> So you are not in their good books :)
15:18:01 <sjmc7> well, i am now. i won’t be if i add one back :)
15:18:24 <TravT> i'd a slightly +1 on py27 using es2 and py34 / py 35 being es 5
15:18:46 <TravT> just to ensure we don't do something that breaks es2
15:18:59 <sjmc7> not so keen on doing that because it overloads the test, but i guess it’d be ok if people understand that we’ve done it
15:19:12 <sjmc7> talking of which
15:19:18 <sjmc7> https://review.openstack.org/390699 got proposed for searchlight-client to enable py35 tests last week
15:19:24 <lei-zh> +1 for both es2 and es5 tests
15:19:36 <sjmc7> we should do the same for python 3.5, so maybe could use py35 for es 5 and py34 for 2
15:19:37 <robcresswell> Well, you'd add a second test. But a single test run for py27 seems sensible
15:20:25 <sjmc7> we might be able to do it properly with a second test; i think the infra objection was that the jobs were custom for searchlight
15:20:37 <sjmc7> if it’s parameterized it’d be ok. i’ll look into it
15:21:07 <sjmc7> since rob’s here, this brings me onto the next topic
15:21:13 <sjmc7> #topic searchlight-ui reviews
15:21:25 <sjmc7> there are a number of SL UI reviews from a while back
15:21:40 <sjmc7> TravT: are they all still valid? i was planning to try to crank through them today
15:21:51 <sjmc7> and is there anything needs to go into horizon itself?
15:23:42 <sjmc7> https://review.openstack.org/#/q/status:open+AND+(openstack/searchlight-ui)
15:23:47 <TravT> let me see...
15:24:42 <TravT> i think tyr's sl-magic search patch can be abandoned
15:25:10 <TravT> the themeable one was dependent on getting some theme work landed in horizon, but i think hurgleburgler didn't get that landed in horizon
15:25:39 <TravT> i think i could update the instance usage aggregation patch
15:27:07 <sjmc7> yeah, i guess any where the dependency didn’t merge are unlikely to be able to get merged
15:27:29 <TravT> i can try running that one again and seeing if i'm wrong
15:27:47 <TravT> robcresswell: did you start moving any instance stuff to horizon master?
15:28:08 <robcresswell> TravT: No not yet, way too much on my plate :)
15:28:14 <TravT> ok.
15:28:22 <robcresswell> TravT: Waiting for some of my current patches to land before I start doing more, so just reviewing right now
15:29:40 <robcresswell> Should be able to work on that in a week or so. Maybe over Christmas when I've got sick of family :)
15:29:50 <sjmc7> :)
15:30:05 <sjmc7> 11am christmas day then?
15:30:32 * TravT pictures rob walking in on everybody eating breakfast and announcing he's got better things to do
15:31:09 <robcresswell> TravT: That's about right
15:32:04 <sjmc7> ok, last up
15:32:09 <sjmc7> #topic ocata-2
15:32:16 <sjmc7> As Rob mentioned, O-2 is next week (doesn’t time fly)
15:32:30 <sjmc7> after that we won’t be adding any significant features
15:32:49 <Kevin_Zheng> https://review.openstack.org/#/c/393617
15:33:23 <sjmc7> yeah, i’ll take another look at that today
15:33:24 <Kevin_Zheng> i got this one up, maybe have a look if you guys have time
15:33:30 <Kevin_Zheng> Thanks
15:33:35 <sjmc7> will do. it looked mostly all the way there
15:33:51 <sjmc7> i won’t do a total freeze on all patches after o-2, but just no big changes that could break stuff
15:34:16 <Kevin_Zheng> Sure
15:34:35 <TravT> so that of course brings up the pipeline
15:34:43 <TravT> looks like lei-zhupdated it
15:34:50 <TravT> lei-zh: ^
15:35:04 <lei-zh> yeah, that's one of the big changes : )
15:35:21 <sjmc7> yeah, i’m gonna take another look at that as well
15:35:27 <robcresswell> \o/
15:35:29 <TravT> Kevin_Zheng: did nova side land for notifications for server groups?
15:35:39 <robcresswell> It would be really great if that could land
15:35:56 <lei-zh> I've updated it today and will add more tests to cover
15:36:04 <Kevin_Zheng> They
15:36:07 <Kevin_Zheng> Have
15:36:12 <TravT> ok, cool
15:36:18 <Kevin_Zheng> But in old style
15:36:52 <sjmc7> you mean unversioned?
15:37:01 <Kevin_Zheng> But I already make it not always call nova API so not that much traffic
15:37:29 <Kevin_Zheng> Yes and the data load is quite limited
15:37:52 <TravT> i'd think server group changes are not very frequent
15:38:45 <Kevin_Zheng> Yes, but I have already done it as security group rules when update and it works well
15:39:23 <sjmc7> ok. i’ll be able to test that today as well
15:39:43 <Kevin_Zheng> :)
15:40:27 <sjmc7> that was all i had on the agenda. if anyone’s got anything else, go for it
15:40:30 <sjmc7> #topic open discussion
15:41:01 <robcresswell> Nothing from me, thanks for the updates :)
15:41:13 <lei-zh> I'll be on vacation next week, but will try to keep an eye on pipeline patch
15:41:34 <sjmc7> yeah, i’m likely to be out for 2-3 weeks as well
15:41:52 <sjmc7> so i will pass the reins to travis
15:42:24 * robcresswell pictures sjmc7 like Santa passing over sleigh reins
15:43:04 <TravT> i think we should plan on cancelling a few meetings...  Dec 22, Dec 29
15:43:29 <sjmc7> yeah, probably
15:43:56 <TravT> although those weeks it would be fine for yingjun, lei-zh, Kevin_Zheng, others to meet if one of you guys want to run the meetings.
15:43:59 <sjmc7> although a lot of folks don’t disappear for a week over christmas
15:44:55 <lei-zh> yeah, we can get two +2s to make things work:)
15:45:34 <TravT> sounds like lei-zh is volunteering to run those weeks.
15:45:38 <TravT> :)
15:45:44 <sjmc7> :)
15:45:59 <sjmc7> ok, if there’s nothing else, thanks everyone
15:46:25 <TravT> thank you!
15:47:31 <lei-zh> thanks, bye:)
15:47:54 <robcresswell> thanks sjmc7
15:48:25 <sjmc7> #endmeeting