17:01:15 #startmeeting qa 17:01:15 Meeting started Thu Jan 18 17:01:15 2018 UTC and is due to finish in 60 minutes. The chair is andreaf. Information about MeetBot at http://wiki.debian.org/MeetBot. 17:01:17 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 17:01:19 The meeting name has been set to 'qa' 17:01:45 #link https://wiki.openstack.org/wiki/Meetings/QATeamMeeting#Agenda_for_Jan_118h_2018_.281700_UTC.29 Today's agenda 17:01:56 o/ who's around for the meeting today? 17:03:01 not so many it seems 17:03:15 let's see if someone shows up otherwise I'll cancel for today 17:05:11 hi, is qa meeting on? 17:06:48 hi 17:06:55 I was waiting for someone to join :) 17:06:58 welcome 17:07:16 hi there 17:07:20 wznoinsk: since no-one else joined I'll skip the formal agenda 17:07:28 do you have anything you wanted to discussed 17:07:32 discuss? 17:07:34 I have one item from some time ago (never had a chance to bring it up): what's the prereq to have https://review.openstack.org/#/c/508554/ and similar changes for RH guest OSes accepted in Tempest? 17:07:35 patch 508554 - tempest - fix v6 tests on NetworkManager-based operating sys... 17:09:18 I'm about to double check and reply to Atilla's feedback, but most likely the code that's there is the correct one (we need both nm and ip commands) 17:09:29 uhm in general we try to avoid as much as possible OS specific things - but it's not possible to completely avoid it 17:10:10 yep, it doesn't break compatibvility with non network manager OSes (i.e.: cirros) tho, such change would cover three distros too 17:10:26 we had a plan in the past to make such guest OS specific things external to tempest.conf, like metadata in a guest os file that can be loaded or so 17:10:31 but it never came to live 17:11:30 andreaf, could we use this metadata to override internal functions of a test tho? 17:12:38 wznoinsk: yeah it should be possible but we never implemented that 17:13:26 andreaf, I'd imagine it could be even more challenging now when tests are in project's repos or project's tempest plugin repos 17:14:16 wznoinsk: the thing is that we don't want to be in the business on tempest side to maintain guest os specific things - but neither we want to prevent people from being able to use different guest OS 17:14:33 andreaf, so is it fair to say that because of lack of a proper soltuion as you said ^ we have to pack these things raw inside and do them right when any solution exists? 17:14:36 is there a specific reason for testing with a network manager based OS? 17:15:12 andreaf, yep, we test with rhel on the host and rhel in guest... gues which company I work for now ;-) 17:17:46 andreaf, definitively I'll follow whatever is the right path to do these things, I'm not kin on kicking off the project you've suggested tho - I'm not sure how many RH specifics will pop out like this more... for the moment only this and we run most of the tempest and neutron tests in my team 17:17:56 s/kin/keen 17:18:17 yeah sure - I'm still not sure using rhel as a guest really helps you validating rhel openstack 17:19:39 for guest specific testing it probably make sense to stick to scenario test (like the one in your patch) 17:19:41 I'd say the more variants of testing the better - more edge cases caught (even on the infrastructure side) 17:20:13 so as long as we don't get a stream of patches everywhere for rhel specific things I don't mind the one patch 17:20:49 an alternative for you could be to have a downstream plugin with rhel guest specific tests 17:21:23 so for https://review.openstack.org/#/c/508554/ as long as it's well documented and it does not break compatibility with cirros I don't mind 17:21:23 patch 508554 - tempest - fix v6 tests on NetworkManager-based operating sys... 17:21:49 but since we'll have to maintain that code it would be nice to have a nice long comment on top of the command explaining what it is about :) 17:22:41 let me do two things first, I'll double check atilla's comment + I'll review how the other tests that potentially do something with NIC's up/down or around behave and whether they could be impacted (I don't have any failing like this at the moment so I'll double check why) 17:22:53 ack on comment 17:23:08 ok sounds like a plan :) 17:23:35 ok, talk to you once it happens ;) thanks 17:23:49 sure np 17:23:53 anything else? 17:24:02 otherwise I'll close the meeting for today 17:24:09 I'm good 17:24:20 ok 17:24:28 ttyl 17:24:33 #endmeeting qa