15:00:39 <TravT> #startmeeting openstack search
15:00:40 <openstack> Meeting started Thu Jun 16 15:00:39 2016 UTC and is due to finish in 60 minutes.  The chair is TravT. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:00:41 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
15:00:43 <openstack> The meeting name has been set to 'openstack_search'
15:00:47 <david-lyle_> o/
15:00:47 <sjmc7> good mornin
15:00:51 <TravT> o/
15:00:55 <sjmc7> david’s on time!!
15:01:00 <lei-zh> o/
15:01:01 <TravT> i also noticed
15:01:04 <TravT> very impressed
15:01:06 <RickA-HP> o/
15:01:15 <TravT> guess your talk with him last week paid off!
15:01:25 <david-lyle> 20 lashes
15:01:43 <sjmc7> i follow 18th century british navy punishments
15:01:54 * TravT afraid to ask
15:02:05 <TravT> o/ yingjun_
15:02:20 <TravT> okay, https://etherpad.openstack.org/p/search-team-meeting-agenda
15:02:23 <TravT> agenda^
15:02:51 <TravT> #topic:     ElasticSearch 1.x support for Newton?
15:02:58 <TravT> looks like this was discussed last week
15:03:11 <TravT> #topic ES 1.x support for newton
15:03:21 <TravT> i seem to be failing at meeting bot
15:04:01 <TravT> anyway, outcome was several people suggesting that we don't officially support 1.x for Newton
15:04:25 <TravT> And there was discussion of sending to ML
15:04:48 <TravT> seems like the discussion on moving devstack to ES 2.x on the ML said just do it.
15:05:01 <TravT> so, maybe we should just submit that up first.
15:05:04 <TravT> and see how it goes
15:05:17 <yingjun_> o/
15:05:38 <sjmc7> yeah, i was meant to do that, wasn’t i
15:05:44 <RickA-HP> Do we have any feedback from actual cuustomers about upgrading ES to 2.X?
15:06:12 <TravT> well, from an HP distro perpsective, we know that it is going out now with ES 2.x
15:06:17 <TravT> but I don't know about others.
15:06:19 <sjmc7> i’ll do it today. the blocker from an openstack perspective would be any services that rely on v1
15:06:22 <TravT> i don't know if the OS survey covers it
15:06:35 <TravT> sjmc7, this seems more like it should go to the operators ML
15:07:18 <RickA-HP> sjmc7: Or more importantly, services that rely on v1 but do not realize it!
15:07:37 <TravT> yeah, this is why we should upgrade devstack
15:07:47 <TravT> we still have one defect to fix in SL
15:08:05 <TravT> or more correctly, get through reviews.
15:08:43 <TravT> https://review.openstack.org/#/c/325698/
15:08:58 <TravT> i started to look at it again yesterday
15:09:08 <TravT> but didn't finish
15:09:29 <TravT> so, let's get that one landed and then get devstack upgraded
15:09:42 <TravT> maybe we can get that done by next week
15:10:12 <TravT> i'm not sure how deprecation policy works for shared infrastructure components
15:10:37 <TravT> do we have to get all services to first declare it is deprecated?
15:10:49 <TravT> dhellmann: are you around?
15:10:51 <sjmc7> i don’t know how officially it’s supported
15:11:04 <sjmc7> it’s only peripherally used
15:11:14 <sjmc7> i don’t know of any core services that rely on it
15:11:23 <lei-zh> if devstack gets upgraded, will other services using 1.x be affected
15:11:40 <TravT> lei-zh yes, and i think that's kind of the point of upgrading it
15:11:49 <TravT> forces 2.x to get tested
15:12:50 <TravT> oh interesting
15:12:51 <TravT> https://review.openstack.org/#/c/320152/
15:12:58 <TravT> see Sean's comment there
15:13:50 <TravT> i think that'd mean each service runs its own ES then.
15:13:56 <TravT> for devstack.
15:14:02 <david-lyle> that's fine and good unless plugins require competing versions
15:14:14 <sjmc7> yeah, that seems a bit weird
15:14:22 <TravT> yeah, i think i'll add a comment asking about that
15:14:24 <david-lyle> containers?
15:14:30 <TravT> hmm
15:14:37 <TravT> maybe a kolla thing?
15:14:42 <sjmc7> um
15:15:13 <TravT> if they make each plugin do this
15:15:19 <TravT> then we could still pull it down.
15:15:24 <TravT> but not install it
15:15:32 <sjmc7> sure, but we won’t know which version we’re gonna end up with
15:15:34 <TravT> just run the script
15:15:48 <TravT> and make sure we are the only ones on that port
15:16:06 <david-lyle> I suppose as long as upper-constraints points to the version we care about
15:16:19 <david-lyle> we can say the other project is in the wrong
15:16:21 <RickA-HP> Having each service use it's own ES is good for developers, but no so much for customers.
15:16:23 <sjmc7> upper constraints doesn’t deal with databases
15:16:36 <david-lyle> it's in global requirements
15:16:49 <sjmc7> the pytohn library’s in global-reqs
15:16:55 <david-lyle> oh, that's just the library
15:16:58 <david-lyle> yeah
15:16:59 <david-lyle> hmm
15:17:10 <sjmc7> yeah. and there is a potential mismatch there also, hence why moving to 2.x only would be nice
15:17:18 <david-lyle> seems like you would want a common version in openstack
15:17:32 <RickA-HP> david-lyle: This is giving me flash backs of porting PalmOS :)
15:17:35 <TravT> david-lyle: i agree
15:17:44 <david-lyle> RickA-HP, :)
15:17:47 <david-lyle> and :(
15:17:55 <TravT> maybe you can also chime in on that review
15:18:05 <sjmc7> i guess from devstack’s perspective they don’t want to support every thing people decide to go off and use
15:18:26 <david-lyle> sure, I think the counter will be that devstack is not openstack proper and that's a problem outside their scope
15:19:05 <RickA-HP> When does something become devstack proper? How many services need to be using it?
15:19:09 <sjmc7> it’s not a big deal to do it ourselves
15:19:16 <david-lyle> RickA-HP, I meant devstack is not
15:19:18 <sjmc7> devstack is part of infra
15:19:41 <RickA-HP> Ok.
15:19:43 <david-lyle> so even if it's consistent in devstack, it won't be for real installations
15:20:18 <TravT> RickA-HP: devstack proper kicked all but the core openstack services out awhile back
15:20:44 <TravT> i think they might want to kick everything out at some point.
15:20:48 <RickA-HP> Hmmm
15:20:56 <TravT> david-lyle didn't they want horizon out?
15:21:02 <sjmc7> anyway - sounds like we need to get buy in from other services (of which ceilo is the onyl one i know of that use elasticsearch at all) and from deployers
15:21:08 <david-lyle> TravT, I don't think so
15:21:11 <david-lyle> re: horizon
15:21:19 <david-lyle> tempest wanted us out
15:21:26 <TravT> ah ok
15:21:34 <sjmc7> there’s only one test to kick out :)
15:21:43 <sjmc7> this is drifting a little off topic
15:21:46 <david-lyle> devstack as an introduction is better with a GUI
15:22:05 <david-lyle> sjmc7, a little?
15:22:27 <david-lyle> I think we just own the es version for devstack
15:22:39 <david-lyle> but research the greater openstack question?
15:22:51 <sjmc7> i can’t think of any other serivces using E-S
15:22:59 <lei-zh> monasca using elasticsearch too?
15:23:00 <david-lyle> ceilo
15:23:05 <sjmc7> yeah, maybe monasca
15:23:10 <david-lyle> unless they stopped
15:23:10 <sjmc7> ceilo sort of shrugged when i asked
15:23:17 <sjmc7> said “sure, go ahead, we’ll see if something breaks"
15:23:27 <sjmc7> monasca supports 2.x already
15:23:29 <david-lyle> we'll shrug back if it does
15:24:18 <TravT> okay, i just left a comment on that review.
15:24:48 <TravT> our action is to get our patches through (one mentioned above)
15:25:14 <TravT> sjmc7 did lakshmiS finish the work to get ES 2.x going for functional tests while I was out?
15:26:28 <TravT> guess i'm off the rocker here or something.
15:26:33 <sjmc7> what specifically?
15:26:44 <sjmc7> yes, that’s what broke everything
15:26:59 <TravT> so that is still check experimental only, right?
15:27:15 <sjmc7> it was voting, but now it’s experimental because it won’t pass
15:27:20 <TravT> right
15:27:32 <TravT> okay, just wanted to verify where we were.
15:27:48 <TravT> i guess we'll need to revisit this once we have that sorted out
15:27:55 <TravT> and that last patch through.
15:28:09 <TravT> #topic Glance snapshot faceting
15:28:13 <TravT> rosmaita: are you here?
15:28:26 <TravT> i know that he said he was at a midcycle or something this week
15:29:14 <TravT> anyway, i was trying to review lakshmi's fix to add snapshot facets for images
15:29:18 <TravT> and i couldn't get it to work
15:29:33 <TravT> i was hoping somebody could confirm whether or not the related bug i filed is valid
15:29:35 <TravT> https://bugs.launchpad.net/horizon/+bug/1592985
15:29:35 <openstack> Launchpad bug 1592985 in OpenStack Dashboard (Horizon) "Instance snapshots no longer filterable" [Undecided,New]
15:29:56 <TravT> or if it is just my env or i have suddenly lost my mind
15:30:48 <rosmaita> TravT: am in glance midcycle meetings, have not had a chance to verify your findings
15:30:58 <rosmaita> (about the instance snapshots)
15:31:00 <TravT> rosmaita: okay, thanks.
15:31:14 <TravT> if anybody else here is able to try with a recent devstack it would be helpful
15:31:38 <rosmaita> TravT: it may depend on what microversion you are using
15:31:50 <TravT> hmm, that's interesting.
15:31:57 <rosmaita> or problematic
15:31:59 <TravT> so i used nova CLI (not specifying microversion)
15:32:12 <TravT> and I tried from horizon (which I don't think specifies one)
15:32:12 <rosmaita> that should use "latest"
15:32:19 <rosmaita> (if you don't specify)
15:32:49 <rosmaita> although i dont' really know if that's true (about the client)
15:33:05 <rosmaita> supposed to use latest if you contact api without including the version header
15:33:20 <rosmaita> nova cli may pass something
15:33:42 <rosmaita> may have to use --debug to see what's being passed by novaclient
15:33:52 <TravT> okay, i can try some more things
15:34:17 <lei-zh> I thought when I first saw that patch, creating snapshot in horizon worked, I'll try to verify it tomorrow
15:34:30 <TravT> lei-zh: it did for me as well
15:34:35 <TravT> but that was about a month ago
15:34:45 <TravT> with my devstack from this week, it was different
15:34:50 <TravT> i definitely appreciate you trying it
15:35:04 <TravT> okay, moving on
15:35:13 <TravT> #topic Versioning conflict
15:35:27 <TravT> fyi that yesterday i saw some errors with versioning conflicts
15:35:32 <TravT> in both swift and designate
15:35:39 <TravT> https://bugs.launchpad.net/searchlight/+bug/1584148
15:35:39 <openstack> Launchpad bug 1584148 in OpenStack Search (Searchlight) "Version conflict during re-indexing" [Critical,New]
15:36:03 <TravT> lakshmi told me that he has seen it with designate, but wasn't able to replicate with swift.
15:36:18 <TravT> so, if anybody else see this error, please mark the bug as confirmed
15:36:57 <TravT> #topic review
15:37:47 <TravT> This one could use some eyes
15:37:48 <TravT> https://review.openstack.org/#/c/317100/
15:38:08 <TravT> just a mentioning it
15:38:24 <TravT> that's all i have for real topics for today
15:38:29 <TravT> anybody else?
15:38:33 <TravT> #open discussion
15:38:39 <TravT> #topic open discussion
15:39:14 <sjmc7> sorry, someone keeps butt dialling me
15:39:26 <TravT> lol
15:40:26 <sjmc7> don’t think i had anything specific
15:40:45 <RickA-HP> Nothing from me.
15:41:11 <lei-zh> I'm doing the pipeline on a forked repo, try to put the demo out next week for some discussions
15:41:21 <TravT> lei-zh: okay, cool!
15:41:55 <TravT> okay, then
15:42:02 <TravT> thanks everybody
15:42:04 <TravT> #endmeeting