15:01:18 <TravT> #startmeeting openstack search
15:01:19 <openstack> Meeting started Thu Mar 17 15:01:18 2016 UTC and is due to finish in 60 minutes.  The chair is TravT. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:01:20 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
15:01:22 <openstack> The meeting name has been set to 'openstack_search'
15:01:52 <lakshmiS_> O/
15:02:11 <TravT> o/
15:02:15 <sjmc7> hey
15:02:16 <RickA-HP> o/
15:02:25 <lei-zh> o/
15:02:30 <stevelle> o/
15:03:07 <TravT> rosmaita sent me a message and said he accidentally scheduled another appointment
15:03:17 <yingjun> o/
15:03:26 <TravT> but said he'll catch up on logs and to let him know what reviews are needed.
15:04:06 <TravT> So, here's our meeting agenda
15:04:11 <TravT> https://etherpad.openstack.org/p/search-team-meeting-agenda
15:04:17 <TravT> as always, add to it if you have something
15:04:26 <TravT> #topic searchlight-ui
15:04:46 <TravT> I extracted out the searchlight patch and created a working horizon plugin yesterday
15:04:55 <TravT> it was on horizon master
15:04:55 <sjmc7> nice!
15:05:28 <TravT> at the mid cycle (really end of cycle) for horizon we decided to do that at least for mitaka to have a UI release vehicale
15:05:33 <TravT> vehicle
15:05:50 <TravT> i'm not sure how long infra and governance will take to accept it
15:05:56 <TravT> but here's the details:
15:06:03 <TravT> upstream repo: https://github.com/ttripp/searchlight-ui
15:06:03 <TravT> pypi: https://pypi.python.org/pypi/searchlight-ui
15:06:03 <TravT> infra: https://review.openstack.org/#/c/293848/
15:06:05 <TravT> governance: https://review.openstack.org/#/c/293850/
15:06:39 <TravT> there are more things i want to do with it / need to do with it... but there always are.
15:06:51 <TravT> i'll see what i can do today.
15:07:06 <TravT> at this point they are pretty minor (some sorting of labels).
15:07:10 <sjmc7> it looks like sleeping wasn’t something you did much of yesterday...
15:07:26 <TravT> yeah, i'm not clear whether or not i'm asleep right now
15:07:47 <sjmc7> :)   is horizon at a point where we could also hook in say the top navbar search from here?
15:07:55 <sjmc7> i’ve lost track of things a bit
15:08:01 <TravT> no, i need to do a bit of refactoring for that.
15:08:06 <TravT> it really isn't much to do
15:08:22 <TravT> but on horizon master we focused on landing enough patches to support the panel
15:08:23 <sjmc7> ok. was just curious, don’t do more work on that account
15:08:25 <TravT> and we got those in
15:08:55 <TravT> i need to verify the devstack plugin script in that github repo
15:09:26 <TravT> i know creating the plugin with sdist and following the instructions in the readme work
15:10:09 <TravT> okay, so next up.
15:10:20 <TravT> #topic mitaka RC1
15:10:29 <TravT> https://launchpad.net/searchlight/+milestone/mitaka-rc1
15:10:48 <TravT> just a tremendous job the last week or so
15:11:30 <TravT> sjmc7, RickA-HP, lakshmiS_ special thanks to the extra efforts
15:11:42 <RickA-HP> TravT: My pleasure!
15:12:10 <sjmc7> not sure about pleasure, but we’re nearly there
15:12:12 <TravT> lei-zh: thanks for your patience with all the rebases
15:12:39 <TravT> so first let's talk blueprint
15:12:47 <TravT> Provide seamless management for zero downtime reindexing
15:13:06 <TravT> this one i would like to mark implemented, but I want to make sure we've got all bugs logged for known issues
15:13:27 <TravT> are we aware of bugs on this that still need to be logged?
15:14:06 <sjmc7> we got the one about double reporting of child plugins?
15:14:14 <RickA-HP> TravT: I need to log two defects: 1. More checks on thew resource_group_name so it doesn't break index creation.
15:14:57 <RickA-HP> 2. Using scan/scroll to index existing doc types that are not specified on the command line.
15:15:02 <TravT> sjmc7: that one is in here: https://bugs.launchpad.net/searchlight/+bug/1557303
15:15:03 <openstack> Launchpad bug 1557303 in OpenStack Search (Searchlight) "searchlight-manage index sync output doesn't tell why more than one type selected" [High,Triaged]
15:15:28 <RickA-HP> sjmc7: I also have a fix for the double reporting. It will be checked in today, along with the next fix for better error handling.
15:15:54 <sjmc7> ok
15:16:04 <david-lyle> o/
15:16:18 * david-lyle once thankful for daylight savings
15:16:20 <TravT> o/ david-lyle
15:16:57 <TravT> david-lyle if you can scroll back, i would appreciate your eyes on the horizon plugin items i put up above
15:17:35 * david-lyle scrolling
15:17:36 <TravT> ok, so RickA-HP if you can please log them and target for mitaka rc1
15:17:45 <RickA-HP> TravT: Yes
15:17:51 <TravT> although i think i have to tag rc1 later today.
15:18:50 <sjmc7> it’s definitely today?
15:19:58 <TravT> i don't know...
15:20:01 <TravT> the message is a little unclear
15:20:03 <TravT> http://lists.openstack.org/pipermail/openstack-dev/2016-March/089001.html
15:20:08 <TravT> dhellmann: are you around?
15:21:01 <david-lyle> the posted schedule indicated this week, but not sure if that's written in stone
15:21:15 <david-lyle> usually there is some flexibility
15:21:22 <TravT> mitaka 3 seemed to have a definitive deadline
15:21:39 <david-lyle> TravT: yes
15:21:43 <david-lyle> that's more firm
15:21:55 <sjmc7> all i’ve seen is “week of”. reason i ask is i’ve got the one neutron patch to finish
15:21:56 <david-lyle> RC-1 is when you don't have any known blockers
15:22:12 <TravT> yeah, that's what i thought as well.
15:22:18 <david-lyle> getting to that point isn't date based
15:22:29 <david-lyle> but at a point you just have to ship something
15:22:34 <sjmc7> crazy talk!
15:22:40 <TravT> okay, so then let's go with that.
15:22:50 <TravT> it would be good for us to have a few days to bang on it
15:23:07 <david-lyle> well there's RC-2 possibilities too
15:23:08 <sjmc7> then RC2 in a week or so?
15:23:17 <david-lyle> but the bar is higher
15:23:32 <sjmc7> there’re only one or two big things that we know of for RC1
15:23:42 <david-lyle> certainly not feature adding for RC-2
15:23:46 <sjmc7> then if we find anything next week we can decide if it’s big enough to warrant fixing
15:23:47 <sjmc7> right
15:24:01 <david-lyle> clean up ship stopping bugs
15:24:12 <david-lyle> and merge translation where applicable
15:24:13 <sjmc7> say that ten times fast!
15:24:31 <david-lyle> great, now I swore at my daughter, thanks sjmc7
15:24:37 <sjmc7> you’re both welcome
15:24:56 <sjmc7> ok. so we’re aiming for asap but tomorrow is probably ok?
15:25:04 <TravT> yes
15:25:09 <david-lyle> that would be my hope
15:25:56 <TravT> okay, so then, bug list
15:26:40 <TravT> we have a number of items on there
15:26:44 <TravT> that are still in new status
15:27:05 <TravT> I think we'll just need to move functional tests to newton 1
15:27:10 <sjmc7> yes
15:27:21 <TravT> #1442806 and 1552809
15:29:19 <TravT> https://bugs.launchpad.net/bugs/1531267
15:29:21 <openstack> Launchpad bug 1531267 in OpenStack Search (Searchlight) "Nova plugin should listen to compute.instance.suspend events" [Low,In progress] - Assigned to Itisha Dewan (ishadewan07)
15:29:22 <TravT> we need to triage it
15:29:41 <sjmc7> yeah. i’ve flip flopped on it twice
15:29:44 <TravT> status is a definitely a field presented in the ui
15:30:16 <sjmc7> if we were notification-only i’d say yes, but since we’re not i’m hesitant. someone needs to see what the effect with and without is
15:31:41 <TravT> okay, well, next week it'll be obvious what didn't make it
15:31:46 <TravT> so we'll just clean up that list then
15:32:11 <TravT> if there are any bugs missing from this list, please set the milestone target appropriately
15:32:24 <TravT> #topic reviews
15:33:36 <TravT> priority today is: https://review.openstack.org/#/c/293771/
15:33:51 <sjmc7> i have oen up to add routers and subnets; as lakshmiS_ pointed out it needs one more notification added for routers
15:34:09 <sjmc7> neutron’s model is very complicated so there is a chance i’ll have missed other things
15:34:28 <sjmc7> yes, that one
15:35:10 <TravT> and anything to do with zero downtime reindexing.
15:35:15 <TravT> this one is really cool.
15:35:36 <TravT> the basics are that now we are able to completely re-index without disruption to the end user
15:35:47 <TravT> took a bit of magic
15:35:55 <TravT> and some changes.
15:36:26 <TravT> sjmc7 or RickA-HP maybe you could talk a bit about the late breaking changes we made
15:36:44 <sjmc7> yeah.. one sec, let me find the link
15:36:55 <sjmc7> https://launchpad.net/bugs/1558240
15:36:56 <openstack> Launchpad bug 1558240 in OpenStack Search (Searchlight) "Disable resource_group_name per plugin" [Critical,Fix released] - Assigned to Steve McLellan (sjmc7)
15:37:22 <sjmc7> we discovered with some horror while testing the swift work that with data spread across multiple indices there are some problems with some query types
15:37:33 <sjmc7> notably parent/child ones, which we are using in several places
15:38:08 <sjmc7> for this release we made the decision to enforce everything on a single index (though that index can be spread across many machines)
15:38:29 <sjmc7> we’ll revisit it in newton
15:39:30 <sjmc7> that is all.
15:40:01 <TravT> all right then
15:40:16 <TravT> #future thoughts
15:40:22 <TravT> #topic future thoughts
15:40:45 <TravT> I've been pondering whether we should craft a cross project spec on notification standards
15:41:10 <sjmc7> the other option is what we discussed with swift
15:41:17 <TravT> true, true.
15:41:30 <lakshmiS_> Notification clients
15:42:00 <sjmc7> getting info from the service APIs in middleware, which i’d like to prototype when i’m done with 60 hour weeks
15:42:34 <sjmc7> but i do think we need to get away from hititng the APIs as soon as possible
15:42:51 <TravT> i think we should spend the first week or so post RC2 doing that and coming up with a game plan to start out Newton
15:43:00 <sjmc7> yep
15:43:08 <sjmc7> i’d like ot take something to the summit
15:43:19 <TravT> getting specs / bp / (swift POC) up will allow us to actually be in the conversation for newton changes we'd like from other project
15:43:48 <TravT> and then of course, submitting patches
15:44:30 <TravT> so, just tuck that away under your hat.
15:44:33 <TravT> to ponder
15:44:59 <TravT> #topic open discussion
15:45:08 <TravT> That's all I had in mind for today
15:45:31 <sjmc7> i didn’t have anything else
15:45:56 <david-lyle> the horizon related bits above seem reasonable
15:46:14 <david-lyle> still willing to help with a plugin repo
15:47:00 <TravT> i went ahead and created one: https://github.com/ttripp/searchlight-ui
15:47:21 <david-lyle> TravT: I meant getting it into the namespace
15:47:27 <TravT> oh, yes, please
15:47:42 <david-lyle> but I'll look at the plugin structure
15:47:45 <david-lyle> too
15:47:54 <TravT> i put up these patches:
15:47:57 <TravT> 09:06 TravT: infra: https://review.openstack.org/#/c/293848/
15:47:57 <TravT> 09:06 TravT: governance: https://review.openstack.org/#/c/293850/
15:48:20 <TravT> i modeled them on a combination of your trove patches doing similar and neutron lbaas dashboard
15:48:31 <david-lyle> TravT: I managed to skip over that
15:48:35 <david-lyle> :)
15:48:44 <TravT> but getting your eye on those would be appreciated as well
15:48:45 <david-lyle> looks like you've got it all worked out
15:48:55 <david-lyle> sure, will check them
15:49:28 <TravT> it seems most dashboard plugins don't do very much in the way of testing.
15:49:39 <TravT> i put more tests than most...
15:49:44 <TravT> in terms of zuul
15:50:05 <TravT> but no integration tests, etc.
15:50:08 <sjmc7> it’s very hard to do functional tests in zuul unfortunately
15:50:11 <TravT> i don't think we want to do that yet
15:50:20 <david-lyle> the trove and sahara have tests and sahara has integration tests
15:50:34 <david-lyle> the dashboard bits
15:50:35 <sjmc7> are they run in the gate, david-lyle ?
15:50:47 <david-lyle> should
15:50:52 <david-lyle> have to double check
15:51:00 <TravT> i should clarify
15:51:01 <david-lyle> the unit tests do for sure
15:51:17 <sjmc7> yeah. i know i had this discussion on murano way back when, but things change fast
15:51:25 <sjmc7> i’ll take a look at sahara
15:51:29 <TravT> most the javascript based ui's don't do integration tests
15:51:46 <TravT> afaict
15:51:51 <david-lyle> yes, sahara-dashboard integration tests run
15:52:46 <david-lyle> TravT: I have a theory on your observation
15:52:50 <david-lyle> :P
15:52:53 <TravT> hehe
15:53:01 <sjmc7> :)
15:53:45 <TravT> okay, let's call it a meeting
15:53:51 <TravT> thanks everybody!
15:53:59 <david-lyle> thanks TravT
15:54:12 <TravT> #endmeeting