15:01:34 <TravT> #startmeeting openstack search
15:01:34 <openstack> Meeting started Thu Sep 10 15:01:34 2015 UTC and is due to finish in 60 minutes.  The chair is TravT. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:01:35 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
15:01:37 <openstack> The meeting name has been set to 'openstack_search'
15:01:48 <rosmaita> o/
15:01:59 <TravT> o/
15:02:12 <lakshmiS> o/
15:02:14 <david-lyle> o/
15:02:14 <sjmc7> o/
15:02:23 <TravT> #link https://etherpad.openstack.org/p/search-team-meeting-agenda
15:02:25 <sjmc7> that's starting to look creepy
15:02:45 <TravT> the agenda or the o/
15:02:50 <sjmc7> :)
15:02:57 <TravT> plz add anything to it.
15:03:08 <TravT> #topic general status updates
15:03:28 <TravT> thanks for some productivity yesterday at the review hour.
15:03:34 <TravT> i know not everybody could  make it
15:03:59 <TravT> but we actually did get 5 patches in yesterday
15:04:36 <TravT> i don't have any other general updates. so...
15:04:44 <TravT> #topic Release
15:05:02 <TravT> been looking at this:
15:05:02 <TravT> https://wiki.openstack.org/wiki/ReleaseTeam/How_To_Release#Swift_intermediary_releases
15:05:46 <TravT> I think that's similar to what we want to do, but I saw in the release tools repo there are some new scripts
15:06:19 <sjmc7> probably nikhil and david have the most experience with the release tooling
15:06:39 <TravT> I was planning to reach out to rel management team next week
15:06:54 <TravT> to see if we need to run it ourselves or not
15:07:05 <nikhil_k_> +1 to that
15:07:10 <TravT> david-lyle: nikhil_k_: they pretty much do it for horizon and glance, right?
15:07:17 <nikhil_k_> if we import os libraries we should run by them
15:07:23 <nikhil_k_> yes
15:07:25 <TravT> being part of the integrated release.
15:07:52 <david-lyle> TravT: yes
15:08:08 <nikhil_k_> sjmc7: the most I've used it the gerrit review system with SHA to tag stuff
15:08:47 <nikhil_k_> TravT: we would need to define requirements similar to global caps. not sure if we have a patch that does it now
15:09:05 <TravT> one thing we'll have to figure out, which maybe nikhil_k_ and david-lyle would know is if we need to reset the milestone target on all our bugs and BPs
15:09:06 <TravT> https://blueprints.launchpad.net/searchlight/liberty
15:09:20 <david-lyle> automated if through relmgr
15:09:27 <david-lyle> they have scripts
15:09:34 <TravT> right now, i've been keeping them up with when they would have released had we done milestone releases
15:09:58 <david-lyle> I think even if you do the release, you would use the same scripts and tooling
15:10:37 <TravT> ok.  well, i'll reach out to the release management team next week to start prep work.
15:11:03 <TravT> also, as we discussed awhile back, i'll set up a seccore team
15:11:23 <TravT> it will be a subset of the core team.
15:11:34 <nikhil_k_> I think so too. or thierry can run the script (with admin rights or whatever) to fix the bug statuses
15:12:00 <TravT> ok, good to know
15:12:34 <TravT> so, if you get an invite to join the seccore team, please accept it if you can.
15:12:47 <TravT> if you specifically have an interest of being on it, let me know.
15:14:01 <TravT> so, lets talk about current BP statuses a bit.
15:14:08 <TravT> #link https://blueprints.launchpad.net/searchlight/liberty
15:14:36 <TravT> lakshmiS how are functional tests coming?  still looking to be good for liberty release?
15:14:59 <lakshmiS> yes i removed WIP status
15:15:02 <lakshmiS> ready for reviews
15:15:21 <TravT> ok.  i filed a bug yesterday that I didn't put up a patch on.
15:15:27 <TravT> but i think it'll affect functional tests
15:15:27 <lakshmiS> we can always add more but i guess there's enough of them now
15:15:51 <TravT> are you testing for "Fields" now?
15:15:52 <lakshmiS> api, load and listener core functionality is tested
15:16:11 <sjmc7> lakshmiS, i'll revierw that again today
15:16:12 <lakshmiS> Fields in ES response?
15:16:16 <TravT> as input.
15:16:26 <TravT> https://bugs.launchpad.net/searchlight/+bug/1493586
15:16:26 <openstack> Launchpad bug 1493586 in OpenStack Search (Searchlight) "Missing _source parameter" [Undecided,New]
15:16:40 <TravT> i have no idea why we are remapping fields to _source
15:16:47 <TravT> but we shouldn't IMO.
15:16:51 <sjmc7> remapping?
15:17:07 <TravT> you can send in fields parameter to searchlight
15:17:09 <sjmc7> if you specify 'fields' e-s won't return the _source doc
15:17:11 <TravT> but it is actually _source
15:17:21 <TravT> but there is a fields parameter for ES
15:17:23 <TravT> that is different
15:17:29 <sjmc7> err... um.
15:17:34 <TravT> confused the heck out of me yesterday
15:17:50 <sjmc7> me too :) i'll have a look if i get time later
15:17:53 <TravT> or a day ago...
15:17:58 <lakshmiS> we do get _source
15:18:25 <TravT> well, possibly i was doing something wrong.  it was nearly midnight
15:18:43 <TravT> but i'm pretty sure we pop fields from the request and send it in as _source
15:19:41 <TravT> anyway, lakshmiS, that's great about the functional tests
15:19:53 <TravT> I'll try to give them a go.
15:19:59 <lakshmiS> ok
15:20:10 <TravT> next bp is designate.
15:20:21 <TravT> ekarlso has been really working through that with sjmc7 and i
15:20:36 <TravT> seems that it is very close
15:21:02 <TravT> i asked to put in parent child relationship for some of the data mapping, which has caused a little churn on it
15:21:41 <TravT> but i think designate will be in for release.
15:22:02 <TravT> bp glance-metadef-plugin
15:22:32 <TravT> lakshmiS found one bug in that
15:22:38 <TravT> and has a fix up
15:22:55 <TravT> i also have been experiencing some pain with metadefs on some horizon work i'm doing
15:23:03 <sjmc7> *hangs head*
15:23:08 <TravT> but haven't debugged that yet.
15:23:43 <TravT> query-perf-tests
15:23:49 <TravT> nikhil_k_: that's yours ^
15:24:28 <nikhil_k_> TravT: I doubt if I can complete them in next 2 weeks. THings have been sitting for a while as Glance was heavy past few weeks
15:24:44 <TravT> well, i don't think it is a release blocker.
15:24:55 <nikhil_k_> But if we really need them, I can try to define some scenarios that we can try to get in for L
15:25:23 <nikhil_k_> sure, we can chat offline when you have a better picture of all other features
15:25:29 <sjmc7> i think it's fine after L. it's to feed into fixing things
15:25:43 <TravT> sounds good... i think perf testing will be a big topic in a lot of areas...
15:25:50 <nikhil_k_> yeah, possibly cross project interaction and some types queries
15:25:53 <sjmc7> yeah. it would be good to do before the summit tho
15:25:56 <TravT> i'm actually a bit more worried about indexing perf than query perf
15:25:59 <sjmc7> yep
15:26:01 <sjmc7> me too
15:26:08 <nikhil_k_> I see
15:27:08 <TravT> I agree that perf scenarios and testing should be really tackled in M.
15:27:41 <TravT> so nikhil_k_, i'll defer that BP to next
15:27:48 <nikhil_k_> thanks
15:28:19 <TravT> next up, of interest to you david-lyle
15:28:28 <TravT> horizon integration.
15:28:45 <TravT> we won't have an in horizon trunk integration for liberty, but...
15:29:06 <TravT> i have a search panel for horizon that's coming along okay.
15:29:38 <TravT> it won't be perfect, but i've been thinking about what to do with it.
15:29:49 <TravT> i was thinking about a few options
15:30:08 <TravT> 1) holding off and proposing to mitaka horizon trunk
15:30:19 <TravT> 2) putting it into contrib on searchlight repo
15:30:24 <TravT> 3) creating a new repo
15:31:18 <TravT> but we need something to actually show searchlight capabilities in liberty
15:31:29 <sjmc7> we've also been fiddling with a single search box type thing
15:31:46 <TravT> yes, i gave a thumbs up to the invision mocks on that
15:31:55 <TravT> went back and forth a bit
15:32:15 <TravT> but the search panel actually brings some qualities that the top nav search doesn't have.
15:32:19 <TravT> and can be plugged in.
15:32:43 <TravT> i have a fear that it'll be the end of mitaka before we can get the top nav figured out in horizon proper.
15:33:13 <TravT> david-lyle must be asleep
15:33:29 <david-lyle> doh, too many streams
15:33:34 <TravT> ^
15:34:27 <TravT> basically i'm asking what I should do with a proof of concept horizon search panel plugin for liberty searchlight
15:35:13 <david-lyle> easiest if probably a different github repo for now or a feature branch, that way you could continue to iterate up to the summit
15:35:26 <david-lyle> and then merge into master when baked
15:35:49 <david-lyle> otherwise the patch approval process will slow you down, I fear
15:36:21 <TravT> yeah, last week I came to the realization that we had to do something like this
15:37:06 <david-lyle> I think it's valid to demo a POC that is targeted to merging
15:37:38 <TravT> well, perhaps I just should keep it as one massive patch and propose to horizon.
15:37:49 <david-lyle> could do that as well
15:37:52 <TravT> it subsumes the searchlight api patch steve already had up.
15:38:23 <TravT> ok, that is pretty easy...
15:38:52 <TravT> i'll start with that.
15:39:00 <david-lyle> never enough runway ;)
15:39:26 <TravT> will be nice to get the code (as ugly as it is at the moment up into a system off mine).
15:39:55 <TravT> okay.  next BP: plugin-config-options
15:40:07 <TravT> ekarlso and sjmc7 were debating starting that yesterday
15:40:17 <TravT> i think ekarlso said he'd put up a patch
15:40:24 <sjmc7> i'm pretty keen to sort out the config stuff before L
15:40:32 <TravT> probably based off what they have in designate
15:40:35 <TravT> me too sjmc7
15:40:36 <sjmc7> it's going to be a problem for actual deployments
15:40:42 <sjmc7> which will no doubt fall into my lap :)
15:40:47 <TravT> lol
15:41:08 <TravT> i look forward to much cursing from you
15:41:20 <sjmc7> an advantage of being english
15:41:27 <TravT> ok, so let's keep that BP for liberty
15:41:42 <TravT> bp enable-cors
15:41:58 <TravT> i'm still very much interested in this, but it still isn't a must
15:42:09 <TravT> at least for horizon use cases
15:42:16 <TravT> since we proxy through horizon atm.
15:42:36 <david-lyle> seems like a nice Mitaka item
15:42:54 <TravT> yeah, most likely.
15:43:17 <TravT> #topic bugs
15:43:20 <TravT> #link https://bugs.launchpad.net/searchlight
15:43:36 <TravT> so far we've done a good job staying on top of them.
15:43:55 <TravT> but there are a few opportunities out there.
15:44:30 <TravT> does anybody have any they'd like to discuss?
15:44:38 <lakshmiS> so what are we deciding on https://bugs.launchpad.net/searchlight/+bug/1490697?  modified fields for glance client 1.0.0
15:44:40 <openstack> Launchpad bug 1490697 in OpenStack Search (Searchlight) "python glance-client for image-members has upgraded" [High,In progress] - Assigned to Lakshmi N Sampath (lakshmi-sampath)
15:44:53 <lakshmiS> functional tests depend on this bug
15:45:07 <TravT> nikhil_k_: sigmavirus24: rosmaita: ^
15:45:19 <sjmc7> ah, yeah. so what was the deal - use glance-client 1.0.0 or no?
15:45:24 <sjmc7> what's global-requirements doing?
15:45:48 <sjmc7> python-glanceclient>=0.18.0 is in global-reqs on master
15:45:57 <sjmc7> no upper cap
15:46:20 <TravT> https://review.openstack.org/#/c/219168/
15:46:26 <lakshmiS> that leaves it open :)
15:46:48 <TravT> nikhil_k_: if you look at comments on bottom of that patch
15:46:52 <sjmc7> does horizon show image members?
15:47:08 <TravT> we just aren't sure if we should put a hard requirement on glance client 1.0 on searchlight
15:47:37 <david-lyle> I don't believe so
15:47:39 <TravT> or if we need some conditional code
15:48:19 <lakshmiS> not sure how other projects deal with backward incompatible code on python clients
15:48:25 <sjmc7> if there's no global-reqs cap, everyone will get 1.0.0
15:48:29 <TravT> horizon does not explicitly show image members or even allow updating them.
15:48:46 <TravT> but it does do filtering based on image members
15:48:55 <sjmc7> we could put a hack in though; look for both attributes
15:48:56 <TravT> but i believe it is implicit
15:48:59 <nikhil_k_> lakshmiS: TravT currently, jokke_ and I are having that conversation on how to deal with the situation
15:49:13 <TravT> if not owned by me and is not public, then must be shared with me.
15:49:34 <nikhil_k_> http://lists.openstack.org/pipermail/openstack-operators/2015-September/thread.html#8079
15:50:04 <sjmc7> worst case, we do the conversion - look for member_id and turn it into member. or vice versa
15:50:34 <lakshmiS> hack would be easy if we dont have a clear solution
15:51:02 <sjmc7> it's not really a hack per se
15:51:11 <TravT> nikhil_k, can you give a two sentence summary
15:51:16 <sjmc7> i'm ok with doing it
15:51:45 <TravT> lakshmiS: i more okay with that than only working with glanceclient 1.0
15:52:13 <nikhil_k_> TravT: we are currently deciding if the CLI should support back-compact changes for v1 API
15:52:14 <lakshmiS> i will meanwhile update the  patch with the hack
15:52:15 <TravT> i also have terrible grammar in irc
15:52:43 <lakshmiS> nikhil_k: if you are going to take a decision please update the patch. I will wait until monday
15:52:44 <sjmc7> let's put the conditional in
15:52:51 <sjmc7> or that :)
15:53:01 <nikhil_k_> TravT: sorry, that's just one line and that's it
15:53:05 <nikhil_k_> lakshmiS: sure
15:53:23 <TravT> tbh, i'd just put in the conditional
15:53:28 <sjmc7> yep
15:54:18 <lakshmiS> TravT: sure
15:54:28 <TravT> ok, cool
15:54:46 <TravT> lakshmiS: i saw you also updated the member notification patch based on feedback from nikhil_k_
15:55:09 <TravT> is it ready for him and other glance guys to review?
15:55:10 <lakshmiS> yes I did fix some of the test cases.
15:55:36 <lakshmiS> there are still 2 tests failing on glance store which i am looking currently (need to understand that part of code which is new to me)
15:55:49 <nikhil_k_> thanks, if you have a link it would help a ton! (got around 100 FF tabs open atm)
15:56:19 <lakshmiS> https://review.openstack.org/#/c/221307/
15:56:26 <nikhil_k_> thanks!!
15:57:29 <TravT> ok, just a couple minutes left
15:57:53 <TravT> if i send out another "review hour" request for next week tuesday or wednesday, would that be helpful?
15:58:27 <sjmc7> sure
15:58:34 <TravT> preference in days?
15:59:15 <lakshmiS> any day is fine with me
15:59:45 <TravT> ok, my coffee IV line needs to be changed and we are out of time anyway.
15:59:52 <TravT> thanks everybody!
15:59:57 <sjmc7> aye. thanks
15:59:59 <TravT> #endmeeting