17:00:51 #startmeeting qa 17:00:52 Meeting started Thu Mar 31 17:00:51 2016 UTC and is due to finish in 60 minutes. The chair is oomichi. Information about MeetBot at http://wiki.debian.org/MeetBot. 17:00:53 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 17:00:55 The meeting name has been set to 'qa' 17:01:07 hi, anyone here today ? 17:01:13 o/ 17:01:28 hi 17:01:36 oh, hi tosky ;) 17:01:53 o/ 17:02:03 o/ 17:02:03 hi all :-) 17:02:13 ok, lets get start 17:02:27 #link https://wiki.openstack.org/wiki/Meetings/QATeamMeeting#Proposed_Agenda_for_March_31th_2016_.281700_UTC.29 17:02:34 ^^^ today agenda 17:02:37 o/ 17:02:45 #topic #Spec reviews 17:03:16 there are some active specs 17:03:28 #link https://review.openstack.org/#/q/status:open+project:openstack/qa-specs,n,z 17:03:46 one is o-h spec 17:03:56 that seems old now, is it necessary still now? 17:04:15 #link https://review.openstack.org/#/c/233187/ 17:04:31 oomichi: I think I asked that exact question in the spec :) 17:04:33 most part seems already implemented, it seems unnecessary 17:04:53 mtreinish: to andreaf? 17:05:14 oomichi, no, in a comment 17:05:38 I mean, yes to andreaf, but on Gerrit 17:05:45 jordanP: ah, I see. 17:05:57 ok, lets wait for andreaf comment 17:06:01 oomichi: yeah I left a comment on it asking why it's still need. I also screwed up with gertty and dropped my -1 by accident 17:06:40 gertty is dangerous, even more if there's a cat around... 17:06:43 mtreinish: yeah, it is nice to put -1 for getting attention 17:06:49 again 17:07:16 btw, I put another spec: https://review.openstack.org/#/c/297473/ 17:07:32 that is for changing json-schema format for fitting swagger 17:07:47 swagger is defined as standard in all openstack projects 17:08:07 oomichi: is that just changing how we write the dictionaries in the schema files? 17:08:08 so I feel it is nice to make our json-schema also fit it 17:08:20 mtreinish: yeah, right. that is a small changes 17:08:33 in schema files 17:08:50 does Swagger "enforce"/check the response as json-schema does ? 17:08:51 oomichi: is there something we gain by doing that? I mean sure swagger is "the direction" but it's not like we're gonna generate api docs from tempest 17:09:17 jordanP: sorry, I cannot catch your meanin 17:09:53 mtreinish: no, that is not current purpose. 17:10:17 I though we use json-schema because it has a python binding that checks that the responses we got match the expected response 17:10:28 I don't think it is nice to generate doc from tempest at this time, but that is a surper long-term direction for me 17:10:57 jordanP: yeah, you are right. 17:10:59 in Tempest we have "import jsonschema", are we going to replace this with "import swagger" ? 17:11:48 anyway, this looks like low priority to me, like a "nice to have" 17:11:52 jordanP: swagger is using jsonshema as the definition. but we defined some part of our jsonschema by our own way 17:11:56 oomichi: right, so I'm wondering is all the churn necessary just to be consistent with an API WG recommendation? 17:12:02 I am trying to change it to swagger way 17:12:39 ok. If that's a relatively small/contained change, then go for it :) 17:12:47 mtreinish: nice point. all jsonschema definition can be changed in tempest for that. 17:13:18 oomichi: so then all the jsonschema will have to be changed? 17:13:32 dmellado: yeah, I hope so. 17:13:46 it is not difficult, but many ;-( 17:14:03 again xD 17:14:05 so it is nice to get some helps for that 17:14:19 dmellado: yeah, nice point. sorry about that 17:14:39 I've one question, would the API WG change (swagger) affect the tempest compatibility when it comes to previous os pversions? 17:15:03 I'm not really savy in terms of this swagger stuff, so I could use some explanation on this 17:15:06 oomichi: right, that's why I'm asking do we really want to do this? I feel like this doesn't buy us anything really useful besides saying we use swagger 17:16:30 mtreinish: humm, *I* really want to buy ;_) 17:16:40 mtreinish: humm, *I* really want to buy ;-) 17:17:03 oomichi: ok, I'll tkae up my concerns in the review 17:17:07 we continue discussing there 17:17:15 yes 17:17:16 +1 17:17:19 mtreinish: thanks, yea, lets move on 17:17:29 thanks all :-) 17:17:44 does anyone have more topic about spec now? 17:18:04 ok, lets move on to the next topic 17:18:22 #topic priority items 17:18:39 #link https://etherpad.openstack.org/p/mitaka-qa-priorities 17:19:07 oomichi: fwiw, there is 1 week left in mitaka :) 17:19:15 most targets are done in mitaka 17:19:25 not a lot of time left to close any last min things 17:19:26 mtreinish: yeah, just one week 17:19:52 do we have any items before mitaka now? 17:20:10 I think "Finalize ssh-auth bp" is done 17:20:18 that's what jlanoux and andreaf said 17:20:20 last meeting 17:20:26 jordanP: cool :_) 17:21:11 ok, how about concentrating on the next cycle? 17:21:26 for the next summit topics 17:21:51 that is next next next topic 17:22:02 lol 17:22:06 xD 17:22:11 sure sure sure 17:22:12 ok, lets move on 17:22:21 #topic Tempest 17:22:53 there are still a lot of tempest patches 17:23:02 #link https://review.openstack.org/#/q/project:openstack/tempest 17:23:37 I dont have any topic about tempest now 17:23:41 do anyone have that? 17:24:15 my patch to replace httplib2 by urllib3 is green: https://review.openstack.org/#/c/295900/ 17:24:29 jordanP: saw it, I think it'd be great to have that in mitaka 17:24:34 other than that, my next match is going to remove our dependency on mox 17:24:42 and not related directly to tempest, but I'd love to have reviews on this 17:24:45 https://review.openstack.org/#/c/274023/ 17:24:46 we should use mock nowadays 17:25:13 jordanP: any other 'deprecated' lib? 17:25:16 jordanP: we use mock in most places IIRC 17:25:20 jordanP: ok, will review it later 17:25:26 there might be a few lingering mox usages from when I first added unit tests 17:25:28 yeagh, replacing mox is going to be triviial 17:25:34 #link https://review.openstack.org/#/c/249100/ 17:25:43 #link https://review.openstack.org/#/c/274023/ 17:25:47 I already have the patch, it's a 30 lines patch. Going to submit it soon 17:26:12 dmellado, you should seek support from Neutron guys on this patch 17:26:13 jordanP: one question 17:26:26 jordanP: I'm doing that too ;) 17:26:43 jordanP: about https://review.openstack.org/#/c/249100/ 17:27:04 that is adding volume test, and most volume test tends to wast the time 17:27:07 the only critical thing I have for tempest this week is: https://review.openstack.org/297322 17:27:20 jordanP: the patch is good from the viewpoint? 17:27:23 we just need to make sure that lands before we can push the next tempest release to mark mitaka 17:27:54 mtreinish: that contains jordanP comment, that seems reasonable 17:27:57 oomichi, this patch is good yes. But let's not merge it before we reach concensus about what should we do about long running tests 17:28:24 and tests that only touch one service 17:28:24 jordanP: ok, nice info. nice to mark it as WIP 17:28:54 jordanP: we do have the slow tag, which will make it only run on periodic/experimental 17:29:20 it's not a WIP, anyway. Ahh yeah, I forgot about the slow tag. Good idea 17:29:45 let's not worry about this patch 17:29:49 yes, please it'll be nice to add that slow tag in the notes too 17:30:02 jordanP: hehe, ok. fine :-) 17:30:15 next cycle I am going to focus on reducing technical debt and complexity in tempest 17:30:22 and try to have faster running tests 17:30:22 jordanP: awesome 17:30:36 jordanP: cool ;) 17:30:47 how about the tempest refactor? 17:30:56 will that be included in your goal jordanP ? 17:31:21 dmellado: ah nice point. 17:31:25 "refactor" is a big vague. what do you have in mind ? 17:31:28 *bit 17:31:35 jordanP: well, it was marked as a point for mitaka 17:31:44 if refactoring needs to be done, refactoring will be done 17:31:58 dmellado, link ? 17:32:04 and I was wondering about the exact meaning too 17:32:06 dmellado: maybe you are saying https://review.openstack.org/#/c/92804/ ? 17:32:11 oomichi: let me check 17:32:12 ah ok 17:32:22 https://etherpad.openstack.org/p/tempest-refactor-ideas 17:32:33 jordanP: yep 17:32:37 #link https://etherpad.openstack.org/p/tempest-refactor-ideas 17:32:59 maybe we can chat about the exact goal in the design summit 17:33:00 oh, that etherpad I made. It was just a title, I didn't put deep meaning behind the title 17:33:01 yeah, there are some good ideas there 17:34:00 that is already put on the session proposals on the etherpad 17:34:09 as design session 17:34:16 oomichi: cool then ;) 17:34:40 ok, anyone have more topic about tempest now? 17:35:23 #topic DevStack + Grenade 17:35:43 dtroyer: are you there for devstack? 17:36:20 the only news here I think is we've branched devstack and grenade for mitaka 17:36:44 mtreinish: nice news:-) 17:36:58 ah, one question about devstack 17:37:29 sometime devstack makes bugs disappear because it changes default setting values 17:37:49 and we faced bugs on production clouds 17:38:08 so I feel it is nice to make devstack small as possible 17:38:20 oomichi: ++ 17:38:23 is it a right direction 17:38:23 that's the current direction indeed 17:38:26 ? 17:38:27 devstack plugins and stuff 17:38:28 oomichi: devstack tries to use the defaults where possible, but sometimes we can't always 17:38:40 (also, btw, devstack-gate should contain as little extra logic that's not in devstack itself) 17:38:50 mtreinish: yeah, that is difficult point 17:39:10 mordred: ah, nice advice, thanks 17:39:40 ok, my question is done, thanks 17:39:48 I didn"t see any question :) 17:39:54 I yes, my bad 17:40:03 the question mark came after 17:40:13 jordanP: yeah, I was not sure that is right direstion to be honest 17:40:30 I think it is the right direction and devstack plugins are awesome 17:40:34 I have a question that may fit on this general topic, but it might should have come up when the topic was tempest (sorry) 17:40:35 we have so many of them, it works 17:41:09 mordred: go for it 17:41:09 jordanP: ok, lets talk about it on open discussion 17:41:17 I have been told that tempest does not do well on clouds that do not have floating ips ... is there anybody working on support for provider-network only clouds and it not would work on such a thing be acceptable? 17:41:38 (assuming everyone knows how evil I tink fips are in the first place) 17:41:39 mordred: it should work fine, but I don't think that's the default configuration 17:41:53 mtreinish: ok. so there should be a config setting I just need to find? cool 17:42:04 I don"t think it should work well 17:42:07 mtreinish: I will look harder and find you for follow up questions if I can't find it 17:42:15 the scenarios assume floating ip now 17:42:17 mordred: likely a couple of options 17:42:21 ah 17:42:22 jordanP: really, that's broken 17:42:26 jordanP: so I might need to update scenarios? 17:42:35 lemme double check, but I think so 17:42:55 mordred: but that could also be the case for scenarios coming from plugins... 17:43:01 mordred: I was gonna point you to: http://docs.openstack.org/developer/tempest/configuration.html#enabling-remote-access-to-created-servers 17:43:29 let's wait for jordanP 17:43:31 dmellado: I think right now we're just trying core stuff, but good point - I'll start looking through those next 17:43:44 mtreinish: awesome docs thing 17:44:00 no I think am wrong :) 17:44:05 sorry about that 17:44:17 mordred, https://github.com/openstack/tempest/blob/master/tempest/scenario/manager.py#L644 17:44:29 the relevant config flag is CONF.validation.connect_method 17:44:43 woot! that's happy making. 17:44:58 best openstack meeting ever. thanks everybody 17:45:04 so that's the same as per mtreinish link ;) 17:45:33 mordred: cool :) 17:45:54 anyone have more items about devstack? 17:46:09 ok, lets move on 17:46:13 #topic Austin Summit 17:46:23 #link https://etherpad.openstack.org/p/newton-qa-summit-topics 17:46:33 we have 8 ideas now for 8 slots 17:46:41 pb is we have a couple of scenarios (2 or 3) that calls directly manager.create_floating_ip(). This should be fixed and use get_server_ip() instead. This is something to look at 17:46:50 so it is nice to get more ideas on the etherpad 17:46:53 oomichi: I counted 9 :) 17:47:11 mtreinish: ah, miss:-( 17:47:51 are there any strong recommendation about the summit topic? ;) 17:48:05 it is nice to vote for each session by writting names 17:48:17 oomichi: when will we make decisions? 17:48:45 mtreinish: you and me, and when..? sorry, I need to know it 17:49:07 oomichi: well we've done it as part of the qa meeting in the past 17:49:20 what's the ETA for the decission before the summit? 17:49:53 that is based on the voting I feel 17:50:06 oomichi, dmellado: I think we'll probably make decisions on like the 14th 17:50:22 we need to have things sorted a week or 2 before summit starts 17:50:46 mtreinish: thanks. OK, I will send a mail to -dev about this later with the deadline 17:51:12 ok, lets move on 17:51:19 to the next topic 17:51:25 #topic critical reviews 17:51:39 please put patch links for that;-) 17:52:35 Not critical, but would love to get eyes on https://review.openstack.org/#/c/293052/ 17:52:44 jswarren: thanks 17:52:54 #link https://review.openstack.org/#/c/293052/ 17:53:22 any more? 17:53:42 jswarren, it would be great if you could ping andreaf about it 17:54:02 jordanP: Ok. Will do. 17:54:30 jswarren: does that conflict with the change we pushed at the midcycle for cleaning up the aliasing in one of those files 17:55:06 mtreinish: trying to remember. 17:55:40 That was https://review.openstack.org/#/c/283806/ 17:55:46 jswarren: yeah I think so 17:56:16 It does not conflict. 17:56:33 ok cool 17:56:36 was just curious 17:57:30 both are now by jswarren now 17:57:42 it is nice to write on dependency? 17:57:51 just an idea 17:58:03 They are actually not dependent. 17:58:30 jswarren: we can avoid some concern like mtreinish's one 17:58:47 jswarren: it is nice to keep it on current way also 17:59:04 just one idea ;) 17:59:26 are there any topic about this more? 18:00:01 ok, thanks all 18:00:05 #endmeeting