15:00:19 <TravT> #startmeeting openstack search
15:00:24 <openstack> Meeting started Thu Mar 24 15:00:19 2016 UTC and is due to finish in 60 minutes.  The chair is TravT. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:00:26 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
15:00:28 <openstack> The meeting name has been set to 'openstack_search'
15:00:39 <TravT> o/
15:00:42 <sjmc7> morning
15:00:48 <TravT> morning
15:00:56 <lei-zh> o/
15:01:20 <TravT> Rick is out sick today, i think
15:01:23 <yingjun> morning/night
15:01:29 <rosmaita> o/
15:01:30 <TravT> o/ yingjun
15:01:40 <TravT> lakshmi should join shortly
15:01:56 <TravT> agenda is here: https://etherpad.openstack.org/p/search-team-meeting-agenda
15:02:08 <TravT> if there is anything you want to add, please go ahead!
15:02:21 <nikhil> o/
15:02:24 <TravT> #topic Mitaka RC1 done
15:02:48 <TravT> We had quite a push to finish things out in RC1
15:02:51 <rosmaita> TravT: good job!
15:03:11 <itisha> Hey!
15:03:28 <nikhil> ohai itisha!
15:03:45 <TravT> https://launchpad.net/searchlight/+milestone/mitaka-rc1
15:03:57 <TravT> we have some cool stuff make it
15:04:19 <TravT> but this coming release, we will need to avoid the FFE's
15:04:54 <TravT> this does mean the repo is tagged and we have a stable/mitaka branch now
15:05:13 <TravT> since we do have some bugs floating into RC2, we'll be following the backport process
15:05:31 <TravT> basically, a fix has to be made to master and merged
15:05:41 <TravT> then we cherry pick it to the stable/mitaka branch
15:06:15 <TravT> #topic Mitaka RC2
15:06:27 <TravT> https://launchpad.net/searchlight/+milestone/mitaka-rc2
15:07:14 <TravT> The tough ones are already underway.
15:07:35 <TravT> the medium and two undecided are just documentation fixes, I think.
15:08:15 <TravT> i was looking at the generated release notes yesterday
15:08:22 <TravT> and the format is a little weird
15:08:31 <TravT> does anybody know where those actually get published to?
15:09:24 <TravT> rosmaita any ideas?
15:09:57 <rosmaita> not sure
15:10:09 <rosmaita> nikhil?
15:10:27 <nikhil> hi
15:10:31 <nikhil> was distracted, sorry
15:10:46 <nikhil> oh the reno stuff
15:10:47 <rosmaita> TravT is wondering where the released notes are published
15:11:18 <nikhil> I think here
15:11:20 <nikhil> #link http://releases.openstack.org/mitaka/index.html
15:11:58 <nikhil> or at least a page related to it
15:12:22 <TravT> yeah, i can see the release but not the notes
15:12:24 <TravT> http://releases.openstack.org/mitaka/index.html
15:12:28 <sjmc7_> sorry, got disconnected
15:12:37 <TravT> no worries
15:12:48 <TravT> well, anyway, i can see the release notes generated in a recent build
15:12:48 <TravT> http://docs-draft.openstack.org/56/296356/3/check/gate-searchlight-releasenotes/940f3a1//releasenotes/build/html/mitaka.html
15:12:59 <TravT> and saw we might have missed a few things
15:13:22 <TravT> so, i'll try to put a patch up to correct that.
15:13:55 <TravT> if there is anything missing from this list for bugs for rc2, please let me know or target the bug yourself:
15:13:55 <TravT> https://launchpad.net/searchlight/+milestone/mitaka-rc2
15:14:39 <sjmc7> there were a couple of patches i think we already moved to rc2 cos they weren;’t ready for 1
15:15:45 <TravT> hmm, okay, well, re-target them if we didn't already.
15:15:56 <TravT> #topic searchlight-ui
15:16:17 <TravT> still waiting for infra to move the repo under openstack NS
15:16:29 <TravT> thierry +1'd the governance change
15:16:53 <TravT> and i addressed a review on the infra patch yesterday.
15:17:23 <TravT> yesterday after server side stuff died down a bit, i started doing some little touchups to the plugin
15:17:33 <TravT> instructions for it should actually work now..
15:17:38 <TravT> https://github.com/ttripp/searchlight-ui
15:17:57 <TravT> have a couple other minor touch ups to do to it
15:18:00 <TravT> for mitaka
15:18:24 <sjmc7> i installed it yesterday, seemed to work fine
15:18:37 <TravT> thanks...
15:18:42 <TravT> david-lyle saw you dropped
15:18:58 <TravT> so just FYI that i've been doing some minor fixes to the searchlight-ui repo
15:19:32 <TravT> okay, that's all for that for now.
15:19:54 <TravT> #topic newton
15:20:28 <TravT> i've started "grooming" the blueprints a bit
15:20:31 <TravT> https://blueprints.launchpad.net/searchlight
15:20:53 <TravT> i think next week we should go through them together and start talking priorities
15:21:21 <nikhil> I've a question after the update is over
15:21:24 <TravT> but if there is something that you'd like to have considered for newton that isn't on there, please add it
15:21:40 <TravT> nikhil: re this or just in general?
15:21:55 <nikhil> regarding this and may be in general :)
15:22:06 <TravT> go ahead!
15:22:33 <nikhil> We (artifacts team) had a chat with app-catalog folks yesterday
15:22:45 <nikhil> on their requirements for artifacts and how they should work
15:22:54 <nikhil> one things came up again and again
15:23:20 <nikhil> they want to scale the registry of artifacts across internet (www -- basically)
15:23:28 <david-lyle> TravT: net split ended up on the wrong side
15:23:58 <nikhil> are folks around then?
15:24:14 <TravT> roll call!
15:24:16 <TravT> i'm here
15:24:18 <sjmc7> still here
15:24:33 <rosmaita> here
15:24:34 <nikhil> I will continue
15:24:46 <nikhil> they were hoping to see if they can use SL or if they should use elastic search for this
15:25:12 <nikhil> mostly they good information out to the users and that's real time-like update
15:25:19 <TravT> if they have any RBAC concerns, then i think taking advantage of searchlight concepts would be very beneficial to them
15:25:20 <sjmc7> what do you mean about ‘scale it across the internet'?
15:25:58 <nikhil> the way app-catalog is setup right now is a website apps.openstack.org and they are planning to run a Glare behind it
15:26:09 <nikhil> the end users can be anyone in the world
15:26:27 <TravT> so, how do they deploy this today?
15:26:28 <nikhil> and they will have read access to listing, download etc
15:26:44 <TravT> do they have multiple datacenters with replicated databases behind it?
15:27:03 <nikhil> it used to be hard coded and now they are using experimental API (not sure though at what capacity exactly)
15:27:12 <nikhil> TravT: nah!
15:27:40 <nikhil> I don't know what their plan is around geo-spatial HA
15:27:56 <nikhil> but they seemed to be on a plan to have a federated like setup
15:28:08 <nikhil> for scientific community for the most part
15:28:32 <nikhil> for eg. different National labs would want to share images, templates etc assets
15:28:41 <nikhil> and keep them versioned and tagged
15:28:50 <nikhil> all of it kept in app-catalog
15:29:11 <nikhil> but they may run a Glare on their end as well (or some other too)
15:29:26 <nikhil> that will get the latest of the assets from app-catalog into the local host
15:29:43 <nikhil> (cloning -- if you may)
15:30:24 <TravT> so who all is on this team?
15:30:35 <nikhil> artifacts or app catalog?
15:30:55 <nikhil> (both are different teams)
15:31:11 <TravT> so is Glare and "artifacts" team the same thing now?
15:31:12 <nikhil> (and app catalog is a consumer of artifacts)
15:31:20 <nikhil> TravT: correct
15:31:38 <TravT> do you know their timeline?
15:31:45 <nikhil> GLARE == GLance Artifacts REpository
15:32:22 <nikhil> TravT: there wasn't a timeline posted. we're still working on the Glare API approval from defcore etc folks
15:32:31 <nikhil> may be this is newton may be ocata
15:32:49 <nikhil> (hence, regarding this and may be in general)
15:32:58 <TravT> okay, it sounds interesting.
15:33:05 <nikhil> anyways, the specific reason I brought this up was -- would we consider adding artifacts support?
15:33:20 <sjmc7> we already have a BP for artifact support, although it’s not very well fleshed out
15:33:26 <TravT> ativelkov brought this up at the last summit
15:33:44 <nikhil> yeah, but the API and architecture is now completely changed (Again!)
15:33:45 <TravT> so we entered a low BP
15:33:53 <TravT> yeah, that's what he mentioned
15:34:01 <TravT> that there needed to be some more work on glare first
15:34:23 <nikhil> it won't be very different from Glance
15:34:29 <nikhil> I mean Glare
15:34:35 <nikhil> (just more generic)
15:35:04 <nikhil> but Glance was a tough one with the prop protections and tags and notifications (Afair)
15:35:30 <TravT> well, if their is a desire for any authentication / authorization / management tooling around ES, then SL has added benefit over starting from scratch with ES
15:35:38 <TravT> their -> there
15:35:57 <nikhil> oh one more thing, they do not have/need keystone support
15:36:18 <nikhil> they use some other sort of authN
15:36:38 <sjmc7> i think we maybe need to take this offline
15:36:47 <nikhil> but I thought the UI and may be horizon integration would be benefitial
15:36:50 <nikhil> sure
15:36:57 <nikhil> I wanted to know: what would be blockers? may be the lessons learnt will make things easier?
15:37:03 <nikhil> that we can discuss offline
15:37:16 <sjmc7> auth would be one we’d need to discuss
15:37:49 <TravT> yeah other auth provider is something we also need to look further into for swift, i think...
15:37:58 <sjmc7> deployment would be another; if it’s not colocated with SL, there’s got to be a mechanism for getting information into SL
15:38:32 <TravT> nikhil, would the timeline for discussing this call for something pre-summit or is this a potential summit session?
15:38:59 <nikhil> TravT: I was thinking at the summit should be good enough. May be even at the contributors meetup would work.
15:39:17 <nikhil> depending on the availability and priorities
15:39:39 <david-lyle> nikhil: Glare doesn't use keystone?
15:39:57 <nikhil> david-lyle: oh sorry, it's app catalog that doesn't need it.
15:39:58 <sjmc7> the app catalog presumably doesn’t
15:40:08 <nikhil> Glare will be exactly like Glance on the basics of openstack.
15:40:11 <david-lyle> ok, less confused now
15:40:25 <sjmc7> so no keystone at all, not just a federated provider?
15:40:47 <nikhil> I prolly shouldn
15:41:01 <nikhil> shouldn't have used the word federated
15:41:01 <david-lyle> community app-catalog isn't running on top of openstack
15:41:02 <sjmc7> yeah, we should maybe not make wild assertions here. we’ll track them down at the summit
15:41:07 <david-lyle> IIUC
15:41:27 <nikhil> david-lyle: yes and now they want to use artifacts (glare) to store assets
15:41:43 <nikhil> sjmc7: correct
15:42:02 <david-lyle> nikhil: where?
15:42:15 <nikhil> I brought this up to understand the project priorities and see if we can try to collaborate and get some momentum on this in newton even if it's just small things
15:42:48 <sjmc7> ok. we’d need to understand the glare data model first off to be able to index anything
15:43:01 <sjmc7> and then look at how they plan to deploy it in this instance
15:43:04 <nikhil> david-lyle: right now (afict) they hard code the links to assets
15:43:27 <nikhil> david-lyle: they will want to run a Glare backing the apps.openstack.org website
15:44:16 * david-lyle still confused, but willing to learn more later
15:44:34 <nikhil> sounds good.
15:44:39 <sjmc7> i think i’ll have that inscribed on my gravestone, david-lyle
15:44:58 <david-lyle> :)
15:45:13 <david-lyle> that's one fun authorization model
15:45:43 <nikhil> sorry about the confusion folks..
15:45:46 <david-lyle> fun as in ¯\_(ツ)_/¯
15:45:51 <TravT> nikhil: thanks for bringing it up
15:45:58 <nikhil> :)
15:46:10 <TravT> it does sound like a good summit conversation
15:46:31 <nikhil> looking forward to it
15:47:15 <TravT> only one other thing i had to mention
15:47:24 <TravT> #topic notification bugs
15:48:07 <TravT> sjmc7 already started and we should continue to get notification bugs / bps filed against other projects.
15:48:48 <TravT> getting as much data via notification as possible is critical to scalability
15:49:01 <sjmc7> yeah. nova’s the big one
15:49:20 <TravT> yeah, we need to figure out the versioned notifications with them.
15:49:29 <TravT> i opened a bp on that
15:49:30 <TravT> https://blueprints.launchpad.net/searchlight/+spec/nova-versioned-notifications
15:49:36 <sjmc7> yes, that too
15:49:41 <TravT> with us...
15:49:59 <TravT> yingjun: have you made any progress on hypervisors?
15:50:33 <yingjun> not yet
15:50:33 * TravT he might be asleep by now
15:50:38 <sjmc7> :)
15:51:20 <TravT> yingjun okay, i think that one will be pretty important to us.
15:51:59 <TravT> okay...
15:52:09 <TravT> #topic open discussion
15:52:42 <yingjun> yes, i’ll put a initial patch recently without the notification
15:53:01 <sjmc7> are there notificaitons at all for hypervisors?
15:53:19 <yingjun> no
15:54:00 <TravT> https://blueprints.launchpad.net/nova/+spec/hypervisor-notification
15:54:14 <sjmc7> ok
15:54:19 <TravT> we'll need to flesh that out a bit... perhaps after yingjun does initial plugin
15:54:31 <TravT> but soon enough to maybe get it noticed in nova land
15:55:38 <rosmaita> i have a question about that
15:55:48 <rosmaita> who is hypervisor info aimed at?
15:55:57 <sjmc7> admins
15:56:02 <rosmaita> i thought the whole point of hte cloud is users don't worry
15:56:24 <rosmaita> hmmm ... you'd hope they'd have some kind of inventory system
15:57:09 <TravT> this is for operators
15:57:17 <rosmaita> ok
15:57:47 <TravT> rosmaita do you still do any nova work?
15:58:04 <rosmaita> sort of
15:58:35 <rosmaita> still mostly glance, but some nova
15:59:13 <yingjun> it would be helpful if someone could move on the hypervisor thing in nova..
15:59:14 <TravT> okay.
15:59:28 <sjmc7> anyone can submit a patch :)
15:59:51 <TravT> maybe rosmaita knows the right person to bribe, though.
15:59:53 <TravT> ;)
16:00:20 <rosmaita> i'll ask johnthetubaguy what he thinks
16:00:27 <TravT> cool, thanks!
16:00:28 <rosmaita> though he'll be away until tuesday
16:00:33 <rosmaita> np
16:00:35 <TravT> guess we're out of time for today.
16:00:37 <TravT> #endmeeting