17:01:09 #startmeeting qa 17:01:10 Meeting started Thu Jan 15 17:01:09 2015 UTC and is due to finish in 60 minutes. The chair is dkranz. Information about MeetBot at http://wiki.debian.org/MeetBot. 17:01:11 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 17:01:14 The meeting name has been set to 'qa' 17:01:49 So who is here today https://wiki.openstack.org/wiki/Meetings/QATeamMeeting#Proposed_Agenda_for_January_15th_2015_.281700_UTC.29 17:02:22 hi 17:03:14 jlanoux: hi 17:03:21 Any one else here? 17:03:39 i'm here 17:03:42 me too 17:03:57 ok, let's get started 17:04:05 1/2 here 1/2 in other meeting :) 17:04:12 #topic Spec Reviews 17:04:25 Any specs to review? 17:04:51 I was shooting to have an update in today but didn't make it in time 17:05:05 I will have an update COB today for cli spec 17:05:22 dpaterson: ok, great 17:06:03 ok, let's move on to blueprints 17:06:27 I think Ken'ichi has made progress on preparing clients for move to tempest-lib 17:06:40 Any other blueprints to bring up? 17:07:06 Looks like this will be a short meeting :) 17:07:17 Any grenade or devstack folks here? 17:07:28 dkranz: sorry, am now 17:07:50 sdague: np. Anything to discuss there? 17:08:20 py26 support with trunk: https://review.openstack.org/#/c/146819/ ? 17:08:38 not much really. The biggest thing is trying to get people to use the devstack plugin interface instead of adding lots of stuff to devstack 17:09:02 sdague: is that an education problem? 17:09:54 sdague: have you been letting stuff in or saying "no", use the plugin interface? 17:10:30 the latter 17:11:07 it's partially education, partially getting a workable pattern for it in the gate. The glusterfs folks are one of the early folks pushing through this. 17:11:47 sdague: Maybe they can be used as a template. I guess that is how we usually get people to adopt new things, but pointing to something that already did it 17:12:06 yep 17:12:10 that's about it 17:12:29 sdague: Maybe an email about this would help. I might have missed it but did see one 17:12:35 sdague: ok, thanks 17:12:51 afazekas: This is quite a contentious issue about py2.6 17:13:31 afazekas: Why do people need to run tempest on py2.6 systems even if their clouds are running on such systems? 17:13:44 afazekas: Have the python clients dropped py2.6? 17:13:49 dkranz: tempest master does not works with py26, the question is do we want to accept patches related to py26 compatibility 17:14:06 afazekas: yes but I think it is more than that 17:14:26 afazekas: If we say that 2.6 is not supported than a lot more people will be submitting things that break it 17:14:40 afazekas: Til now this has not been a big issue but it will get bigger 17:14:44 dkranz: probbaly it is simper then creating a py2.7 env and ignoring the *-manage tests 17:15:43 afazekas: If we are going to keep 2.6 working by accepting patches, it would be better to have a gate job. 17:15:54 afazekas: But many of our cores do not want that 17:16:15 afazekas: I think we need a ml discussion and some kind of decision. 17:16:35 dkranz: it is expexted to break 3 times per cycle, a random failure nosse from any additional job causes more issue then accepting 3 patches 17:17:02 dkranz: as a non tempest core I am pretty sure openstack as a whole sort of made that decision back in atlanta 17:17:11 dkranz: I don't understand why we would need to bring the discussion back up again 17:17:13 afazekas: I am saying that I expect it to break more than that once people realize that breaking it is ok 17:17:35 clarkb, tempest does not have stable/juno branch or stable/icehouse branch 17:17:40 clarkb: I meant withing tempest 17:17:49 afazekas: dkranz that is orthogonal 17:18:06 openstack has sai "we do no support python 2.6 except where necessary for previous releases" 17:18:24 thats the short tl;dr of the current stance as interpreted by me at least 17:18:44 clarkb: I'm actually not opposed to that 17:18:46 tempest python2.6 support is not necessary for previous releases 17:19:08 clarkb: Was this a tc decision that we can just point to? 17:19:17 dkranz: it was an atlanta summit session 17:19:30 there should be an etherpad linked off the wiki 17:19:43 * afazekas wasn't @ atlata 17:19:51 clarkb: I don't think that is good enough as our issue is that our "customers" are complaining about it which is what I was trying to address 17:20:03 (note I say python 2.6 support is not necessary for previous releases because we haven't run tempest on python2.6 against any of the current stable releases upstream) 17:20:27 clarkb: you mean for tempest? 17:20:34 yes specifically talking tempest here 17:20:41 clarkb: right, and that was silly 17:20:49 clarkb: because we have kept it working anyway 17:20:59 #link https://etherpad.openstack.org/p/juno-cross-project-future-of-python 17:21:20 that has the outcomes of the session about which versions to support 17:22:11 so the crux of this argument is red hat uses rhel6 for test nodes internally, right? 17:22:29 what is the policy with the pythonclients ? Tempest should be handled similarly. 17:22:42 sdague: This was not just coming from Red Hat. Suse also. 17:22:50 afazekas: pythonclients and other libs like oslo continue to support python2.6 until juno eol 17:22:57 but I disagree that tempest and python clients are similar 17:23:15 tempest is a utility you run it against a cloud from whatever python install 17:23:27 clarkb: They are not. But our policy is either "we all drop 2.6 now" or "some of us do" 17:23:31 its not a lib linked into long lived applications that we want to support for a reasonable period of time 17:23:40 dkranz: no that isn't the policy because thats impossible 17:23:47 clarkb: I was just clarifying that if it is the latter, then it is up to the tempest team to decide 17:24:39 clarkb: there is disagreement on the patch so this needs to be discussed an closed 17:24:40 honestly the python clients support requirement is only because our testing situation is wonky, and we let master uncapped clients with older servers 17:24:46 But not at this meeting 17:25:05 dkranz: right I am trying to say this is already discussed and closed. suse and red hat were both in that room 17:25:07 so I feel like we've been warning people forever that python26 was getting dropped from tempest 17:25:16 there wasn't any major disagreement to the decision that was made 17:26:01 afazekas: I think we should discuss offline why this is so important 17:26:42 In any event we should create a new tempest tag to be "eol of py2.6" 17:27:02 dkranz: that would have been before there were tempest releases 17:27:29 dkranz, ok, I had issue with bleeding edge systems, Please have a look https://review.openstack.org/#/c/146039/ 17:27:30 sdague: I meant "last tempest that worked with 2.6" 17:27:42 sdague: which was not too long ago 17:28:31 afazekas: So I put that review on the agenda at the end. I do not understand it well enough to give +2 17:29:40 dkranz: The short version the double 'l' makes me happy (and f21), and does not makes unhappy the previous change author :) 17:31:36 afazekas: well, we don't gate on these system differences. Can you get a review from him? 17:33:05 dkranz: I can try , On the cirros bug you can see what was difference what I also inspected.. 17:33:13 afazekas: thanks 17:33:22 #topic Open Discussion 17:33:32 Anything else any one wants to discuss? 17:33:55 dkranz: sorry I forgot that: #link https://review.openstack.org/#/c/142809/17 17:34:28 jlanoux: ok, looks like it is passing now. I will complete review today 17:34:50 dkranz: yes, finally! thanks. 17:34:57 If some one else could review it as well that would be great 17:36:31 jlanoux: Looks like the move of RestBody to service may have already merged. 17:36:53 jlanoux: If so you will need to rebase and change the client import 17:37:33 jlanoux: yup, it merged 17:37:38 dkranz: it did merged, that's why my change failed 17:37:51 dkranz: I alreday updated that 17:37:56 jlanoux: great! 17:38:11 Anything else? 17:38:59 ok, thanks every one. I think py2.6 will be gone. 17:39:43 #endmeeting