15:00:02 <TravT> #startmeeting openstack search
15:00:03 <openstack> Meeting started Thu Aug 25 15:00:02 2016 UTC and is due to finish in 60 minutes.  The chair is TravT. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:00:04 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
15:00:07 <openstack> The meeting name has been set to 'openstack_search'
15:00:41 <TravT> o/
15:00:48 <RickA-HP> o/
15:00:51 <lei-zh> o/
15:00:52 <sjmc7> o/
15:00:55 <yingjun_> o/
15:00:57 <Kevin_Zheng> o/
15:01:32 <TravT> okay
15:01:36 <rosmaita> o/
15:01:41 <matt-borland> o/
15:01:43 <TravT> so, agenda is here as usual: https://etherpad.openstack.org/p/search-team-meeting-agenda
15:01:46 <TravT> feel free to add
15:01:52 <TravT> i've added a few things as usual
15:02:00 <TravT> #topic Newton Milestone 3 has 1week left
15:02:07 <sjmc7> :O
15:02:27 <TravT> last week we discussed feature freeze for the policy work.
15:03:00 <TravT> we didn't mailing list that, but i could send out a message announcing it
15:03:13 <sjmc7> this would be for the RBAC change?
15:03:26 <sjmc7> to shift rbac to policy?
15:03:31 <TravT> Phase 2 and phase 3 of the policy support
15:04:08 <sjmc7> i’m gonna try and take a first stab at that today. if it looks really gnarly we’ll maybe pass on it; i don’t want to be introducing too much terror this close to the release
15:04:11 <TravT> i think neutron rbac (if that's what you are talking about) should be done as well if not merged by next thursday
15:04:34 <sjmc7> no, talking about phase 2
15:04:48 <RickA-HP> I'm finishing up a test for RBAC and will submiut a new patch today.
15:04:59 <TravT> okay
15:05:34 <TravT> lei-zh: looks like you have a patch up for pipeline
15:05:40 <TravT> https://review.openstack.org/#/c/359972/
15:05:59 <TravT> i haven't looked, but we'll have to really be careful considering it this late
15:06:29 <lei-zh> yeah, I add some test cases, but it'not enough, still work on test cases to ensure it works
15:07:42 <TravT> ok, well, everybody plz give a quick look through.
15:07:47 <TravT> we need to especially evaluate it for risk
15:08:01 <sjmc7> yeah, will do
15:08:16 <lei-zh> I think it's too late for the feature, but it's helpful to give some feedbacks
15:08:37 <lei-zh> so I can continue work on it
15:08:44 <RickA-HP> And feedback you shall get :)
15:09:05 <TravT> ok, great.
15:09:23 <lei-zh> thanks:)
15:09:37 <TravT> So let's do a quick scan on BPs targetted for newton
15:09:37 <TravT> https://blueprints.launchpad.net/searchlight/newton
15:10:22 <TravT> Kevin_Zheng: I just marked nova-server-groups-plugin as implemented.
15:10:33 <TravT> But we'll need a follow on for supporting notifications
15:10:55 <TravT> we did that for flavors and hypervisors as well
15:11:30 <TravT> There is still some ongoing work on the UI.
15:11:50 <TravT> I'd like to propose allowing feature freeze exception on it
15:12:07 <TravT> a few polishing activities
15:12:26 <TravT> there are some bugs in it that i've discovered over the last day or two
15:12:32 <TravT> and it uncovered a bug in Searchlight
15:12:46 <sjmc7> that’ll learn you for testing things
15:12:50 <TravT> most bug fixes can go through RC1
15:13:06 <RickA-HP> TravT: Does "it" for freeze excerptions mean the UI work, the server groups notification or both?
15:13:19 <TravT> server group notifications aren't feature freeze
15:13:22 <TravT> they are out of our control
15:13:28 <TravT> because of a dependency on nova
15:13:38 <TravT> so, there is also a string freeze period
15:13:53 <TravT> so, the UI work will have to freeze on strings shortly as well
15:14:13 <TravT> i believe this also impact log messages localized
15:14:54 <TravT> hard string freeze is 3 weeks from now, i think
15:15:12 <TravT> we've enabled the translation jobs on the UI
15:15:29 <TravT> but there is no guarantee that they'll get picked up by translators, at least for a bit
15:15:46 <sjmc7> maybe we could run a sweepstake
15:16:27 <TravT> jumping around a bit
15:16:33 <TravT> but this is the bug i found using the UI
15:16:34 <TravT> https://bugs.launchpad.net/searchlight/+bug/1616674
15:16:34 <openstack> Launchpad bug 1616674 in OpenStack Search (Searchlight) "OS::Nova::Server ignores build --> error status update" [Undecided,New]
15:16:47 <TravT> I probably wouldn't have noticed it if weren't for some recent patches from tyr_
15:16:51 <TravT> and matt-borland
15:16:55 <sjmc7> ah. that’s probably my fault
15:17:12 <TravT> where the status of an item is detected and a progress bar shows up if it is an in transition status
15:17:19 <matt-borland> ah
15:17:41 <TravT> yeah, i'm going to up the priority on that to critical
15:18:01 <TravT> unless somebody can't reproduce
15:18:02 <tyr_> UI reload doesn't affect it so it appeared to be a missing update to SL index.
15:18:13 <TravT> tyr_: yes, this is a bug for server side
15:18:22 <TravT> your ui helped to me to spot
15:18:47 <sjmc7> i’m sure it’s a server side bug
15:20:10 <TravT> I will need to tag searchlight, ui, and client next week.
15:20:17 <TravT> I'd like to do it by end of day wednesday
15:20:42 <TravT> I'm not sure if we actually have new things for the client
15:20:52 <TravT> I'll have to look
15:20:56 <Kevin_Zheng> sure
15:21:14 <Kevin_Zheng> sorry for the bad network
15:21:15 <tyr_> we wanted to provide the total hit count
15:21:24 <TravT> tyr_: i put a patch up on that last night
15:21:32 <tyr_> stellar!
15:22:20 <TravT> i also put up a patch hiding field matches and another allowing sort by relevancy, date, or alphabetical server side
15:22:53 <tyr_> also was looking at hiding/configuring some of the columns that are shown by default.
15:23:18 <TravT> yes, the field matches patch also hides that by default
15:23:30 <tyr_> to make sure we provide a "consumer" experience, but also don't destroy some of the columns that we use a lot as devs.
15:23:33 <TravT> and we still need to do the same for the visibility column
15:23:34 <tyr_> perfect. thx
15:24:06 <TravT> i also got 80% done with having a toggle between table rows (list view) and tile view.
15:24:45 <tyr_> tile view? Sounds exciting. Targeting Newton?
15:25:28 <TravT> perhaps.
15:25:58 <TravT> I also think i see how to get the pagination in around smart table... i really want that to verify pagination works as expected.
15:26:53 <TravT> there is still a huge user experience gap right now though with the actions
15:27:12 <TravT> i hope that you and matt-borland can figure out how to best handle that
15:27:36 <tyr_> huge gap? Can you elaborate a bit?
15:27:37 <TravT> it is rather confusing to be able to take action on some things within table and not others.
15:28:40 <tyr_> Ooo, you mean we need more content?
15:28:43 <TravT> hehe
15:28:44 <TravT> yep
15:28:56 <tyr_> 100% agree :D
15:29:07 <TravT> but i really think we need that action shim
15:29:16 <TravT> as a stop gap
15:29:19 <tyr_> +1
15:29:58 <tyr_> Unfortunate to need to supply that from searchlight, but byob I guess.
15:30:07 <TravT> yes, it is...
15:30:20 <TravT> but things get extremely hung up in horizon land
15:30:29 <tyr_> its very well isolated, so I don't think it is a problem to include
15:31:08 <TravT> anything we "shim" in for content / actions in SLUI, we need to plan to try to get that into horizon proper next release
15:32:35 <tyr_> oh, we could use a hand verifying that searchlight ui is correctly themable, if anyone has some CSS urges they'd like to express.
15:32:35 <TravT> okay
15:32:57 <TravT> i was going to ping hurgleburgler on it...
15:33:05 <tyr_> that'd be awesome
15:33:56 <TravT> it does work fine switching from material to default
15:34:01 <TravT> i haven't tried anything else
15:34:29 <TravT> we better move on
15:34:37 <TravT> #topic barcelona summit
15:34:51 <TravT> does anybody know if they are going to be able to go?
15:35:56 <sjmc7> chances are not high for me
15:36:16 <lei-zh> nope, still wait for topic vote result, 50% perhaps
15:37:15 <TravT> i'm tempted to ask for 1 fishbowl and working session.  The fishbowl would abuse the concept a bit and actually turn it into a "writing plugins for searchlight" teaching seminar for both backed and UI.
15:37:36 <TravT> working session possibly propose to be a joint one with nova for cells v2
15:38:01 <TravT> they've told us we can't have as many as austin
15:38:09 <TravT> one less days and more projects
15:38:24 <TravT> thoughts?
15:39:14 <sjmc7> if you can get nova involved that’d be good
15:39:31 <RickA-HP> Have we reached out to Nova? Are they receptive?
15:39:32 <sjmc7> see if they’re serious about it, and indicate that the notifications are a bit of an impediment to us
15:40:08 <TravT> well, i went to their meeting this week and was waiting for "open discussion", but got distracted right at the end so missed it
15:40:36 <TravT> i can't image why anybody would get distracted during an IRC meeting... ;-)
15:41:09 <RickA-HP> TravT: Can you repeat that, I wasn't paying attention.
15:41:26 <TravT> huh?
15:41:27 <TravT> :)
15:41:39 <TravT> okay, let's move on
15:41:43 <TravT> #topic open discussion
15:41:57 <TravT> yingjun_: Kevin_Zheng anything?
15:42:14 <yingjun_> nothing from me
15:42:47 <Kevin_Zheng> me neither
15:43:02 <rosmaita> i got something
15:43:06 <TravT> Kevin_Zheng: question
15:43:12 <TravT> sorry rosmaita go ahead
15:43:41 <rosmaita> just want people to look at this and say what they think: http://docs-draft.openstack.org/87/357287/2/check/gate-searchlight-api-ref/5c73760//api-ref/build/html/
15:43:57 <TravT> oh, i'm sorry... i didn't get to that since last week
15:44:02 <rosmaita> not sure the docs team will like all those POST /v1/search calls
15:44:15 <rosmaita> but we should decide if *we* even like it first
15:44:25 <rosmaita> no rush on this, so don't worry
15:44:44 <rosmaita> https://review.openstack.org/#/c/357287/
15:44:47 <sjmc7> oo, that’s nice
15:45:04 <TravT> that is looking good
15:45:43 <rosmaita> feedback i need is how many more calls does it make sense to put in there?
15:45:53 <sjmc7> keystone’s a good example rosmaita
15:46:00 <TravT> i also owe you that...
15:46:12 <sjmc7> they have examples for various authentication options (unscoped, domain, project)
15:47:04 <rosmaita> sjmc7: haven't looked at that, will do
15:47:35 <sjmc7> but i would keep the number pretty low since we have to maintain it; maybe a really simple query_string type query and a little more complicated one. the important thing is the various things that can be in the response
15:47:45 <sjmc7> results, highlighting etc
15:47:50 <sjmc7> so the bare minimum that shows that, i guess
15:48:02 <TravT> I would add in an example term search, boolean search.
15:48:07 <TravT> i still owe you the sample calls.
15:48:23 <sjmc7> i’m slightly against putting loads of exampel searches in if the request payload is all that differs
15:48:41 <sjmc7> but maybe look around and see what other projects do
15:48:46 <TravT> i think a few examples should be done.
15:48:48 <rosmaita> there are a bunch of examples in the current dev docs
15:49:12 <TravT> because most projects have APIs with request parameters fully specified
15:49:19 <sjmc7> yeah. i’m not gonna get worked up about it, just don’t want to maintain stuff in loads of places
15:49:22 <TravT> in our case, most the magic is in the payload...
15:49:24 <tyr_> rosmaita: TravT, sjmc7 and I took several iterations to figure out a good way to query multiple items. That might be a useful example to add.
15:49:49 <sjmc7> ok - but again, not sure if that’s API documentation versus elasticsearch DSL documentation
15:49:59 <TravT> fair enough...
15:50:15 <TravT> maybe at the top, we should give a link to a few query types
15:51:09 <TravT> rosmaita: been pretty slammed, but if i don't have something for you by tomorrow please feel free to ping me directly
15:51:23 <rosmaita> well, this isn't a real rush thing
15:51:34 <rosmaita> we can publish independent of the release milestones
15:51:44 <tyr_> in my example, not API so much as it is suggestions for "optimal" queries given how we expect searchlight to be deployed.
15:51:48 <rosmaita> so i'm not too worried about getting feedback
15:52:01 <rosmaita> i mean, i am, just doesn't have to be immediate
15:52:09 <TravT> I'd also be happy to push the base patch through sooner than later and add on
15:52:49 <rosmaita> ok, maybe we should put up an etherpad with ideas for what we want in the base merge
15:53:02 <TravT> okay, makes sense.
15:53:17 <TravT> i guess you can add a link to that in your patch
15:53:28 <rosmaita> https://etherpad.openstack.org/p/searchlight-api-ref
15:54:03 <TravT> great.
15:54:11 <TravT> i do have a question for Kevin_Zheng if he hasn't fallen asleep
15:54:36 <tyr_> I'm very interested in defining 3-5 common tasks that people know (or even guess) their users do when maintaining their cloud. If you have any data, please pass it on.
15:54:42 <TravT> Re: the healthcheck middleware, is this something that would be required for your company before they'd consider deploying SL
15:55:43 * TravT poor guy may have fallen asleep
15:56:15 <RickA-HP> TravT: I think earlier Kevin alluded to having network problems.
15:56:36 <TravT> okay, well we are pretty much out of time
15:56:58 <TravT> thanks everybody!@
15:57:06 <RickA-HP> Bye
15:57:07 <lei-zh> bye
15:57:29 <TravT> #endmeeting