15:00:39 <TravT> #startmeeting openstack search
15:00:41 <openstack> Meeting started Thu Dec  3 15:00:39 2015 UTC and is due to finish in 60 minutes.  The chair is TravT. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:00:42 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
15:00:45 <openstack> The meeting name has been set to 'openstack_search'
15:01:42 <TravT> o/
15:01:42 <nikhil> o/
15:01:47 <sjmc7> ello
15:01:51 <yingjun> hi
15:02:24 <TravT> maybe a smaller crowd here today.
15:02:57 <TravT> Here's the agenda: https://etherpad.openstack.org/p/search-team-meeting-agenda
15:03:02 <nikhil> what has happened to the world!?
15:03:12 <rosmaita> o/
15:03:35 <rosmaita> nikhil: don't ask! bad shit is going on everywhere
15:03:44 <TravT> yeah, seriously
15:03:47 <nikhil> how are things in chicago and fort collins?
15:03:54 <sjmc7> cold
15:04:00 <TravT> can't complain too much... just a bit cold
15:04:08 <TravT> nikhil, are you still in blacksburg?
15:04:13 <nikhil> rosmaita: oh man! every-time I read news my faint in humanity sinks :(
15:04:20 <nikhil> faith*
15:04:29 <nikhil> TravT: yeah
15:04:39 <TravT> yingjun: where do you live?
15:04:40 <nikhil> sjmc7: hope the snow situation is better
15:04:46 <nikhil> hope lakshmi is doing fine (with flood situation)
15:05:02 <yingjun> China, Changsha
15:05:29 <nikhil> hi yingjun, we've never met (even virtually)
15:05:43 <TravT> cool, we're really happy to have you join in the fun here!
15:05:45 <yingjun> HI, nikhil
15:05:54 <nikhil> fun indeed
15:06:11 <yingjun> :), nice to join
15:06:22 <nikhil> TravT: glad to hear, things are better there.
15:06:33 * nikhil shuts up
15:07:00 <TravT> allright, let's get on with some topics. we seem to be missing a couple faces, but that happens.
15:07:12 <TravT> #topic m1 is today
15:07:27 <TravT> So, there's this: https://openstack.nimeyo.com/66888/openstack-dev-release-preparing-your-mitaka-milestone-tag
15:08:05 <ekarlso> hola
15:08:09 <ekarlso> i'm lurking :P
15:08:11 <TravT> howdy ekarlso!
15:08:47 <TravT> i'll look at taking care of submitting a few things up
15:08:58 <TravT> but did have a couple questions
15:09:10 <TravT> are we aware of any release notes that we missed before landing reno?
15:09:30 <sjmc7> no, don’t think so
15:10:13 <TravT> possibly we should have had one for this patch
15:10:17 <TravT> https://review.openstack.org/#/c/245981/
15:10:37 <TravT> but i'm not thinking we need to stress over it.
15:10:49 <rosmaita> +1 for no stress
15:10:51 <sjmc7> :)
15:11:01 <nikhil> :)
15:11:35 <TravT> so then the next question is version...
15:11:55 <TravT> we had previously discussed whether to go to 1.0 with mitaka
15:11:59 <TravT> i think that is the desired goal
15:12:05 <TravT> but right now setup.cfg os 0.2
15:12:51 <sjmc7> i suppose that depends what we want to call 1.0
15:12:55 <TravT> in reading that email, it seems like we should propose 0.2b1 because that is in setup.cfg
15:13:10 <TravT> but that might mess things up if we jump to 1.0 at the end of mitaka
15:13:45 <rosmaita> TravT: i am glad you are PTL, i am having a hard time understanding the process
15:14:08 <nikhil> process is very dynamic
15:14:17 <TravT> lol, yeah, there have been quite a few changes this last bit in how the release team wants to do things
15:14:28 <nikhil> not sure if we need to invent a new word for openstack process
15:14:47 <rosmaita> i think i used a good word earlier
15:14:53 <TravT> :D
15:14:58 <nikhil> I may decide to use urban dictionary finally again after 3 years
15:15:11 <nikhil> rosmaita: you go for it then
15:15:26 <TravT> \o?
15:15:29 <nikhil> TravT: I think it should be fine with the .2b1
15:15:33 <TravT> that's my head scratching emoticon
15:15:46 <nikhil> nice!
15:15:56 <TravT> nikhil: yeah, i think so too.  we'll sort out 1.0 decision at the end of this release.
15:16:07 <nikhil> cool
15:16:13 <rosmaita> works for me
15:16:17 <nikhil> (was just about to enter) TravT: why do you anticipate it to be an issue?
15:16:29 <nikhil> but cool now
15:16:29 <TravT> for me, 1.0 is a signal that we'll start formally handling deprecations and adhere to stability of plugin interfaces
15:16:53 <nikhil> gotcha
15:16:59 <TravT> nikhil: because this process is new that doug put out (including reno) and nobody has experience with it.
15:17:25 <nikhil> makes sense
15:17:47 <TravT> ok, well, i'll work on the tag etc today and ping in room if needed later.
15:18:19 <TravT> so along these lines
15:18:32 <TravT> #topic searchlight client launchpad project
15:18:49 <TravT> yingjun: has been working on the client
15:18:55 <TravT> \o/
15:18:55 <sjmc7> i vote strongly not to have another launchpad project
15:19:01 <sjmc7> and thanks yingjun for all the hard work
15:19:04 <yingjun> +1
15:19:08 <TravT> i sent out an email http://lists.openstack.org/pipermail/openstack-dev/2015-November/080525.html
15:19:09 <nikhil> yingjun: ++
15:19:25 <rosmaita> yingjun: +++
15:19:30 <nikhil> sjmc7: haha, nice. I think that's fine to have everything consolidated
15:19:44 <nikhil> so, +1 from me too
15:20:25 <sjmc7> i can’t see us ever having a huge number of client-specific issues open
15:20:34 <nikhil> btw
15:20:43 <nikhil> TravT: do we plan to release client in M?
15:20:47 <TravT> i'm happy to not use it for now. my only question was if this would cause us issues with release managment since the client doesn't have to follow the same release schedule.
15:21:14 <TravT> nikhil: that would be the goal
15:21:18 <sjmc7> ah. that, i don’t know
15:21:37 <TravT> i added release onto that email, but only got crickets from them.
15:21:39 <nikhil> TravT: I think you can rely on reno for rel mgmt but for milestone tracking no
15:21:47 <TravT> i think they are dealing with changing the processes for every project right ow
15:22:09 <nikhil> TravT: you prolly want to be aware about https://review.openstack.org/#/c/226157
15:22:17 <nikhil> I brought that up in glance mtg
15:22:32 <nikhil> and it adds a new clause for client back compat
15:22:54 <TravT> can you summarize for us?
15:23:01 <nikhil> so, the client release timing would be essential
15:23:05 <nikhil> sure
15:23:19 <nikhil> it's mostly saying that all clients and libraries need to support back compat
15:23:29 <nikhil> well it mentions that you can deprecate it
15:23:34 <nikhil> but need to give that some time
15:23:42 <nikhil> the example suggests 4 cycles
15:23:54 <nikhil> but I don't anticipate if that would ever be true
15:24:04 <nikhil> so, when we release something
15:24:16 <nikhil> we want to ensure that we absolutely want that API to exist for a long time
15:24:36 <nikhil> if we need to have breaking changes the `process` will be quite painful
15:24:42 <nikhil> so, for SL client
15:24:48 <nikhil> if we need to have it in M
15:25:16 <nikhil> we prolly should test a beta and commit to client API (CLI stuff) in M official release
15:25:21 * nikhil done
15:25:49 <nikhil> sorry I missed an imp point -- that sort of ties in with the LP project
15:26:15 <nikhil> to determine the commitment that the project is giving to the API in a official project page (wherever that might be)
15:26:28 * nikhil done for reals
15:27:04 <TravT> thx nikhil
15:27:14 <TravT> did doug add reno to the python glanceclient?
15:27:39 <nikhil> no
15:27:47 <nikhil> and I don't think they expect
15:27:56 <nikhil> but some folks from IBM are requesting
15:28:20 <nikhil> and that is a discussion item for me with mr. flavio
15:28:55 <nikhil> that I haven't gotten to yet (with backlog list growing)
15:29:11 <TravT> ok, so nikhil, does everything you just said and your experience with glance client keep your vote at a +1 to not have a separate launchpad project for searchlight client?
15:29:29 <nikhil> TravT: nah, that bring me to 0
15:29:34 <nikhil> +1 for consolidation
15:29:35 <sjmc7> i’m tremendously confused :)
15:29:47 <nikhil> -1 for explicit communication
15:30:15 <nikhil> sjmc7: sorry, I realized a bit before that, I did that
15:30:26 <TravT> we might need the project just for somewhere to upload the releases of it
15:30:31 <nikhil> TravT: well, what I meant to say is
15:30:54 <david-lyle> I think it should be separate, to be contrarian
15:31:03 <TravT> o/ david-lyle
15:31:06 <nikhil> I gave some tradeoffs to you all and we can choose to prefer one way over other based on what we want to deal with later
15:31:50 <TravT> david-lyle: are you being contrarian for fun or is that your real opinion?
15:31:55 <david-lyle> the issues and releases are unique
15:32:18 <david-lyle> once people start piling on bugs and bps, it will get noisy
15:32:24 <david-lyle> and harder to track
15:32:54 <TravT> the separate releases is the thing that makes me lean towards having its own project
15:32:55 <david-lyle> for consumers and devs
15:33:32 <TravT> i think we need to record a vote here... but before doing, any more points to consider?
15:33:54 <sjmc7> yeah, i’m leaning that way as well now
15:34:02 <nikhil> I was told in a x-prj metg
15:34:10 <nikhil> that we are going to have some other tool for bugs
15:34:12 <nikhil> not LP
15:34:23 <nikhil> not sure how much has that been moved fwd
15:34:33 <david-lyle> nikhil: they've said that for 5 releases now
15:34:39 <david-lyle> I'll believe it...
15:34:45 <nikhil> :)
15:34:53 <TravT> i'm wondering how that'll help things at this point
15:35:05 <david-lyle> change for the sake of change
15:35:16 <nikhil> the opinion they had was projects can have LP for their own tracking purposes
15:35:38 <nikhil> but some <tool> for bugs, spec for BPs, and reno for rel
15:35:42 <nikhil> that was my impression
15:35:56 <nikhil> but as a prj we can choose to have LP
15:36:08 <nikhil> and like I said, there are tradeoffs to either
15:36:14 <nikhil> with my +0
15:36:31 <david-lyle> until the replacement is more than handwaving, launchpad is what we have
15:36:55 <TravT> i fail to see how specs (gerrit) on its own actually is good enough for BPs. (no assignees, priority, etc)
15:37:28 * TravT worries that the wheels are coming off the bus for process management tools
15:37:48 <david-lyle> we're on 3/4 now
15:38:07 <david-lyle> nothing mass confusion and chaos won't fix
15:38:09 <david-lyle> :(
15:38:16 <TravT> okay, to stay on track with current topic, i'm going to put out a irc vote.
15:38:38 <nikhil> :D
15:38:51 <TravT> #startvote (Should we use a separate LP project for python-searchlightclient)? yes, no
15:38:52 <openstack> Begin voting on: (Should we use a separate LP project for python-searchlightclient)? Valid vote options are yes, no.
15:38:53 <openstack> Vote using '#vote OPTION'. Only your last vote counts.
15:39:10 <david-lyle> #vote yes
15:39:30 <rosmaita> #vote yes
15:39:33 <nikhil> I am fine w/ either
15:39:36 <TravT> #vote yes
15:39:40 <sjmc7> #vote yes
15:39:47 <TravT> yingjun?
15:40:12 <yingjun> fine for either too
15:40:30 <TravT> ekarlso: you still lurking and care at all?
15:41:06 <TravT> #endvote
15:41:06 <openstack> Voted on "(Should we use a separate LP project for python-searchlightclient)?" Results are
15:41:07 <openstack> yes (4): TravT, david-lyle, rosmaita, sjmc7
15:41:31 <TravT> okay, so we'll keep it...
15:42:16 <yingjun> I’ll update the doc..
15:42:23 <TravT> thank you!
15:42:36 <TravT> #topic client command brainstorming
15:42:54 <TravT> yingjun: has a patch up with the first command on it
15:42:56 <TravT> https://review.openstack.org/#/c/249076/
15:43:36 <TravT> i added a comment about starting to sketch out the full suite of commands (so we have a basic roadmap of where we are going even if the first patch is only the first command)
15:43:55 <TravT> but i think we should spend some time brainstorming that out
15:44:42 <TravT> so, we don't have to do it here
15:44:45 <TravT> but here is an etherpad
15:44:46 <TravT> https://etherpad.openstack.org/p/searchlight-cli-brainstorming
15:45:46 <TravT> so, please take a look and add thoughts
15:45:47 <ekarlso> tarya :p
15:45:49 <ekarlso> TravT: ..
15:46:12 <ekarlso> TravT: just busy busy -,,-
15:46:18 <TravT> no worries
15:46:29 <TravT> #topic specs
15:46:31 <ekarlso> anyone started the cli yet that I can take a look at ?
15:46:53 <TravT> ekarlso: yingjun has started it
15:47:05 <sjmc7> https://review.openstack.org/#/c/247565/ ekarlso
15:47:17 <sjmc7> and https://review.openstack.org/#/c/249076/
15:47:33 <ekarlso> ok, i'll go look alter :)
15:47:47 <TravT> thank you!
15:47:52 <TravT> re: specs
15:48:06 <TravT> sjmc7 and I have been having some fun with two that are "essential" for mitaka
15:48:32 <TravT> they are admittedly kind of heavy to read
15:48:56 <TravT> but we do need to work through them and come to some conclusion so we can work them out
15:49:06 <TravT> and get implemented this release
15:49:42 <TravT> if you can take a look please do.  sjmc7 is actually going to be in fort collins next week, so we'll talk through them more in person next week
15:50:06 <TravT> Admin search fields unless already addresed above
15:50:06 <TravT> https://review.openstack.org/#/c/245357/
15:50:06 <TravT> Zero downtime re-indexing
15:50:08 <TravT> https://review.openstack.org/#/c/245222/
15:50:45 <TravT> sjmc7: anything specific to bring up today?
15:51:11 <sjmc7> about those? not really, except if anyone’s interested to look them over
15:51:34 <sjmc7> though i did also want to briefly mention the elasticsearch client version palava
15:51:58 <TravT> #topic elasticsearch client versioning
15:52:09 <sjmc7> that’s the one!
15:52:33 <TravT> mention away sjmc7
15:52:41 <sjmc7> apologies yingjun, i should’ve mentioned i’d already proposed a patch to openstack/requirements at https://review.openstack.org/#/c/252075/
15:52:59 <sjmc7> capping the client version at 1.9.0
15:53:21 <yingjun> sjmc7, no worries
15:53:25 <sjmc7> gordon chung seems ok with it
15:53:31 <sjmc7> says they’ve only tested with 1.7
15:53:48 <sjmc7> and i don’t think anyone else uses it
15:54:07 <sjmc7> i’ll pick some random infra people today to add as reviewers
15:54:18 <TravT> yeah, we need to get this through
15:54:44 <sjmc7> we’ll need to do a small amount of work to be 2.0-compatbile
15:54:59 <sjmc7> i’ll file a ticket
15:55:10 <TravT> because searchlight doesn't work in out of the box devstack at the moment because of the 2.0 client release.
15:55:28 <david-lyle> is the query syntax different between versions?
15:55:44 <sjmc7> david-lyle - there’s one thing that isn’t supported in 2.0, which is deleting by document type
15:55:56 <sjmc7> that’s the only thing i’ve found we do that doesn’t work as expected
15:56:18 <david-lyle> ok, that's not too bad
15:56:36 <sjmc7> and actually, a fresh devstack install should be ok unless i’ve missed something
15:56:45 <sjmc7> unless the data already exists
15:56:54 <sjmc7> but yes, needs fixing
15:57:32 <TravT> https://bugs.launchpad.net/searchlight/+bug/1515412
15:57:32 <openstack> Launchpad bug 1515412 in OpenStack Search (Searchlight) "python elasticsearch must match server version" [High,Triaged] - Assigned to Liyingjun (liyingjun)
15:57:58 <TravT> ok, well, let's get the version capped for now, unless yingjun / sjmc7 come up with a searchlight fix first.
15:58:13 <sjmc7> i think the version cap’s sensible anyway
15:58:21 <sjmc7> while devstack and most people will be using 1.x
15:58:27 <TravT> i think for now...
15:58:39 <TravT> but it can't stay that way.
15:58:58 <TravT> ok, thanks yingjun and sjmc7
15:59:04 <TravT> out of time.
15:59:13 <TravT> #topic parting words
15:59:17 <TravT> any last words?
15:59:42 <TravT> thanks everybody!
16:00:08 <TravT> #endmeeting