22:00:31 <mtreinish> #startmeeting qa
22:00:32 <openstack> Meeting started Thu Feb  5 22:00:31 2015 UTC and is due to finish in 60 minutes.  The chair is mtreinish. Information about MeetBot at http://wiki.debian.org/MeetBot.
22:00:33 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
22:00:35 <openstack> The meeting name has been set to 'qa'
22:00:42 <mtreinish> hi, who's here today?
22:00:46 <masayukig> \o
22:00:51 <mtreinish> #link https://wiki.openstack.org/wiki/Meetings/QATeamMeeting#Proposed_Agenda_for_February_5th_2015_.282200_UTC.29
22:00:52 <dkranz> \o
22:00:54 <gmann> hi
22:00:55 <mtreinish> ^^^ Today's agenda
22:00:58 <riwinters> hi
22:02:00 <mtreinish> dtroyer, jogo, sdague: around?
22:02:08 <mtreinish> ok, let's get started
22:02:21 <mtreinish> #topic Spec Reviews
22:02:52 <mtreinish> ok, before we discuss the one I put on the agenda, does anyone have any open specs they'd like to discuss?
22:03:00 <andreaf> o/
22:03:55 <andreaf> mtreinish: sorry just saying I'm here, no spec to discuss
22:04:03 <mtreinish> andreaf: heh I figured :)
22:04:05 <dtroyer> o/
22:04:22 <mtreinish> I did review this one earlier: https://review.openstack.org/144329 if someone else wanted to jump on it
22:04:44 <mtreinish> if there aren't any other specs to bring up, I guess we can talk about the one I put on the agenda
22:05:04 <mtreinish> #link https://review.openstack.org/152683
22:05:35 <mtreinish> so that's a rough draft of some ideas I had for how we'll change things in the QA program in the world under a big tent
22:06:02 <mtreinish> I'm not sure where the doc should live eventually, but I figured it would be better to start righting down my ideas to start a discussion
22:06:18 <mtreinish> I'm sure it'll come up at summit, but better to start early :)
22:06:32 <dkranz> mtreinish: I have a concern about the big tent thing here
22:07:01 <dkranz> mtreinish: It might help if non-core projects could contribute some integration tests to make sure we do not break them
22:07:15 <mtreinish> dkranz: in general or in my wip spec?
22:07:15 <dkranz> mtreinish: where "we" is the core projects you listed in the spec
22:07:27 <mtreinish> ah, ok
22:07:52 <dkranz> mtreinish: tempest could still contain integration tests for non-core projects
22:08:39 <mtreinish> dkranz: well that's the thing I was trying to outline, I don't really think that scales.
22:08:44 <dkranz> mtreinish: such tests don't have to be gating
22:10:06 <dkranz> mtreinish: not sure
22:10:14 <mtreinish> dkranz: then I'm not sure I understand, if they're not gating it doesn't prevent things breaking so why put them in tempest?
22:10:21 <andreaf> dkranz: you mean run them post-merge or?
22:10:36 <dkranz> andreaf: yes, post-merge. I not in tempest, where would they go?
22:10:44 <mtreinish> dkranz: there is also a cross-project spec from sdague on testing in general for openstack:
22:10:57 <mtreinish> #link https://review.openstack.org/150653
22:11:06 <dkranz> mtreinish: I had not seen that but will take a look.
22:11:19 <andreaf> dkranz, mtreinish: so tempest could contain tests that we run in the gate but also test that we run periodically like we discussed in Paris I believe
22:11:44 <dkranz> mtreinish: I am just concerned that we will go from trying to test a complete stack well, and failing, to not testing one at all
22:12:16 <mtreinish> andreaf: sure, but I still think we should define what kinds of tests belong in tempest.
22:12:30 <mtreinish> and make it easy for people to write things outside that scope in other places
22:12:33 <dkranz> mtreinish: I thought we said tempest was integration tests
22:12:47 <Rockyg> ++ dkranz
22:12:56 <dkranz> mtreinish: the question is which projects can have integration tests in tempest
22:13:12 <dkranz> mtreinish: and I think you are suggesting it is only the new small set of "core" projects
22:13:18 <dkranz> mtreinish: and I'm not sure that is a great idea
22:13:28 <mtreinish> dkranz: we did, but integration between what? When you have N^50 projects how can we keep on top of that
22:13:30 <Rockyg> Agree with dkranz
22:13:54 <mtreinish> dkranz: in my view thos projects are the base where most things are built on top of
22:14:19 <mtreinish> and if we ensure that's stable and works well it'll mean the greater ecosystem built on top of them will work better
22:14:20 <Rockyg> I think there is a need to define what/how periodic intergation jobs would be put together/run
22:14:39 <dkranz> mtreinish: for heat, yes, but things like ceilometer are not "on top"
22:15:19 <mtreinish> dkranz: I didn't put ceilometer on that list, because there are other issues associated with that. But I put that list up as a starting point
22:15:30 <mtreinish> I intended that list to be organic
22:15:37 <dkranz> mtreinish: ok, fair enough
22:16:24 <dkranz> mtreinish: I'm just suggesting that list could be larger than the co-gated list
22:16:38 <mtreinish> dkranz: anyway feel free to put comments in the spec I'm sure we won't reach a conclusion on this quickly :)
22:16:45 <dkranz> mtreinish: sure
22:17:12 <mtreinish> dkranz: yeah, and I'm not sure it should be more than that
22:17:38 <mtreinish> ok if there aren't any other specs to discuss, I guess we can move on
22:18:06 <Rockyg> Just a quick advert for the other review you mentioned:  https://review.openstack.org/144329
22:18:39 <Rockyg> We really need it to get finallized so we can get DefCore on Icehouse, Juno out before the summit
22:19:03 <andreaf> mtreinish, Rockyg: I'll review it tomorrow
22:19:22 <mtreinish> andreaf: ok, thanks
22:19:33 <mtreinish> with that I guess we'll move on
22:19:37 <mtreinish> #topic Blueprints
22:20:00 <mtreinish> ok so today was K-2, I think ttx just pushed the email out for the other projects
22:20:18 <mtreinish> I don't think we landed any bps for this milestone (not a big deal)
22:20:28 <mtreinish> but does anyone have any status updates on in progress bps?
22:20:44 <mtreinish> #link https://blueprints.launchpad.net/tempest/
22:20:44 <andreaf> I do
22:20:51 <mtreinish> andreaf: ok go ahead
22:21:10 <andreaf> #link https://blueprints.launchpad.net/tempest/+spec/resource-cleanup
22:21:59 <ekhugen> andreaf sorry I'm late to the meeting, but what's going on with the resource-cleanup blueprint
22:22:23 <andreaf> there are a lot of contribution to that, but we're hitting a race condition issue on the juno jobs
22:22:44 <andreaf> ekhugen, perhaps you have an update on that I just saw your patchset #27 or so passed the tests
22:22:55 <ekhugen> andreaf I think I'm narrowing it down, it's to do with the positioning of the credentials setup and client setup
22:23:09 <mtreinish> andreaf: a race on juno, is there a bug open for that? I can dig into it tomorrow
22:23:10 <ekhugen> they can't be in two separate methods
22:23:35 <ekhugen> mtreinish yes, there's a bug open from a long time ago
22:24:34 <mtreinish> ekhugen: do you have a link?
22:24:44 <Rockyg> Just wann point out this one is also high priority for Defcore.  No pressure here ;-)
22:24:49 <ekhugen> https://bugs.launchpad.net/tempest/+bug/1357055
22:24:58 <gmann> ekhugen: its strange, having those in separate methods should not be any issue?
22:26:00 <ekhugen> gmann andreaf actually my latest patchset failed again, so forget what I said about separate methods
22:26:51 <ekhugen> but it is something with credentials being in a separate method, because that was the part that caused it to fail
22:26:56 <mtreinish> ok, I'll try to dive into that one tomorrow
22:27:09 <mtreinish> I don't think we'll be able to sort it out in the meeting :)
22:27:49 <mtreinish> andreaf_: was there anything else on this bp? I'm assuming you've got another one to discuss too
22:27:49 <ekhugen> okay, thanks mtreinish
22:29:18 <mtreinish> ok does anyone else have a bp to discuss? (we might have lost andreaf_)
22:30:33 <mtreinish> ok then let's move on
22:30:44 <mtreinish> #topic Devstack
22:30:55 <mtreinish> dtroyer: any thing new in the world of devstack?
22:31:06 <mtreinish> besides the failed neutron as default switch? :)
22:31:30 <dtroyer> that's the big thing…we reverted default this morning back to nova-net
22:31:48 <dtroyer> but also getting a working linuxbridge config appears to be non-trivial
22:32:24 <mtreinish> dtroyer: oh, of that I have no doubt, I've never viewed setting anything up in neutron as trivial...
22:32:45 <dtroyer> in other news, I'm continuing to work on the venv solution, three prelim reviews up already
22:33:28 <mtreinish> dtroyer: cool do you have links?
22:33:54 <dtroyer> sdague is working away at helping folks set up for the external plugins, and a new pkg format to emulate what distro package managers do for things that are not distro packaged
22:34:39 <dtroyer> https://review.openstack.org/151811 and https://review.openstack.org/151812 remain…looks like the first needs a recheck
22:35:43 <mtreinish> dtroyer: ok thanks
22:35:53 <dtroyer> sean is also working on some docs regarding direction for DevStack (https://review.openstack.org/#/c/150789/)
22:36:08 <dtroyer> that's all I have
22:36:11 <mtreinish> #link https://review.openstack.org/#/c/150789/
22:36:23 <mtreinish> dtroyer: yeah I saw that, I haven't had a chance to comment on that doc yet
22:36:36 <mtreinish> dtroyer: what do you mean by new pkg format
22:37:20 <mtreinish> I'm not sure I follow, I get it's for installing non-python things which aren't in a distro. Is it just standarizing how we do that?
22:38:26 <dtroyer> basically yes.   an example would be for docker to get any installs downloaded and running during stack.sh.  This differs from the existing plugins partially by timing, but also in that these are not openstack projects (might be sourceforge though)
22:38:59 <dtroyer> there is some overlap between internal plugins, external plugins and the pkg stuff, we'll sort it as we go
22:39:15 <mtreinish> ah ok, would we also be able to levarage this to install newer versions of things than in the in distro pkgs?
22:39:41 <dtroyer> possibly.  I think the libvirt-from-tar bits might work here
22:39:57 <dtroyer> although that paricular example has other complexities
22:39:58 <mtreinish> ok, cool
22:40:10 <mtreinish> does anyone have anything else to discuss on devstack?
22:41:11 <mtreinish> ok then let's move on
22:41:18 <mtreinish> #topic tempest-lib
22:41:53 <mtreinish> so oomichi left a lot on the agenda about his work to do the service client migration
22:42:27 <mtreinish> he also has a spec up to document the approach where we're having a good discussion about it:
22:42:30 <mtreinish> #link https://review.openstack.org/#/c/150356/
22:43:07 <mtreinish> the other thing on this is I pushed out a new tempest-lib release yesterday because there was an issue where tempest was using an exception class which wasn't in a tempest-lib release
22:43:30 <mtreinish> http://git.openstack.org/cgit/openstack/tempest-lib/tag/?id=0.2.0
22:44:02 <mtreinish> that was all I had on tempest-lib
22:44:11 <mtreinish> does anyone have anything else to discuss on this?
22:44:35 <andreaf> just mention that I'm working on migrating the auth framework
22:44:49 <andreaf> I did most of the preparation work - under review
22:45:01 <mtreinish> andreaf: ah, yeah sry I forgot. Do you have a link?
22:45:08 <dkranz> mtreinish: just one question
22:45:14 <andreaf> once that's approved I'll add auth.py to tempest-lib
22:45:31 <dkranz> mtreinish: If we don't think any projects are going to use the tempest lib service clients, why bother to move them?
22:46:01 <mtreinish> dkranz: I don't think all the services are going to use them. Like I know nova has no interest in them at this point
22:46:02 <dkranz> mtreinish: I am pretty sure marun was planning to use them.
22:46:16 <mtreinish> dkranz: yeah he and heat are the only ones I know about at this point
22:46:28 <dkranz> mtreinish: so why not move them "on demand"?
22:47:17 <mtreinish> dkranz: we could do that (I think I brought that up in the spec review too :) ) My only concern is the consistency
22:47:21 <mtreinish> but we can work around that
22:48:07 <dkranz> mtreinish: Given how hard it will be to change something once it moves, it seems better to wait until some one actually wants to use it so they can change it before moving.
22:48:15 <mtreinish> dkranz: fwiw, I think having the service clients in the lib with a consistent interface will just be a good thing to do
22:48:36 <mtreinish> dkranz: the other aspect is there has been some discussion as using them as a base client layer for non testing purposes
22:48:41 <marun> mtreinish: nova is just behind :)
22:48:44 <mtreinish> but that's a secondary concern
22:48:46 <marun> mtreinish: they'll catch up eventually
22:49:36 <mtreinish> marun: heh
22:49:45 <mtreinish> ok is there anything else on this topic?
22:50:38 <mtreinish> #topic Bugs
22:50:44 <mtreinish> #link https://etherpad.openstack.org/p/Tempest-bug-report
22:50:50 <mtreinish> gmann: anything on the bug front?
22:51:30 * mtreinish just realized we should probably put devstack and grenade on that etherpad too
22:52:14 <gmann> mtreinish: nothing much. currenttly we have only 1 new bug which Salvatore is looking
22:52:20 <mtreinish> ok cool
22:52:58 <mtreinish> we probably should try to work through the 190 open bugs. If there are only 78 in progress that means there are a lot of triaged bugs without anything happening on
22:53:03 <andreaf> mtreinish: https://bugs.launchpad.net/tempest/+bug/1357055 is the bug I was mentioning before - related to cleanup of resources
22:53:27 <mtreinish> andreaf: cool thanks, ekhugen pasted it when your connection dropped :)
22:53:43 <gmann> mtreinish: there might be lot of invalid case as lot is changed in tempest (xml tests etc)
22:54:10 <mtreinish> gmann: ok, we should probably schedule a time to go through that and do a purge then
22:54:19 <mtreinish> but there isn't a huge rush
22:54:37 <mtreinish> anyway ~5min left, so let's move
22:54:46 <gmann> mtreinish: in triage rotation we can start looking in those when there is not much to triage?
22:54:59 <mtreinish> gmann: sure that works
22:55:04 <mtreinish> #topic Critical Reviews
22:55:21 <mtreinish> ok does anyone have any reviews they'd like to get extra eyes on
22:55:29 <mtreinish> #link https://review.openstack.org/#/c/137239/
22:55:35 <mtreinish> ^^^ That one was in the agenda
22:55:45 <masayukig> mtreinish: yeah, I've wrote that.
22:56:05 <masayukig> This patch has several things such as "Support for IPv6", refactoring, fixing comments etc.. So I'd like to know others opinion.
22:57:05 <andreaf> I also have one
22:57:11 <andreaf> #link https://review.openstack.org/#/c/151206/
22:57:20 <mtreinish> masayukig: hmm it does seem like a hodgepodge in the commit message
22:57:29 <dkranz> mtreinish: That patch does have the test renaming issue
22:57:43 <dkranz> mtreinish: we need to decide if we are going to allow that before unique ids are in
22:57:54 <dkranz> mtreinish: I mean masayukig's patch
22:58:35 <masayukig> mtreinish: yes.
22:59:10 <mtreinish> dkranz: does it? But that's a good point if it does
22:59:55 <mtreinish> anyway we're at time, we can discuss this in -qa
22:59:59 <mtreinish> thanks everyone
23:00:07 <mtreinish> #endmeeting