17:00:32 #startmeeting qa 17:00:33 Meeting started Thu Dec 17 17:00:32 2015 UTC and is due to finish in 60 minutes. The chair is mtreinish. Information about MeetBot at http://wiki.debian.org/MeetBot. 17:00:34 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 17:00:37 The meeting name has been set to 'qa' 17:00:48 hi, who's here today? 17:00:49 hello 17:00:55 o/ 17:01:08 #link https://wiki.openstack.org/wiki/Meetings/QATeamMeeting#Proposed_Agenda_for_December_17th_2015_.281700_UTC.29 17:01:15 ^^^ today's agenda 17:01:49 o/ 17:02:51 let's get started 17:02:57 #topic Specs Reviews 17:03:03 #link https://review.openstack.org/#/q/status:open+project:openstack/qa-specs,n,z 17:03:29 o/ 17:03:29 we've got a couple of open specs still 17:03:35 #link https://review.openstack.org/173334 17:03:52 and 17:03:54 #link https://review.openstack.org/233187 17:04:05 although there both from andreaf now that I look at it :) 17:04:15 and 1 still needs to be updated 17:04:38 but just a reminder that we should still keep an eye on specs. We don't want things to bog down there 17:04:59 yes I need to update the specs - but I think they could use reviews anyways 17:05:15 I have time for reviews tomorrow.. will do 17:05:47 regarding test-resources, afazekas proposed to use testfixtures and testresources there, but it's not clear to me how they would fit really 17:06:47 regarding the o-h one, I will update it, and we probably could close it as well then as the 4 pages are up and running now 17:07:19 andreaf: do we even want to bother anymore with the o-h spec? 17:07:47 as for the testfixtures testresources, yeah I'm not sure where that would fit in. Unless he's saying make the test interface to the resources a fixture 17:08:06 mtreinish: I thought it was nice to have it there, but it's probably more interesting have docs on o-h side 17:08:35 andreaf: +1 for docs 17:09:16 andreaf: yeah I think just having docs would be better 17:09:31 mtreinish: also it would be great to have your review on the resources spec :) 17:09:40 it would have been better if we wrote a spec in ft collins when we started the project, but it's a bit late now 17:09:46 andreaf: sure I'll try to take a look post-meeting 17:09:51 ok is there anything else on specs? 17:09:55 there is also another spec I've started implementing, #link https://review.openstack.org/#/c/92804/ - this is very very old 17:10:29 but rather than trying to get the spec up to date I started implementing it, as it's easier to understand what I had in mind there 17:10:40 o/ sorry for the late arrival 17:10:51 dmellado: np 17:11:02 #link https://review.openstack.org/#/c/258566/ - this is the first patch, I'd love more reviews / opinions on that 17:11:19 it's a client manager refactor / split 17:11:39 to provide a client manager that can go to tempest-lib and lazy load of clients 17:11:58 andreaf: this is really needed 17:12:15 andreaf: wow, that is an older spec. Like you first submitted it when we started specs 17:12:39 mtreinish: :D indeed 17:13:16 andreaf: should we try to convert the manila client to see how it's working? 17:14:10 mkoderer: sure - I'd love that - the register interface is not implemented yet, but it shouldn't take me long 17:14:30 andreaf: are you working on that tomorrow? we can try to spilt the effort 17:14:59 I also think this will help with the clients refactors, as it's painful to carry around many clients where it used to be one 17:15:15 mkoderer: yes, I will be working on this topic tomorrow most likely 17:15:30 andreaf: ok cool I will ping you 17:15:32 mkoderer: let's sync in IRC 17:15:59 ok, let's move on 17:16:12 #topic Priority Items 17:16:19 #link https://etherpad.openstack.org/p/mitaka-qa-priorities 17:16:37 Does anyone have anything to discuss on any of the priority work items? 17:17:26 andreaf, jlanoux: how goes the ssh auth stuff? 17:18:11 mtreinish: the patch to fix the deprecation part is done 17:18:23 #link https://review.openstack.org/#/c/253444/ 17:18:42 after that, I plan to do the wrapper around compute and remote_client 17:18:59 mtreinish: there is also work to be done to take some of the neutron scenario tests functionality out and back into the validation framework 17:19:40 andreaf: do you have an example? 17:19:41 mtreinish: but honestly I would even consider that an extra - post spec work - 17:19:50 sure, that wasn't part of the spec 17:20:04 I'd just like to see the spec finally finished :) 17:20:09 it's been a long road 17:20:24 mtreinish: well the scenario tests have a lot of validation helpers, everytime I try to follow the code through them I get lost 17:20:47 mtreinish: and it would be good to make sure they all as much as possible honour all validation settings 17:21:06 andreaf: right, I agree 17:21:11 self.var usage is confusing , we should do more explicit parameter passing 17:21:29 afazekas: +100 17:21:44 afazekas: yes, there is a bug open for that (or there was) there were patches starting to clean it up 17:21:47 but it's still a mess 17:22:55 ok, is there anything else on priority efforts? 17:22:58 two cli patches could use reviews 17:23:07 #link https://review.openstack.org/#/c/247240/ 17:23:17 #link https://review.openstack.org/#/c/252428/ 17:23:52 I'll have a look at them this afternoon 17:24:47 ok, let's move on 17:24:55 #topic Tempest 17:25:12 there are 2 items on the agenda for tempest 17:25:28 mkoderer: yours is up first :) 17:25:32 ok :) 17:25:44 IMHO we have an issue with logging and tempest CLI 17:25:59 the this is that logging.conf isn't really used 17:26:07 wasn't the log environment not working anymore? 17:26:10 yep 17:26:38 so all the scripts log to tempest.log but nothing is shown on stdout 17:26:48 for instance stress 17:27:13 I started to clean it up here https://review.openstack.org/#/c/258955 17:27:52 but the question is: do we want a default logging.conf that defines stdout logging for all cmd output 17:28:37 or we use "print" for all the user output 17:29:02 mkoderer: I'm thinking if we want to do stdout we should just use print 17:29:18 or manually write to sys.stdout 17:29:23 mtreinish: that fine for me 17:29:43 I feel like the logging should probably be left configurable for users 17:30:19 mtreinish: so I will fix stress and cleanup cmd then 17:30:26 ok, cool 17:30:38 Guess I am a bit confused, some scripts are using LOG. and some print() 17:30:53 dpaterson: yeah, that's the issue :) 17:30:54 Are you saying we should change all to print()? 17:31:16 dpaterson: if the tool wants something to go to stdout (like final output of a command) it should use print 17:31:18 dpaterson: no.. log messages we use LOG but for user output we use print 17:31:44 dpaterson: for instance the statistics of the stress test run should go to stdout 17:32:41 ok I push some patches tomorrow 17:33:03 ok, then the other topic under tempest was from me 17:33:20 about removing the ec2 tests from tempest 17:33:21 mkoderer: I'm fine with refactoring all cmd scripts, I was trying to get them to parity with cliff first 17:33:31 #link https://review.openstack.org/#/c/222737/ 17:33:35 I noticed the inconsistancies too 17:34:11 basically nova has deprecated their in-tree ec2 implementation, and want to remove it soon 17:34:21 dpaterson: yeah cool let's have a review on it tomorrow 17:34:32 but the tempest tests are blocking that. We also said the only reason we keep the tests in tree is because nova has it in tree 17:34:59 mkoderer: sure, tx 17:35:00 mtreinish: then it looks like a straightforward move to remove them too 17:35:13 so I made a tempest plugin with the code and put it up on github (also the ec2-api project has their own tempest plugin) to preserve the tests 17:35:25 just in case someone want to use them for something in the future 17:35:32 dmellado: yep 17:35:36 mtreinish: +1 17:35:40 +1 17:35:44 I've brought it up on the ML before too 17:36:00 I just wanted to bring it up on the meeting, because we should be good to land the removal patch 17:36:18 I'll push a tag after we remove it to signify the removal 17:37:29 ok, does anyone else have anything else to discuss on tempest this week? 17:37:49 mtreinish: I had some comment on that patch, I'll re-review tomorrow if that's ok 17:38:10 andreaf: that's fine. (although dims respun the patch to jsut remove the options) 17:39:34 ok, if there's nothing else let's move on 17:39:45 #topic DevStack + Grenade 17:40:00 does anyone have anything to discuss about devstack or grenade? 17:40:06 sdague, dtroyer: ^^^ 17:40:42 the extras.d patch landed 17:40:44 finally 17:40:51 yay 17:41:02 it only took a week of rechecks :) 17:41:12 yeh 17:41:32 I also brought over some code from d-g to make our apt_get_update more resilient to mirror breaks 17:41:40 that seems to be helping, as is backported 17:42:06 cool 17:42:17 I've also got a new angle on the "can't ssh" to test nodes issue 17:42:30 which is that there is a race in the fixed ip allocator in nova 17:42:45 where we might allocate an ip that has an oustanding lease 17:43:03 so the guest won't get a valid ip, and dhcp fails 17:43:09 there is a nova patch up for that 17:43:16 I can't wait to see how that impacts things 17:43:26 cool, that does sound like it could impact things for the better 17:43:38 do you have a link to the patch? 17:44:59 ok, does anyone else have anything to discuss on devstack or grenade this week? 17:46:11 ok, then lets move on 17:46:15 #topic Critical Reviews 17:46:36 does anyone have any reviews they'd like to get extra eyes on? 17:48:06 https://review.openstack.org/#/c/257557/ 17:48:59 afazekas: yeah that does feel like something that should be a neutron side fix 17:49:13 afazekas: I'll follow up with armax or someone to get them to weigh in on it 17:49:37 mtreinish: Actually it is not, but IMHO the neutron agent scheduler tests should be in the neutron repo 17:49:48 These kind of behaviors kind of internal. 17:50:11 afazekas: oh, yeah that definitely feels like an internal neutron thing 17:50:30 are there any other reviews? 17:51:33 ok, then lets open up the floor 17:51:38 #topic Open Discussion 17:51:53 does anyone have something they'd like to discuss that wasn't on the agenda? 17:53:23 oh, one thing is I'm gonna cancel the next 2 meetings 17:53:37 because of the holidays I expect not too many people will be around 17:53:38 +1 from me, almost everybody would be out for Christmas 17:53:57 I'll send something to the ML about that today 17:54:15 ok, if there isn't anything else I'll end here 17:54:19 thanks everyone 17:54:22 #endmeeting