22:00:46 <mtreinish> #startmeeting qa
22:00:46 <openstack> Meeting started Thu Mar 19 22:00:46 2015 UTC and is due to finish in 60 minutes.  The chair is mtreinish. Information about MeetBot at http://wiki.debian.org/MeetBot.
22:00:47 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
22:00:50 <openstack> The meeting name has been set to 'qa'
22:01:01 <mtreinish> hi who's here today?
22:01:05 <masayukig> \o
22:01:06 <ajmiller> hi
22:01:08 <oomichi> hi
22:01:11 <gmann> hi
22:01:22 <mtreinish> #link https://wiki.openstack.org/wiki/Meetings/QATeamMeeting#Proposed_Agenda_for_March_19th_2015_.282200_UTC.29
22:01:23 <dtroyer> o/
22:01:24 <dkranz> o/
22:01:26 <mtreinish> ^^^ today's agenda
22:02:15 <mtreinish> jogo, andreaf: courtesy ping
22:02:19 <mtreinish> ok let's get started
22:02:24 <mtreinish> #topic Specs Reviews
22:02:33 <andreaf> o/
22:02:38 <jogo> o/
22:02:46 <mtreinish> are there any open spec reviews to discuss today?
22:03:17 <mtreinish> dtroyer: I did want to ask there are a couple of open devstack specs (1 of which has been implemented)
22:03:22 <mtreinish> what did you want to do with them?
22:04:06 <dtroyer> I need to go through them and get caught up
22:04:20 <mtreinish> dtroyer: ok, no worries. I was just curious
22:05:18 <mtreinish> ok is there anything else on specs? otherwise we can move on
22:06:21 <mtreinish> #topic Blueprints
22:06:52 <mtreinish> so today was k-3
22:07:17 <mtreinish> I have been meaning to go through and update the targets for a number of open bps
22:07:24 <mtreinish> I'll try to take care of that soon
22:07:34 <mtreinish> does anyone have an open bp to discuss?
22:08:33 <andreaf> mtreinish: nothing special only quickly updates
22:09:21 <mtreinish> andreaf: well the auth migration is a big step for tempest-lib
22:09:22 <andreaf> mtreinish: resource-cleanup is still close to be completed, there's only a few patches and a little bit of reorganizing credentials code in base classes that I wanted to do
22:10:40 <andreaf> mtreinish: yes indeed that's almost done now
22:10:59 <andreaf> mtreinish: the patch to move auth with unit tests in ready for review
22:11:22 <andreaf> actually it's going through gate now
22:11:44 <mtreinish> andreaf: yeah masayukig +A'd it
22:11:46 <andreaf> and I'm working on the patch to remove auth from tempest - pending tempest-lib release
22:13:05 <mtreinish> gmann: on your schema refactor bp, is that done? I lost track of the patches at some point
22:13:12 <andreaf> on multi-auth bp, almost finished, the main patch outstanding is https://review.openstack.org/#/c/133029/, I need to incorporate jamielennox comments in there
22:13:50 <mtreinish> andreaf: cool, that's been a longstanding one, it'll be good to have keystone v3 only gate runs
22:13:56 <gmann> mtreinish: all patches are  ready for review
22:14:12 <mtreinish> gmann: do you have a link?
22:14:49 <gmann> #link https://review.openstack.org/#/q/status:open+project:openstack/tempest+branch:master+topic:bp/rearrange-nova-response-schemas,n,z
22:15:01 <mtreinish> thanks
22:15:19 <mtreinish> ok, are there any other bps to discuss today?
22:15:39 <andreaf> mtreinish: well having a full v3 job is a separate bp, there's some work to be done to get a job running with v2 turned off...
22:15:54 <andreaf> mtreinish: I just noticed javelin uses v2 only for instance
22:16:14 <andreaf> mtreinish: but that's kind of out of the original scope of the bp
22:16:15 <mtreinish> andreaf: oh, my mistake. the multi-auth bp is just to make everything support v3
22:17:46 <mtreinish> ok, let's move on
22:17:51 <mtreinish> #topic Devstack
22:18:24 <mtreinish> dtroyer: anything exciting to discuss this week?
22:18:42 <dtroyer> we spent a bit of time earlier this week looking at speeding up some parts of devstack
22:19:33 <dtroyer> I put up a demo of batching OSC commands in keystone that saved a little bit
22:20:05 <mtreinish> dtroyer: the big time sink is still the wheel compilation though right?
22:20:45 <dtroyer> yeah, that dwarfs the seconds I managed to recover
22:21:14 <dtroyer> I've wanted to do the batching anyway though as an example of how to do that
22:21:36 <dtroyer> I did spend some time trying to run OSC in a daemon-like mode so it doesn't always have to be reloaded but we can still get looked-up values.
22:21:55 <dtroyer> again, that's useful but not worth the small return here just yet
22:22:07 <mtreinish> dtroyer: sure, but it's still cool
22:22:27 * jogo would like to see OSC faster itself
22:22:37 * jogo would like a magical unicorn too
22:22:54 * dtroyer would like to use a compiled language with real binaries
22:23:16 <mtreinish> dtroyer: for the wheel compilation do you think it would be worth just turning that off if there aren't any venvs enabled?
22:23:30 <jogo> greghaynes: is looking into the wheel stuff
22:23:54 <jogo> mtreinish: probably
22:24:15 <mtreinish> jogo: just in the mean time I'm not sure we need to build the wheels if we're not using them
22:24:33 <dtroyer> possibly, but we've already started pulling out python distro packages and that build time will still happen
22:24:54 <mtreinish> dtroyer: oh, right I forgot about that
22:25:16 <dtroyer> so some of that rests on what sort of python package policy we want to use, we've been back and forth a couple of times in 3 1/2 years
22:25:41 <jogo> mtreinish: I'll look into if it makes sense to turn off for now or not
22:26:11 <dtroyer> speaking of venvs, I refreshed the venv patches and added a Grenade tweak to handle getting the database modules installed by pip on the target side
22:26:50 <greghaynes> jogo: Hey, yep. I have patches up that are waiting for infra reviews to get mirrors setup
22:27:04 <jogo> greghaynes: \o/
22:27:39 <greghaynes> Things have been exploding today so havent been bugging them too hard ;)
22:27:56 <mtreinish> dtroyer: cool, grenade venvs could have some interesting implications for doing upgrades with decoupled versions
22:29:18 <mtreinish> ok, is there anything else to discuss on devstack?
22:29:35 <dtroyer> not here
22:30:27 <mtreinish> ok, then let's move on
22:30:39 <mtreinish> #topic Grenade
22:30:49 <mtreinish> ajmiller: you put a topic on the agenda
22:30:56 <ajmiller> Yes, thanks.
22:31:45 <ajmiller> The issue is that in neutron-lbaas we implemented a devstack plugin, and that works fine, but the final patch to remove the lbaas-specific code from devstack broke grenade when it merged.
22:31:57 <ajmiller> So we reverted it
22:32:36 <ajmiller> Then I messed around for a while hacking in grenade, but that seems to be a loosing proposition without doing more comprehensive work in grenade to fully support the plugins.
22:33:16 <ajmiller> I'm looking for opinions, guidance, whatever.  Also, we will be doing a similar plugin for octavia (currently a stackforge project)
22:33:34 <dtroyer> because grenade doesn't run stack.sh on the target side, all that has to be done 'by hand' there
22:33:48 <ajmiller> #link https://review.openstack.org/#/c/161436/
22:34:06 <mtreinish> dtroyer: yeah, you should be able to hack it in that way
22:34:21 <dtroyer> I think this is an issue mostly because it's a service that is on the base side and now not present on the target side.
22:34:41 <dtroyer> or present in a different way
22:34:51 <mtreinish> ajmiller: sdague did make a ML post about defining a plugin interface for grenade, but I don't think anyone has had bandwidth to start on that
22:36:00 <ajmiller> Yes, saw the post.  I don't think its a disaster for us if this patch doesn't make it into kilo, it might be a bit of a maintenance hassle, but the basic functinality works.
22:36:19 <ajmiller> Although I plan on doing more testing.
22:36:52 <mtreinish> ajmiller: we definitely will need to have a concrete story around grenade plugins before too long, especially since we're moving for more things to go that direction
22:37:11 <mtreinish> ajmiller: I just think right now no one has had the bandwidth to really take a bite out of it
22:37:31 <jogo> mtreinish: maybe that is something we can flesh out next week?
22:37:43 <jogo> mtreinish: a concrete plan to follow for when we have time
22:38:06 <ajmiller> mtreinish: Great, understood.  It would be great if we can get a plan started.
22:38:08 <mtreinish> jogo: yeah it was on my original work items list, but I can't remember why I took it off
22:38:46 <mtreinish> jogo: but, yeah we'll definitely talk about it, especially since all 4 grenade cores will be in a room
22:39:21 <jogo> cool
22:39:38 <ajmiller> I'll be happy to consult, help, whatever I can, since I've done 1 devstack plugin and will probably be doing another fairly soon.
22:39:42 <mtreinish> ajmiller: sorry I don't have anything more concrete. I'll try to keep you in the loop whenever we start the discussion
22:40:12 <mtreinish> ajmiller: I think whatever we end up with will probably turn into a qa-spec before we start implementing it
22:40:50 <ajmiller> mtreinish: That sounds like a good plan.
22:41:40 <mtreinish> ok, is there anything else to discuss on grenade?
22:43:01 <mtreinish> #topic Bugs
22:43:11 <mtreinish> #link https://etherpad.openstack.org/p/Tempest-bug-report
22:43:21 <mtreinish> gmann: anything new on bugs this week?
22:44:30 <gmann> mtreinish: not much on this
22:44:44 <gmann> actually i did not get chance to look into bugs this week
22:44:51 <gmann> rohan had  triage rotation this week and poornima next
22:45:01 <mtreinish> gmann: ok, no worries
22:45:16 <mtreinish> does anyone else have anything to discuss about bugs?
22:46:09 <mtreinish> #topic Critical Reviews
22:46:31 <mtreinish> ok, does anyone have any open reviews they'd like to get some extra eyes on?
22:46:54 <andreaf> mtreinish: I'm still incorporating jamielennox feedback, but it should soon be ready: https://review.openstack.org/#/c/133029/
22:47:10 <mtreinish> #link https://review.openstack.org/#/c/133029/
22:47:44 <mtreinish> andreaf: ok sure
22:48:34 <dkranz> https://review.openstack.org/#/c/162721/
22:49:01 <dkranz> https://review.openstack.org/#/c/164225/
22:49:05 <dkranz> Some non-admin stuff
22:49:41 <dkranz> masayukig: can you please remove your -2 from https://review.openstack.org/#/c/164300/ ?
22:50:07 <andreaf> dkranz: +A
22:50:09 <masayukig> dkranz: sure
22:50:21 <dkranz> andreaf, masayukig : thanks!
22:50:42 <andreaf> dkranz: I had reviewed that before, not sure why I didn't vote
22:50:43 <mtreinish> dtroyer: oh interesting on that first one we were generating admin creds everywhere for no reason
22:50:45 <masayukig> dkranz: done
22:51:14 <dkranz> mtreinish: yeah. As soon as andreaf gives the go-ahead that he is done, I will finish the non-admin cleanup
22:52:02 <mtreinish> ok are there any other reviews to bring up?
22:53:13 <mtreinish> then let's open the floor for 8 mins
22:53:16 <mtreinish> #topic Open Discussion
22:53:35 <mtreinish> does anyone have any other topic to discuss?
22:54:02 <dkranz> mtreinish: is there going to be any opportunity for remote participation next week?
22:55:22 <mtreinish> dkranz: unfortunately I don't think so. We're just going to be in a regular conference room, I don't have a good microphone or anything
22:55:43 <mtreinish> dkranz: but I'm sure we'll all be on irc if we need to reach out for anything
22:56:06 <dkranz> mtreinish: my thought was that I could actually hack on something. Just need to know what the plan ends up being
22:57:47 <mtreinish> dkranz: sure, we can definitely try to keep you in the loop if you want to work in parallel, that shouldn't be an issue. I thought you were asking about some kind of active telepresence (or an open audio feed)
22:58:07 <dkranz> mtreinish: nope
22:58:16 <dkranz> mtreinish: that's fine
22:58:49 <mtreinish> dkranz: fwiw, I'll be creating an etherpad for tracking the work (probably tomorrow)
22:58:59 <dkranz> mtreinish: great
22:59:54 <mtreinish> oh, and in the last min. just a heads up now that we're at k3 it's probably time to start thinking about design summit
23:00:09 <mtreinish> so I'm planning on pushing out a planning etherpad next week to brainstorm ideas
23:00:20 <mtreinish> I think the approach we took for paris worked out pretty well
23:00:40 <mtreinish> anyway we're at time, thanks everyone
23:00:42 <mtreinish> #endmeeting