17:00:20 <oomichi> #startmeeting qa
17:00:20 <openstack> Meeting started Thu May 26 17:00:20 2016 UTC and is due to finish in 60 minutes.  The chair is oomichi. Information about MeetBot at http://wiki.debian.org/MeetBot.
17:00:21 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
17:00:23 <openstack> The meeting name has been set to 'qa'
17:00:25 <hogepodge> o/
17:00:31 <oomichi> hi
17:00:36 <dmellado> hi \o
17:00:36 <jordanP> hi
17:00:38 <fnaval> o/
17:00:44 <mtreinish> o/
17:00:45 <oomichi> who's here today?
17:01:06 <oomichi> ah, so fast "hi" from many people :-)
17:01:11 <jordanP> hehe
17:01:21 <oomichi> #link https://wiki.openstack.org/wiki/Meetings/QATeamMeeting#Agenda_for_May_26th_2016_.281700_UTC.29
17:01:28 <oomichi> ^^^ today agenda
17:01:50 <oomichi> ok, lets start the meeting
17:02:02 <oomichi> #topic Specs Reviews
17:02:11 <andreaf> o/
17:02:18 <oomichi> #link https://review.openstack.org/#/q/status:open+project:openstack/qa-specs,n,z
17:02:52 <oomichi> so there are specs which contain red mark
17:02:57 <oomichi> meaning -1
17:03:12 <oomichi> does anyone want to discuss an open spec review?
17:03:23 <vishnoianil> #info vishnoianil
17:03:30 <mtreinish> oomichi: I think probably should abandon specs > 1yr old
17:03:34 <oomichi> I will update tempest-resource spec later anyways
17:03:35 <mtreinish> if they have a -1
17:03:45 <mtreinish> or a -2
17:03:55 <oomichi> mtreinish: +1 for doing that as cleanup
17:04:00 <mtreinish> looking at the list about 1/2 look very stale
17:04:01 <dmellado> oomichi: yes, there are some from 2014
17:04:06 <oomichi> at least 2 years ago
17:04:11 <mtreinish> dmellado: heh
17:04:32 <jordanP> yep abandoning those would be great
17:04:38 <oomichi> ok, I will do that later with some comments
17:04:58 <oomichi> #action oomichi will drop very old specs which not updated 1 years
17:05:14 * dmellado is guilty of one, but can't abandon it due to gerrit issues, though
17:05:41 <oomichi> are there any discussion here?
17:05:54 <mtreinish> I -1'd https://review.openstack.org/#/c/314704/ this morning mostly because I think it's too soon to adopt that project
17:06:01 <mtreinish> but getting other opinions on that would be good too
17:07:09 <oomichi> mtreinish: are you owner of this spec now?
17:07:24 <mtreinish> oomichi: no, I just rebased it after I reorganized the dir structure a little
17:08:32 <oomichi> mtreinish: this test seems useful for deployers, but it seems different from the gate(devstack)
17:08:51 <jordanP> yep I kinda agree. Not because it's too soon but because it's not typical enough, imo. Most OS clouds run with "default" policy and it could add complexity to support those use cases.  And they have the option of running tempest with users that have the permitted/expected permissions
17:09:03 <mtreinish> oomichi: right it's an extra project for testing custom policy configurations in real deployments
17:09:07 <oomichi> mtreinish: we are using default policy
17:09:17 <jordanP> yeah, it feels like it doesn't really belong to tempest
17:09:39 <mtreinish> jordanP: it's not for tempest, he's asking it be adopted as a seperate qa project
17:09:46 <jordanP> ahh by bad
17:10:02 <oomichi> mtreinish: interesting :-)
17:10:22 <oomichi> mtreinish: will you use tempest from the separated project like refstack?
17:10:31 <mtreinish> oomichi: no, it's a tempest plugin
17:10:47 <dmellado> mtreinish: but from the gate perspective, would this make sense?
17:10:50 <oomichi> mtreinish: ah, I see
17:11:02 <dmellado> I mean, would you change the default policy there?
17:11:26 <mtreinish> dmellado: no, not at all. We test with the defaults in the gate I don't expect that to change. This would just be an extra thing for people who want to test their custom things
17:11:59 <dmellado> got it
17:12:33 <oomichi> mtreinish: that means testers need to pass their own policy to the test project before testing?
17:12:34 <mtreinish> dmellado: although for this projects own gate testing, maybe it would do some weird custom setups :)
17:13:19 <oomichi> mtreinish: oh, that would be huge custom setup
17:13:22 <dmellado> ;)
17:13:31 <dmellado> that'd be interesting to see
17:13:38 <mtreinish> oomichi: no, see: https://github.com/rlrossiter/cinnamon-role/blob/master/README.md#how-to-use it uses a modified preprov creds to indicate what creds to use on which tests
17:15:29 <oomichi> mtreinish: that seems testers need to write the test code?
17:15:44 <oomichi> I am seeing the sample code
17:16:23 <mtreinish> oomichi: you can write your own test code, or reuse existing tests
17:16:33 <oomichi> mtreinish: the existing tests
17:17:00 <fnaval> this is interesting as it can be used with role based access type tests
17:17:31 <oomichi> mtreinish: yeah, nice to reuse tempest tests with small costom code
17:17:53 <mtreinish> personally I think this is a useful util, because custom policy is something I've hit on every production cloud I've encountered so I'm sure they'd like a simple mechanism to test it works
17:18:41 <oomichi> mtreinish: yeah, I agree. policy tends to be customized on each cloud
17:19:01 <oomichi> then original tempest is difficult to be passed on the default
17:19:31 <oomichi> that is a nice project for them
17:20:06 <oomichi> are there any topics about specs?
17:20:24 <oomichi> lets move to the next topic
17:20:31 <hogepodge> mtreinish: oomichi: trying to wrap my head around that topic from an interop perspective
17:20:37 <hogepodge> will think about it and comment later
17:20:45 <mtreinish> hogepodge: oh custom policy is the worst for interop
17:21:05 <mtreinish> hogepodge: it makes it next to impossible (you can just plain turn apis off)
17:21:10 <oomichi> #topic Tempest
17:21:11 <mtreinish> but it's something that exists and people use it
17:21:30 <mtreinish> and this plugin wouldn't be something you could use for defcore anyway
17:21:48 <oomichi> yeah, it will be a nice project for interop
17:21:50 <hogepodge> mtreinish: yup, my opinion is if you modify default policy you aren't openstack any more
17:22:01 <jordanP> that's bold
17:22:23 <hogepodge> (From a trademark and interop perspective... but let's move on)
17:22:52 <oomichi> there are many dev items as the agenda
17:23:07 <oomichi> #link https://wiki.openstack.org/wiki/Meetings/QATeamMeeting#Agenda_for_May_26th_2016_.281700_UTC.29
17:23:31 <oomichi> Id like to check the status
17:23:48 <oomichi> andreaf: the doc cleanpup is already done?
17:24:15 <oomichi> tempest cli seems good progress
17:24:23 <andreaf> oomichi: i've done a couple of patches but there's more to be done possibly
17:24:41 <oomichi> andreaf: ok, I got it. thanks for doing that
17:24:57 <oomichi> dmellado: how about test class hierarchy?
17:24:58 <mtreinish> oomichi: things on the cli are moving:
17:25:01 <mtreinish> #link https://review.openstack.org/#/q/topic:bp/tempest-run-cmd
17:25:06 <mtreinish> that starts the basics of the run command
17:25:13 <mtreinish> it'll be expanded from there
17:25:16 <dmellado> oomichi: we drafted a list of helpers to decide upon
17:25:21 <dmellado> https://docs.google.com/spreadsheets/d/1Y241ELbiM0djdFeII8mAiL8ZEQZwpBQlsV_2xbJUMMQ/edit?pref=2&pli=1#gid=0
17:25:31 <dmellado> and there's a patch around
17:25:37 <oomichi> mtreinish: yeah, good progress
17:25:37 <mtreinish> although those patches are blocked on me pushing an os-testr release (which I'm gonna do today)
17:25:39 <dmellado> https://review.openstack.org/#/c/320495/2
17:25:58 <dmellado> but I haven't had time to do upstream due to internal work, almost since Austin, so I'd love to get some more folks on this
17:26:24 <mtreinish> dmellado: you probably can pull horizon off that list, david-lyle has been working on ripping that out into a plugin
17:26:25 <oomichi> dmellado: thank you also
17:26:38 <oomichi> dmellado: we need more patch for removing neutron wrappers?
17:26:46 <dmellado> oomichi: I was thinking about maybe some of the Intel folks to help
17:26:59 <dmellado> oomichi: will check the list and update it, will ping you about that
17:27:15 <oomichi> dmellado: cool :-)
17:27:19 <dmellado> mtreinish: thanks for the info
17:27:33 <andreaf> oomichi: OO wrappers are a separate topic, there is a patch WIP on that #link https://review.openstack.org/#/c/320495/
17:27:39 <mtreinish> dmellado: http://git.openstack.org/cgit/openstack/tempest-horizon/
17:27:59 <dmellado> mtreinish: ack, I'll pull them off ;)
17:28:40 <andreaf> dmellado, mtreinish: done
17:28:50 <dmellado> about the tempest plugins too, should we try to enforce the projects to move them to split repos?
17:29:01 <dmellado> I see a ton of benefits from that
17:29:12 <dmellado> but this was offtopic in here, so sorry ;)
17:29:25 <oomichi> dmellado: yeah, that is a nice recommendation for using it as branchless
17:29:33 <dmellado> oomichi: exactly
17:29:48 <dmellado> if we tie tempest tests to a branched project, they wouldn't be branchless anymore
17:29:49 <mtreinish> dmellado: we already strongly recommend it. We're not really in a position to force it
17:30:16 <mtreinish> the way the plugin interface is constructed there really isn't a mechanism to enforce only seperate repos
17:30:22 <oomichi> mtreinish: yeah, right. they can implement the own tests as they like
17:31:16 <jordanP> we should enforce separate repos imo
17:31:21 <jordanP> *shouldnt
17:31:31 <dmellado> mtreinish: I'm aware on that we can't enforce it, as of now,and I was wondering if there's any kind of track
17:31:43 <dmellado> about the number of in-tree vs repo plugin projects
17:32:14 <dmellado> jordanP: in any case it's not like if we can ;)
17:32:28 <jordanP> yeh
17:32:55 <andreaf> dmellado: in tree is the vast majority atm
17:33:11 <dmellado> andreaf: ack, thanks!
17:33:15 <mtreinish> dmellado: http://codesearch.openstack.org/?q=tempest.test_plugins&i=nope&files=setup.cfg&repos=
17:33:37 <mtreinish> that's every tempest plugin in openstack projects
17:34:13 <mtreinish> there are tempest plugins outside of openstack too
17:34:29 <oomichi> (sorry, my pc was freezed just before..)
17:34:40 <fnaval> i'm currently have an open review for a tempest plugin as well for neutron-lbaas
17:34:41 <dmellado> mtreinish: interesting, but I'm fine with that data, thanks too!
17:34:58 <andreaf> mtreinish: I've seen that code search before
17:35:04 <hogepodge> mtreinish: well, that list includes every plugin but isn't all plugins? I don't think defcore has actually written one :-D
17:35:12 <dmellado> fnaval: out of curiosity, did you get ping on that by acruz?
17:35:33 <fnaval> dmellado: it was by a few folks from that channel
17:35:50 <dmellado> fnaval: would you mind putting the link on the commit?
17:35:55 <hogepodge> (oh nevermind, it's just in the search list) sry
17:36:07 <mtreinish> hogepodge: it shows every project on gerrit that has a plugin merged
17:36:17 <fnaval> amuller and emeilino i think: https://review.openstack.org/#/c/321087/
17:36:31 <oomichi> fnaval: are you working for cleaning neutron-lbaas now?
17:36:39 <fnaval> i still have to make modifications on post_test_hook.sh
17:36:56 <oomichi> fnaval: I guess the project has a lot of duplicated code from tempest
17:37:04 <fnaval> oomichi: i'm working on neutron-lbaas testing efforts mostly
17:37:15 <fnaval> oomichi: oh, that was cleaned up awhile back
17:37:24 <oomichi> fnaval: great :)
17:37:29 <fnaval> with the tempest to tempest-lib back to tempest thing
17:37:37 <fnaval> =)
17:37:52 <fnaval> i became quite well versed with tempest after that ordeal
17:38:22 <fnaval> not sure if we're following the list on the meeting wiki: but I wanted to bring up the name filitering with auth feature
17:38:33 <dmellado> fnaval: let me know if I can help you on that
17:38:58 <fnaval> dmellado: the setting up the tempestplugin?  sure, i'll take note
17:39:01 <oomichi> fnaval: yeah, please go ahead
17:39:16 <oomichi> fnaval: maybe this: https://review.openstack.org/#/c/284273
17:39:22 <fnaval> so, i've made significant progress with it and thanks for everyone's comment
17:39:58 <fnaval> im unsure of the protocol of picking up tasks so I basically want to make sure that it is a feature that is wanted in tempest
17:40:25 <fnaval> i fear that it might not get in because it wasn't 'approved' by the cores or ptls
17:40:53 <fnaval> i'm still getting used to the process so please forgive any errors that I may make along the way
17:40:58 <oomichi> hogepodge: are you done about interopthing on the agenda today?
17:41:12 <fnaval> #link https://review.openstack.org/#/c/284273
17:41:20 <fnaval> so it should be ready for review ^^
17:41:32 <hogepodge> oomichi: just refactoring or moving the tests in that list I dropped in the agenda
17:41:33 <mtreinish> fnaval: I think it's probably a valid patch, it's just going through the review process can take some time
17:41:46 <fnaval> also not sure if i should have brought it up in this section of the meeting or under the critical reviews part
17:41:48 <mtreinish> fnaval: I -1'd that this morning there is a backwards incompat change
17:42:06 <hogepodge> that require admin access. So they need to move, or be refactored where possible to not require admin. I hope to push the initial patches soon
17:42:21 <fnaval> mtreinish: cool - i'll take a look and ping you if I have any issues
17:44:02 <mtreinish> hogepodge: ok cool, I'll keep an eye out for it
17:44:20 <mtreinish> ok, oomichi is having network issues. I think we should move on
17:44:29 <jordanP> yes, 15min left :)
17:44:36 <mtreinish> although I don't think anyone can change topics but chair
17:44:49 <mtreinish> #topic Devstack + Grenade
17:45:08 <mtreinish> well does anyone have anything to talk about devstack or grenade this week?
17:45:22 <mtreinish> (even if the topic didn't update)
17:46:24 <mtreinish> ok. I'll take that as a no
17:46:31 <mtreinish> then let's move on to the next topic
17:46:43 <mtreinish> #topic OpenStack-Health
17:47:04 * oomichi back from networking issue.. :-(
17:47:04 <mtreinish> the only update I have on this for today is we're now also polling elasticsearch
17:47:26 <mtreinish> it uses elastic-recheck to search for known bugs on failed runs and report that on the ui
17:47:37 <mtreinish> the only place doing that is the table on the home page right now
17:47:43 <mtreinish> but it'll be expanded in the future
17:48:37 <mtreinish> that was all I had on openstack-health
17:48:42 <mtreinish> is there anything else on this topic?
17:50:08 <mtreinish> ok, let's move on then
17:50:15 <mtreinish> #topic Critical Reviews
17:50:27 <mtreinish> oomichi: ^^^ you need to do that to make it official, since you're the only chair
17:50:45 <jordanP> not a critical review but a general reminder about review process: we should be hard/though on code duplication. If someone tries to add a new utility method like _delete_volume or _create_volume_type, chances are that they are reinvinting the wheel
17:51:00 <mtreinish> does anyone have any reviews they'd like to get extra eyes on?
17:51:09 <oomichi> #topic Critical Reviews
17:51:15 <mtreinish> jordanP: yeah, has there been more proliferation of that kinda thing?
17:51:18 * oomichi back again..
17:51:28 <andreaf> #link https://review.openstack.org/#/c/226297/
17:51:43 <jordanP> I can't really say but it's bugs me a lot when I see those kind of things
17:52:02 <fnaval> #link https://review.openstack.org/#/c/284273
17:52:14 <jordanP> it usualy never pass but we have to stay alert
17:52:55 <jordanP> fnaval, what about it ?
17:53:00 <jordanP> mtreinish, left a valid comment imo
17:53:10 <fnaval> although it has a -1 at the moment, any other comments would be great
17:53:19 <andreaf> jordanP: yes this goes along the effort of refactoring and consolidating all the helpers around in scenario as well
17:53:25 <jordanP> I don"t like it because I like to group args together, but he's right
17:53:32 <fnaval> something that mtreinish may  not have picked up
17:54:21 <oomichi> jordanP: yeah, and sometime this kind of thing tend to become the same as corresponding project implementation
17:54:33 <dmellado> jordanP: +1 about that and that's something we'll try to tackle at the refactor too
17:54:37 <oomichi> then tempest code is different between projects tests
17:54:45 <oomichi> sometimes
17:55:50 <fnaval> jordanP: I don't disagree on it; but it's something that I don't fully understand and need to find out why
17:56:28 <jordanP> it's python, kwargs can also be filed "positionnaly"
17:56:36 <jordanP> (sorry it's hard to explain)
17:57:45 <fnaval> ah ok. if the kwargs being passed in are positional vs keyword.
17:57:50 <fnaval> then it breaks.
17:58:14 <jordanP> yeah, have a look at this code:
17:58:15 <jordanP> In [1]: def a(b,c):
17:58:15 <jordanP> ...:     print b,c
17:58:15 <jordanP> ...:
17:58:15 <jordanP> In [2]: a(c=1, b=2)
17:58:15 <jordanP> 2 1
17:58:45 <oomichi> jordanP: nice explanation :-)
17:58:46 <jordanP> (though it's not exactly related to your patch  but informative too)
17:58:52 <andreaf> oomichi: I had one last topic on the agenda that we did not cover, but it's a bit late, maybe next week  - about name for service client methods
17:59:20 <oomichi> yeah, nice to talk about it after the meeting
17:59:35 <oomichi> sorry for my today leading
17:59:45 <oomichi> the time comes
17:59:55 <fnaval> but doesn't the **kwargs expect the params being passed in to be a dict?  and that order doesn't matter?
18:00:08 <fnaval> jordanP: in your code snippet, I still see that as valid
18:00:21 <oomichi> please talk remaining on qa channel if we have
18:00:21 <oomichi> #endmeeting