09:00:51 #startmeeting qa 09:00:52 Meeting started Thu Oct 15 09:00:51 2015 UTC and is due to finish in 60 minutes. The chair is oomichi. Information about MeetBot at http://wiki.debian.org/MeetBot. 09:00:53 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 09:00:55 The meeting name has been set to 'qa' 09:01:06 who's here for today's meeting 09:01:13 o/ 09:01:15 hey 09:01:20 hi guys 09:01:36 o/ 09:01:48 #link https://wiki.openstack.org/wiki/Meetings/QATeamMeeting#Proposed_Agenda_for_October_15th_2015_.280900_UTC.29 09:01:56 ^^^ today agenda 09:02:26 Ok, let's get started 09:02:44 #topic Mitaka Summit Prep 09:02:45 hi guys o/ 09:03:00 #link https://etherpad.openstack.org/p/mitaka-qa-summit-topics 09:03:19 we have proposals for our design sessions 09:03:53 and we need to select sessions 09:04:12 mtreinish is choosing them now 09:04:30 yea 09:04:38 oomichi, do we need to select or we can have them all ? 09:04:56 Fishbowls are full now 09:05:21 jordanP: maybe we can select sessions for Work Rooms 09:05:32 oomichi: in the etherpad , how about the 1st fishbowl? is that a typo? 09:05:47 or it'll be filled by another topic 09:05:57 dmellado: oops, yeah I am not sure that :( 09:06:32 oomichi, dmellado: think there are 4 fishbows, so one is not assigned 09:07:04 andreaf: ah, maybe that is right 09:07:15 Is Fishbowl name of room?? 09:07:17 I think if any session did not get room then we have last day full 09:07:20 we can select them by writting name on "interested" 09:07:36 oomichi, dmellado: more people attends the fishbowls, so I guess mtreinish needs to select the topic suitable for that 09:07:52 o/ 09:07:54 shuquan: it's the type of session 09:08:10 andreaf: ty for the explanation 09:08:19 shuquan: in fishbowl sessions there is a presenter in a ~larger room 09:08:32 oh, i see 09:08:41 andreaf: thanks. 09:08:42 yeah, fishblowl is like https://www.google.co.jp/search?q=fishbowl+room&safe=off&hl=ja&biw=1276&bih=520&source=lnms&tbm=isch&sa=X&ved=0CAYQ_AUoAWoVChMIr5O6opDEyAIVQp6mCh0tLgpW 09:08:51 the work sessions are moderated by one person still, but usually everyone sits at the same table 09:10:05 most sessions which are proposed are available on summit 09:10:18 by numbers of room and proposal 09:11:07 ok does anyone have anything to discuss on this topic? 09:11:17 andreaf: quick question, for the existing resources session would you address also the account creation? aka accounts.yaml 09:11:19 ? 09:11:23 that'd be also nice 09:12:16 dmellado: in what sense? 09:13:03 andreaf: I was thinking into something for addressing the mapping between accounts and resources 09:13:27 just an idea, anyway 09:13:58 dmellado: yes I was thinking we could have a common interface in the tests to access global, access wide and tenant wide resources 09:14:25 andreaf: ack, that'd be great 09:14:25 dmellado: but at this stage it's just a vague idea, I only added a sentence about it in the spec 09:14:40 I'll take a look and add comments ;) 09:14:50 dmellado: thanks - https://review.openstack.org/#/c/173334/ 09:15:25 ok, let's move on 09:15:40 #topic Specs Reviews 09:15:54 #link https://review.openstack.org/#/q/status:open+project:openstack/qa-specs,n,z 09:16:13 about "Restructure Scenarios" 09:16:24 so I was suppose to respin this spec 09:16:26 jordanP: yeah, please go ahead 09:16:42 but I have no optimal idea on what to do yet so my plan is 09:16:50 jordanP: yep, I thought we might want to discuss that 09:16:55 to cleanup scenarios gradually 09:17:18 mriedem and I started to remove uneeded class variable 09:17:20 jordanP: I saw you started a series of patches to decouple scenario method 09:17:22 to use local vars instead 09:17:29 andreaf, yes 09:17:42 I'll continue on that and then 09:17:47 jordanP: yes that's useful if we want to create a library of helper methods for scenario tests or so 09:17:48 I want to remove code duplication 09:17:57 like there's 6 "boot_server" method 09:18:12 jordanP: yea, that will be nice 09:18:16 jordanP: + 09:18:16 having a common lib for all scenario tests 09:18:16 everything should be in manager.py and it should have only one method 09:18:18 will be great 09:18:30 because there are quite some code embedded in the tests themselves 09:18:34 jordanP: +1 09:18:38 then maybe make some methods static 09:18:39 jordanP: yes we already create one which is meant to be common to API and scenario tests, but it's not applied to scenario tests yet 09:18:57 so my general plan is to improve all this gradually :) 09:19:03 no spec then :) 09:19:12 most of this is general good practise 09:19:13 jordanP: so that mean necessary spec 09:19:19 jordanP: do you link those together for review? 09:19:26 jordanP: http://git.openstack.org/cgit/openstack/tempest/tree/tempest/common/compute.py#n30 09:19:26 jordanP: with common topic etc 09:19:33 jordanP: I guess we should write a spec, or maybe modify that 09:19:44 in order to have those 'common sense' rules applied xD 09:19:59 that seems cleanup thing, and it is fine to go forward without spec 09:20:12 yeah, it's mostly cleaning and it should not be controversial 09:20:23 but oomichi jordanP how about the common methods? 09:20:40 common methods for now, should be in manager.py 09:20:48 as a first small step 09:21:02 jordanP: so basically you suggest some cleanup from the code at least at first 09:21:08 if we remove instance variable and side effects, reasoning about that manager.py should be way easier 09:21:26 ad then maybe use tempest/common. but latter 09:21:34 +1 on that 09:21:37 yeah, +1 09:22:03 do you want to add any action for that for now? 09:22:11 gmann, no common topic. i forgot about that. i will 09:22:40 jordanP: Thanks, it will quick to review those 09:22:40 I am going to continue my effort on "remove instance variables" 09:22:59 then I'll start a "remove methods duplication" work 09:23:31 and then we will see what we have and what should be the next move 09:23:53 jordanP: I'll take a look at that too, so let's sync later after the meeting ;) 09:23:56 +1 09:24:11 great 09:24:23 cool 09:24:29 does anyone have anything to discuss on this topic? 09:24:42 about qa-spec 09:24:44 oomichi: I have a couple of specs 09:24:52 andreaf: please :) 09:24:53 oomichi: #link https://review.openstack.org/#/c/173334/ 09:25:37 this is the test resources once - I revamped and added some more ideas, so it would be great to get some feedback on it 09:26:03 it's going to be discussed at the summit so the more idea / comments the better 09:26:26 yeah, nice to get more ideas before the summit for productivity 09:26:28 I'll take a look at it 09:27:18 andreaf: I didn't read this yeat, can you introduce small overview? 09:27:25 the main ideas are: cleanup tempest.conf of any configuration which requires a running cloud (e..g image_id), have a resources.yaml file to specify test resources and their properties, and allow test to run against multiple resources (i.e. testscenarios) 09:28:08 aren't we replacing a big tempest.conf by a big resources.yaml ? 09:28:51 andreaf: what I was thinking also was about to use the accounts.yaml to track who belongs the resources (maybe using the tag) so we'd be able to feed that to the cleanup script. I'll give it a thought into how to combine the efforts 09:28:55 jordanP: tempest.conf will keep the list of services active / non-active, available extensions, timeouts, log paths etc 09:29:18 this work is huge, basically we could have to add 1 or 2 tags to each test and rewrite some tests also,no ? 09:29:23 andreaf: "run against multiple resources" you means running multiple instance of tempest with separate resource file ? 09:29:51 gmann, like running this specific test with a Cirros image and an Ubuntu image 09:29:57 gmann: I mean a test could run multiple time against a matrix of input configurations, e.g. all the images and flavours 09:30:13 yea 09:30:25 mmm that'd be quite useful 09:30:28 gmann: I use that a lot for instance to test guest images in public clouds and distributions 09:30:48 ohhk 09:31:15 one problem with that though is the test uuid - we would have one uuid for what could be considered as multiple tests 09:31:48 andreaf: these combinations are selected randomly during test operation? 09:31:55 andreaf: for that how about adding resource info as metadata with tests 09:32:17 oomichi: the idea is to have filters on resources, so the combinations are of all parameters passing the filters 09:32:42 so all images matching "ssh-able" with all flavours matching "smaller than X in disk size" for instance 09:32:44 andreaf: cool idea :) 09:32:59 979558 09:33:01 790613 09:33:06 sorry guys 09:33:25 dmellado: credit card details? :D 09:33:29 825111 09:33:33 nope 09:33:34 :D 09:33:37 luckily xD 09:33:39 heh 09:34:16 oomichi: if there's nothing else on this spec, I have another one 09:34:26 andreaf: thanks for explanation, I'd like to see it deeply later 09:34:37 andreaf: please 09:34:45 oomichi: for openstack-health #link https://review.openstack.org/233187 09:35:09 the idea is to formalise a bit the work which is already being done on openstack-health 09:35:22 is there any kind of roadmap to go with openstack-health? 09:35:25 but it would be nice to get everyone involved and also other folks opinion 09:35:50 dmellado: well only as far as the spec goes 09:36:14 I added a new openstack-health folder in the qa-specs repo 09:36:39 the idea is that people should be able to propose next pages / features for the dashboard in there via spec 09:36:40 andreaf: yeah, that is nice to add new folder 09:36:41 ty andreaf I'll have a look (althought I'm not quite savy on that, it took me even a while to get it installed xD) 09:37:17 andreaf: can we see openstack-health dashboard now? 09:37:23 I am not really involved in openstack-health now but it looks great and super powerful. Great potential 09:37:45 I keep saying that week after week, i know ! 09:37:50 jordanP: lol 09:38:38 andreaf: I cannot catch this topic. it is nice to get overview before seeing the spec 09:39:04 oomichi: ok 09:39:33 oomichi: well the idea is to build a dashboard that will help us visualise and debug issues in the various CI pipelines 09:39:40 andreaf, I am going to read that spec carefully 09:40:00 oomichi: we are starting with gate because we already have the test result data for that pipeline 09:40:22 oomichi: but the plan is to extend to periodic - to finally get some visibility and trackability on periodic tests 09:40:45 andreaf: already available site of the dashboard? like http://status.openstack.org/zuul/ 09:41:01 oomichi: not yet, but it will be linked in there 09:41:11 andreaf: that includes periodic jobs only or all 09:41:19 andreaf: ah, thanks. I got it 09:41:27 oomichi: there's WIP to write a pupper module for it and to deploy it in infra - it shouldn't be long 09:41:49 andreaf: +1 09:41:56 gmann: I guess it's to be defined, based on the data we have an what we see useful 09:42:03 looking forward to seeing it 09:42:19 there's a link in the spec with an etherpad of the work in progress patches 09:42:25 and things to be done 09:42:26 andreaf: ok 09:42:32 if anyone wants to contribute 09:42:55 #link https://etherpad.openstack.org/p/openstack-health-tracking 09:43:02 cool 09:43:34 that seems last stage(page) 09:43:43 oops js xD 09:43:44 nice progress :) 09:44:07 is there a demo about openstack-health?? 09:44:24 shuquan: you can download it and run it to check it 09:44:27 it'll go with mock data 09:44:28 there is a link for that, I'm trying to find it 09:44:35 (at least last time I checked) 09:44:45 oops, time is comming 09:44:56 oomichi: should we move forward? 09:45:01 can we move to the next topic? 09:45:33 sorry for my time management 09:45:41 oomichi: sure, np! ;) 09:45:44 #topic Tempest 09:46:06 I'd like to introduce current situation 09:46:15 compute service client dev is nice progress 09:46:25 almost done on https://etherpad.openstack.org/p/tempest-lib-service-client-migration 09:46:29 #link https://etherpad.openstack.org/p/tempest-lib-service-client-migration 09:46:41 glad if reviewing the remaining ;) 09:47:07 one different topic is "Generate API doc from temepst.log" 09:47:17 #link https://review.openstack.org/#/c/235092/ 09:47:26 ^^^ is a prototype 09:47:51 we will be able to get API doc from tempest.log for applying the same way for all projects 09:48:12 I will prepare a qa-spec for that 09:48:22 and glad to discuss it on the summit 09:48:26 +1 09:48:30 oomichi: +1, let me know if you need help on that 09:48:38 dmellado: thanks :) 09:49:04 we will have a lot of tasks for fixing docstring of service clients 09:49:17 so any helps will be very helpful 09:49:29 that is all from me 09:49:50 does anyone have anything to discuss on this topic? 09:49:55 about tempest? 09:50:46 let's move on 09:50:48 I wanted to know the current situation from the cleanup script with the accounts.yaml, but I'll review andreaf's about resources first 09:50:50 ;) 09:51:12 so let's move on 09:51:12 dmellado: cool ;) 09:51:19 #topic OpenStack Health 09:51:30 maybe this is done by the previous discussion 09:51:38 oomichi: yep 09:51:40 #topic Critical Reviews 09:51:56 anyone has critical reviews now? 09:52:21 I have https://review.openstack.org/#/q/status:open+project:openstack/tempest-lib+branch:master+topic:bp/migrate-service-clients-to-tempest-lib,n,z 09:53:03 oomichi, will review. That's an easy one :) 09:53:13 jordanP: thanks so much! 09:53:15 +1 09:53:26 dmellado: thanks:) 09:53:27 oomichi: cool, I'll have a look 09:53:33 also https://review.openstack.org/#/q/status:open+project:openstack/tempest+branch:master+topic:use_lib_service_client,n,z 09:53:43 andreaf: thank you also ;) 09:54:04 gmann: yeah, that also is easy one ;) 09:54:05 and these from previous list - https://review.openstack.org/#/c/234563/ & https://review.openstack.org/#/c/234568/ 09:54:30 #link https://review.openstack.org/#/q/status:open+project:openstack/tempest+branch:master+topic:use_lib_service_client,n,z 09:54:32 these 2 actually holding other approved one :) 09:54:41 #link https://review.openstack.org/#/q/status:open+project:openstack/tempest-lib+branch:master+topic:bp/migrate-service-clients-to-tempest-lib,n,z 09:54:41 oomichi: Thanks 09:55:01 more other reviews? 09:55:28 ok, let's move on 09:55:36 #topic open discussion 09:56:06 I have a plan to make a qa party Sunday before the summit 09:56:17 or Monday 09:56:25 cool 09:56:29 oomichi: nice idea 09:56:29 :D 09:56:31 awesome! 09:56:41 :) 09:56:55 thanks, I will ping you guys later ;) 09:57:07 one idea is https://www.google.co.jp/search?q=fishbowl+room&safe=off&hl=ja&biw=1276&bih=520&source=lnms&tbm=isch&sa=X&ved=0CAYQ_AUoAWoVChMIr5O6opDEyAIVQp6mCh0tLgpW#safe=off&hl=ja&tbm=isch&q=%E9%87%A3%E3%82%8A%E5%A0%80%E3%80%80%E5%B1%85%E9%85%92%E5%B1%8B 09:57:33 we can do fishing on restraunt 09:58:03 just one idea, and I'd like to get opinions about that later ;-) 09:58:20 that sounds fun! xD 09:58:27 I never fished in my life :) 09:58:35 I hope I'll still eat something 09:59:02 :D 09:59:05 jordanP: same here :) 09:59:06 jordanP: won't you pity the fish if you catch it? xD 09:59:34 probably 09:59:38 yeah, we need to consider it 09:59:47 it will great fun, Japanese style party :) 10:00:00 anyway, I'd like to get ideas later 10:00:04 ;) 10:00:11 sorry, the time comes 10:00:18 thanks all 10:00:22 #endmeeting