15:00:48 <TravT> #startmeeting openstack search
15:00:52 <openstack> Meeting started Thu Aug 20 15:00:48 2015 UTC and is due to finish in 60 minutes.  The chair is TravT. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:00:54 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
15:00:57 <openstack> The meeting name has been set to 'openstack_search'
15:01:04 <rosmaita> o/
15:01:24 <nikhil_k> o/
15:01:34 <TravT> Let's give a couple mins.
15:01:35 <lakshmiS> o/
15:01:37 <david-lyle> o/
15:01:41 <sigmavirus24> o/
15:02:00 <TravT> I don't think sjmc7 will be here at least at the start.  There is an HP meeting in seattle that he had to go to.
15:02:27 <sjmc7> morning
15:02:27 <TravT> and there he is
15:02:31 <TravT> :)
15:02:35 <sjmc7> sorry, bit under the weather
15:02:46 <TravT> https://etherpad.openstack.org/p/search-team-meeting-agenda
15:02:56 <TravT> Add anything you want as usual
15:03:15 <TravT> #topic Testing
15:03:33 <TravT> As mentioned last week, we put up a review for python 3 in the gate.
15:03:37 <TravT> It went through.
15:03:50 <TravT> that's all.
15:03:57 <sjmc7> :)
15:04:01 <TravT> Functional tests.
15:04:19 <TravT> lakshmiS how are they coming?
15:04:27 <TravT> you were going to put up an initial patch?
15:04:48 <lakshmiS> ready to go. will put up the patch before the weekend.
15:05:03 <TravT> Great!
15:05:19 <TravT> any discussion needed on them with the group here?
15:05:36 <lakshmiS> nothing specific i can think about
15:05:50 <TravT> ok, well, i'm looking forward to seeing them.
15:05:57 <lakshmiS> even though i tried to cover the RBAC filters that rosmaita initially gave
15:06:30 <sjmc7> there were some in that patch i had in glance a while back
15:06:35 <sjmc7> if that's useful
15:06:45 <lakshmiS> sjmc7: yeah i ported them
15:07:09 <sjmc7> great, thanks
15:07:43 <TravT> #topic reviews process
15:08:08 <TravT> so, we're only a couple weeks out in liberty
15:08:26 <TravT> i'm still hopeful we can have a release.
15:08:54 <TravT> but i also know that having time to do reviews is difficult for a number of people right now.
15:09:34 <TravT> i've been trying to catch up on them.
15:09:47 <TravT> and they are time consuming to fully test.
15:10:14 <TravT> but wanted to open up some discussion on ideas for how to facilitate getting reviews down in timely manner.
15:10:49 <TravT> and also what is / isn't okay process-wise
15:11:08 <TravT> In the early phases of this project.
15:11:26 <sjmc7> i know one thing we've tried to stick to is to have +2s from multiple companies, even though there are a majority of HP folks on the core list
15:12:14 <sjmc7> if people are ok with approving changes with 2 +2s even if it's just HP folks just as the summit deadline approaches that might help, especially for bug fixes
15:12:43 <david-lyle> I think for now that may be a good relaxation of the review guidelines
15:12:50 <rosmaita> i agree
15:13:04 <rosmaita> i keep hosing my devstack, which is slowing my reviews, so i am part of the problem here
15:13:14 <david-lyle> we're still working on critical mass
15:13:17 <nikhil_k> same with me
15:13:39 <nikhil_k> I agree too, but would be good if someone can test them functionally. I know TravT is very particular about it and best to keep that practice
15:13:45 <sjmc7> yeah, i've been testing them
15:13:54 <sjmc7> it is time consuming btu esp given our tests it's necessary
15:14:06 <TravT> yeah, i seem to be spending a large portion of my days just playing QA recently.
15:14:36 <nikhil_k> :)
15:15:05 <lakshmiS> we need downstream applications using searchlight to do that for us:)
15:15:13 <TravT> lol
15:15:30 <sjmc7> we've got people itching to use it :)
15:15:46 <TravT> I'd still like to ensure that everybody has an opportunity to weigh in at least conceptually.
15:15:58 <lakshmiS> a +1 would be good to have
15:16:13 <TravT> yes, just to indicate the idea seems right.
15:16:18 <sjmc7> yeah - for features or bp implementations i'd definitely like additional input
15:17:16 <sjmc7> but for pure bugfixes it seems less important
15:17:34 <TravT> yes, and we have several high and critical's sitting out there right now with code proposed.
15:17:45 <TravT> I also will consider the functional tests to be critical.
15:18:21 <TravT> because i don't want to see any regressions.
15:19:35 <TravT> any other input on this specifically?
15:20:48 <TravT> lakshmiS, last week you had mentioned the idea of having an hour or so available each week for people to interact on specific reviews through end of liberty
15:21:05 <nikhil_k> that would be nice!
15:21:14 <lakshmiS> yes lets schedule it
15:21:55 <TravT> how about Tuesday or Wednesdays at this same time?
15:22:09 <TravT> in the searchlight room
15:22:35 <nikhil_k> i don'tmind sparing a couple hours tomorrow just on reviews if someone is available
15:23:16 <TravT> nikhil_k, there is a question on the nova patch I have for you when we get to it later in this meeting.
15:23:18 <rosmaita> tuesday would be better for me
15:23:27 <nikhil_k> wed looks better for me, otherwise
15:23:37 <lakshmiS> both are good with me
15:24:03 <nikhil_k> i am travelling for work mon, tue most likely
15:24:35 <TravT> i can do either.
15:24:44 <TravT> is wednesday out for you rosmaita
15:24:46 <TravT> ?
15:25:39 <sjmc7> either's ok with me, wednesday maybe a bit easier
15:26:19 <david-lyle> I' fine with either
15:26:49 <TravT> david-lyle: i'm still trying to figure out if i can psych myself up for your added early morning horizon meeting on wednesday. :(
15:26:57 <sjmc7> :)
15:27:20 <nikhil_k> actaully, I may be online at 1500utc on tuesday but not 100% sure
15:27:43 <TravT> If possible, let's shoot for Tuesday
15:27:58 <TravT> I'll check in on Monday with everybody
15:28:37 <david-lyle> just not 2000 or 2100 UTC on Tues those are taken with TC and X-Project
15:29:06 <TravT> that is way too late for lakshmiS anyway
15:29:08 <TravT> http://everytimezone.com/#2015-8-20,180,cn3
15:29:18 <david-lyle> ok, just making sure
15:29:19 <TravT> that's the time slot shooting for on Tuesday.
15:29:34 <TravT> if you can't make it nikhil_k, we'll bump to wednesday
15:29:52 <TravT> ok, on to but reviews
15:29:56 <nikhil_k> thanks TravT
15:30:01 <TravT> #topic bugs
15:30:04 <lakshmiS> it says 8:30 PM local time for me which is same as now
15:30:29 <TravT> so, this one:
15:30:30 <TravT> Fix for Authentication not happening https://review.openstack.org/211047
15:30:43 <rosmaita> TravT: i cann do wed, but have team mtg at 15:30
15:31:11 <TravT> I kind of like tuesday better than wednesday
15:31:23 <TravT> just so we have a day between that and this meeting
15:31:48 <TravT> so, this one: Fix for Authentication not happening https://review.openstack.org/211047
15:32:01 <TravT> even with allowing 2 +2s from same company, it needs another.
15:32:14 <TravT> it is a critical bug
15:32:30 <TravT> so, nikhil_k, david-lyle, sigmavirus24, rosmaita, if one of you could take a look
15:33:13 <nikhil_k> ack
15:33:29 <TravT> note that i co-authored, so I can't +2 it.  my +1 was only to indicate it was ready for reviews.
15:33:48 <david-lyle> looking, can try out the patch after this
15:34:01 <sjmc7> thatnks david
15:34:06 <rosmaita> i am looking at that, but hosed my devstack
15:34:09 <TravT> thanks david-lyle
15:34:25 <sjmc7> i did test it, generated configs and docs looked good to me
15:35:08 <TravT> there are a couple more critical highs.
15:35:17 <TravT> Fix RBAC filters https://review.openstack.org/212563
15:35:26 <TravT> I tested that one yesterday with images and the nova servers patch.
15:35:34 <sjmc7> yeah, this was a good one
15:35:37 <TravT> tried about 8 different queries on it.
15:35:45 <TravT> i put those in a paste in my review comments.
15:36:18 <TravT> laskhmiS, would be good to consider that with any of the RBAC functional tests you are doing
15:36:41 <TravT> so maybe you could take a look lakshmiS?
15:36:47 <sjmc7> we should provide guidance on writing these in the docs, or abstract some of it to the base class at some point. it's easy to screw up
15:36:49 <lakshmiS> ok i will check if its covered
15:37:16 <TravT> Your sorting patch, sjmc7     Sorting https://review.openstack.org/#/c/206268/
15:37:24 <TravT> I will look at this one
15:37:30 <sjmc7> yeah.. just needs reviews
15:37:39 <sjmc7> it should be pretty straightforward
15:37:59 <sjmc7> passthrough of the e-s DSL for sorting
15:38:19 <TravT> ok.
15:38:32 <TravT> On to BPs
15:38:36 <sjmc7> you actually could do it already by passing it inside the 'query' parameter
15:38:43 <sjmc7> but this makes it more explicit
15:39:36 <TravT> #topic nova plugin
15:39:49 <TravT> Nova instances plugin https://review.openstack.org/198852
15:39:57 <TravT> i've been harassing sjmc7 on this one quite a bit
15:40:00 <TravT> and it is very close
15:40:07 <sjmc7> i do feel quite harassed
15:40:29 <TravT> question for rosmaita and nikhil_k
15:40:47 <TravT> currently the following fields are filtered if you aren't admin
15:40:57 <TravT> OS-EXT-SRV-ATTR:host
15:41:08 <TravT> OS-EXT-SRV-ATTR:hypervisor_hostname
15:41:13 <TravT> are there others we should add?
15:41:27 <sjmc7> there's a hostId too
15:41:30 <sigmavirus24> Sorry, in the middle of a sad sad video call
15:41:32 <sjmc7> which it occurs should probably be protected
15:41:48 <TravT> sigmavirus24: need some kleenexes?
15:41:49 <sjmc7> :)  sigmavirus24. the best kind!
15:42:00 <sigmavirus24> no
15:42:06 <sigmavirus24> send the strongest of alcohol
15:42:12 <TravT> lol
15:43:40 <TravT> rosmaita: nikhil_k: if you could add a comment on that patch in this file about fields to filter for admin
15:43:41 <TravT> https://review.openstack.org/#/c/198852/14/searchlight/elasticsearch/plugins/nova/servers.py
15:43:55 <nikhil_k> TravT: don't quite have them ready but can check
15:44:17 <TravT> sure. line 30.
15:45:01 <TravT> other than a few other nits on that patch and some ideas for future things, that patch worked quite nicely for me
15:45:11 <TravT> so, maybe we can get nova support in by end of this week.
15:45:30 <nikhil_k> TravT: so, I am not sure about the extensions and their indexing
15:46:11 <nikhil_k> am aware that they were being proposed to be purged or moved to regular api but not sure if we can ever come up with a list of all the ext attrs that are admin only
15:46:44 <nikhil_k> they depend on which ext are enabled by the deployer and we will be playing cat and mouse
15:46:49 <nikhil_k> so, sorry some context:
15:46:54 <TravT> nikhil_k: in some of my comments on that patch, i suggested that we add a plugin config for whitelisting / blacklisting fields.
15:47:12 <nikhil_k> OS-EXT-* are OS Extensions added to nova api
15:47:26 <sjmc7> until we do that we could have a whitelist of extension fields in the code
15:47:39 <sjmc7> so we don't accidentally scrape up any we shouldn't
15:47:47 <nikhil_k> that sounds good
15:48:33 <TravT> here's a comment I put on the patch:
15:48:35 <TravT> It really also occurs to me that when we do that the nova plugin should support having separate extension files for all these extension fields. The extensions enabled would all be invoked during serialize function and also would be called to add to the mapping. This way deployers could add their extension without modifying source just like they do for nova. BP needed.
15:48:54 <sjmc7> yeah. maybe not a liberty timescale thing at this point
15:49:56 <nikhil_k> ah cool
15:50:13 <TravT> nikhil_k, you mentioned having review time tomorrow.
15:50:23 <nikhil_k> yes
15:50:31 <TravT> maybe you could do another pass on the nova patch.
15:50:42 <nikhil_k> sure
15:50:50 <nikhil_k> on my list now
15:50:51 <TravT> hopefully sjmc7 will be able to address my current comments before tomorrow and i'll do another pass
15:51:03 <TravT> ok.
15:51:12 <sjmc7> yeah, will do
15:51:31 <nikhil_k> darn, evernote app broke on my m/c
15:51:49 <TravT> ok, did want to mention that following our BP review last week I opened the client blueprint we discussed.
15:51:56 <TravT> https://blueprints.launchpad.net/searchlight/+spec/initial-openstack-client-search-plugin
15:51:58 <TravT> just FYI
15:52:15 <TravT> #topic open discussion
15:52:30 <sjmc7> i've actually got to run, meeting i have to attend in the office
15:52:54 <TravT> ok.
15:53:02 <sjmc7> anything anyone needs, i'll be in the SL channel in a biit
15:53:03 <TravT> i think everybody is quite busy these days.
15:53:08 <david-lyle> http://lists.openstack.org/pipermail/openstack-dev/2015-August/072082.html
15:53:21 <david-lyle> is a thread about keystone and pagination
15:53:41 <david-lyle> but the current stance by keystone on the v3 API makes it impossible to index
15:53:42 <sjmc7> i saw that one :)
15:53:44 <sjmc7> good reading!
15:54:02 <TravT> oh interesting.
15:54:06 <david-lyle> so, that's awesome
15:54:07 <TravT> i didn't see it
15:54:21 <david-lyle> this is not new, but progress is feeling unlikely
15:54:22 <TravT> that is awesome. :(
15:54:43 <lakshmiS> promising
15:55:36 <david-lyle> the other option is build based on other indexed data
15:56:36 <TravT> david-lyle?
15:56:43 <TravT> trying to figure out what you mean.
15:56:50 <david-lyle> me too :P
15:57:14 <TravT> ok, well, i'll read through that thread
15:57:15 <david-lyle> no, you could go do a secondary request based on results from nova, etc
15:57:48 <david-lyle> nova data will have project ids, you could request data on project ids returned with the data
15:58:06 <david-lyle> but that's an incomplete index
15:58:14 <david-lyle> so maybe worse :(
15:58:32 <TravT> we could just run a scheduled job and periodically get all projects from keystone.
15:58:54 <david-lyle> well no, you can't actually
15:59:00 <TravT> yay
15:59:12 <david-lyle> you'll get the first 500 or 1000 and an error saying there are more
15:59:18 <david-lyle> but no way to index into them
15:59:47 <lakshmiS> will have to go to their db then:)
15:59:49 <david-lyle> unless you dynamically build up a filtering scheme until you get below the MAX return # threshold
16:00:30 <TravT> okay, we are out of time, but this was good to bring up
16:00:33 <nikhil_k> Thanks!
16:00:35 <TravT> we can continue
16:00:41 <TravT> on it in the future.
16:00:44 <TravT> #endmeeting