07:58:44 #startmeeting qa 07:58:45 Meeting started Thu Dec 14 07:58:44 2017 UTC and is due to finish in 60 minutes. The chair is gmann. Information about MeetBot at http://wiki.debian.org/MeetBot. 07:58:46 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 07:58:48 The meeting name has been set to 'qa' 07:58:54 who all here today 08:00:59 * gmann hoping people are not on Christmas vacation 08:01:22 let's wait for another 5 min 08:01:25 o/ 08:01:29 andreaf: hi 08:01:36 hi good morning 08:01:41 morning 08:01:57 o/ 08:02:07 gmann: thanks for following up on the cinder v3 stuff, I'll try to review that later as well 08:02:08 felipemonteiro_: hi 08:02:29 andreaf: thanks, i will push devstack patch also meanwhile. 08:02:48 let's start the meeting 08:03:00 #topic Previous Meeting Action review 08:03:11 no AI from last meeting 08:03:21 #topic Office Hours Summary 08:03:31 #link http://eavesdrop.openstack.org/meetings/office_hours/2017/office_hours.2017-12-07-09.00.html 08:03:39 we did around 10 bug triage 08:04:13 we have 8 new bugs in current situation 08:04:53 we have 7 high importance bugs seems and most of them triaged 08:05:18 let's discuss those in next office hour in case they are just stuck somewhere due to review/feedback etc 08:05:36 anything else on office hour thing? 08:05:49 does folks feel they are useful? 08:06:06 I'm happy with them so far 08:06:35 yea me too 08:06:39 but I wonder if we should keep it as is, or make it weekly 08:07:00 and i am looking for extending it to more than just bug triage 08:07:11 and replace the meeting eventually 08:07:19 yeah that's what I was thinking 08:07:37 but at the end it takes the full hour just to do some bug triage 08:07:43 andreaf: we can do that but let's make it more broad so that people gets used to discuss more things there 08:08:00 after that we can decide on weekly basis 08:08:43 gmann: andreaf \o/ 08:08:48 gmann: well in principle office hours are open already for people to join in and discuss about anything they might want to discuss 08:08:53 chandankumar: o/ 08:09:11 yea 08:09:27 weekly makes sense 08:09:35 but start from new year 08:09:41 let's see for few more months or at least till queens to replace meeting 08:10:21 andreaf: how about alternate TZ office hour ? did you decided on that? 08:11:10 gmann: heh that would replace the current meeting basically 08:11:22 gmann: I think it would make sense since there's not much attendance to meetings anyways 08:11:47 i feel office hour is good in that TZ too in case people want to join and discuss 08:12:07 at end we just replace both meeting with office hour with same time as meeting used to be 08:12:23 gmann: but I guess it's only good if people attends :) 08:12:46 if not replacing meeting for office hours does not necessarily help :) 08:13:10 yea so we can just start and finish in case no attendee as it is on qa channel only 08:13:11 keep the meeting as it is and make the office hour weekly 08:13:28 office hours is 9 utc? 08:13:29 chandankumar: than its too much time on meetings :) 08:13:33 so... as a huge huge noob 08:13:52 andreaf: i mean we gradually replace meeting only and at end only office hours in both TZ 08:13:59 out of curiosity, what is the sort of thing one would bring to office hours for discussion 08:14:02 mguiney: yea :) 08:14:33 mguiney: anything people want to dicsuss, bugs, fatures, help, feedback etc 08:14:37 because that's my main hesitation on that front- a sort of "i dont even know what do" 08:14:43 * mguiney nods 08:14:52 cool. thank you for the clarification 08:15:43 mguiney: it is kind of open things to discuss anything on need basis and its on qa channel to get most of the qa people there 08:16:12 * mguiney nods 08:16:17 ok, thank you :) 08:16:17 my idea is that with office hours being more informal (no fixed agenda) it may feel more open / welcome for people to pop in and ask questions / help if they want 08:16:30 yea. +1 08:16:42 there will not be any agenda on that 08:16:46 that's something people can do at any time in this channel anyways, but during office hours we guarantee that some cores will be in 08:17:46 makes sense. would it be a good place for people to pop up if they were looking to dip a toe in a little more, so to speak, or would that be better suited to another... platform? venue? 08:17:56 (i dont quite know the correct wording) 08:17:57 andreaf i was thinking to change time of current office hour to 8 UTC. it is little late in office sometime 08:18:42 andreaf: but i am favor of having office hour in both TZ 08:19:02 mguiney: yeah that would be fine for a good place to get started 08:19:12 * mguiney nods 08:19:31 cool. again, sorry to ask, just didn't know 08:19:43 gmann: uhm earlier than 8 UTC I'm definitely not going to make it :P 08:20:00 andreaf: yea it is little early for you 08:20:23 8 AM na ? 08:20:32 gmann: but if you masayukig and zhufl are around it should be fine 08:20:35 * gmann cannot attend 7 AM API-WG meeting due to early morning 08:20:41 andreaf: ok 08:21:21 andreaf: ok, let's make it 8 UTC. 08:21:37 let's move next 08:21:47 #topic Tempest 08:22:08 1 things i want to bring about cinder testing 08:22:28 we have 2 approach here to achieve the v3 API testing as default in gate 08:22:29 +1 that's pretty urgent 08:22:41 1. https://review.openstack.org/#/q/topic:disable_volv2+(status:open+OR+status:merged) 08:22:52 2. https://review.openstack.org/#/q/topic:volume-default-testing-v3+(status:open+OR+status:merged) 08:23:15 1st from smcginnis and 2nd one i just pushed today 08:23:43 on second one, i need to push devstack patch and project-config patch to make v2 job 08:23:53 feedback are welcome on those 08:24:20 gmann: yeah I will try to review them today 08:24:21 v2 API is going to be removed from cinder and v3 will be the current one 08:24:26 andreaf: thanks 08:24:33 we need to switch main API testing to v3 as soon as possible 08:24:41 yea +1 08:24:49 on master at least 08:25:22 gmann: i am not sure this also needs to be changed https://github.com/openstack/tempest/blob/master/tempest/cmd/verify_tempest_config.py#L241 08:25:28 my idea is to do like same way we do for compute 08:25:32 using v3 to get extension list 08:25:52 chandankumar: yea we will cleanup those, first we get the agreed direction 08:26:01 gmann: thanks :-) 08:27:18 * gmann hope andreaf does not notice lib interface change in 2nd approach which are just version change but backward incompatible 08:27:38 * andreaf will definitely not notice now :D 08:27:43 anyways i will make that backward compatible if we agree on direction 08:27:49 :) 08:28:04 let's move next 08:28:06 Zuul V3 patches 08:28:17 #link https://review.openstack.org/#/q/status:open+project:openstack/tempest+branch:master+topic:zuulv3-tempest 08:28:22 ^^ all these are merged 08:28:42 andreaf: your turn, you are doing lot of work around that 08:29:12 yeah well not much left on tempest though I have a patch series in devstack 08:30:02 #link https://review.openstack.org/#/q/status:open+branch:master+topic:test_matrix_role 08:30:14 #link https://review.openstack.org/#/q/status:open+project:openstack-dev/devstack+branch:master+topic:get_devstack_logs 08:30:24 thanks 08:30:39 What I want to achieve here is feature parity between new style "dsvm" jobs and legacy one 08:30:48 so we can then switch over 08:30:52 1 question, we will keep general tempest job on project config side or we need to move to tempest ? 08:31:16 I will be on tempest side I think 08:31:22 that's what I'd like to see at least 08:31:41 unless there is any specific concern with folks to maintain that 08:31:46 yea, i was hoping that 08:32:10 but at least many of basic can be, other like ceph etc can be on ceph repo side 08:32:42 so in terms of single node job the things left to achieve parity now are mostly in the area of missing log files and cleanup of console log 08:33:11 k 08:33:22 and conf files etc also 08:33:25 but there are two other areas which need attention still 08:33:48 one is multinode which I'm working on with jeblair, since in multinode case we need to be able to set different configs on different nodes 08:34:10 for the list of base services that's partly handled by the test matrix 08:34:19 but there's more to it 08:34:42 the logic is currently in d-g - basically in there we say now if node is subnode then... 08:35:17 but using ansible and having all nodes in the inventory it would make sense to have something like host/group vars to control node specific config / behaviour 08:35:59 another big piece which is missing once this is done is support for grenade in zuulv3 native jobs 08:36:23 I haven't even started to look into that tbh, so if anyone wants to do so it would be nice :) 08:37:32 another thing that is missing which is the reproducer.sh script which I think is a very handy thing to have but that's something for the infra team really 08:38:01 on the zuulv3 side mordred has been working on this series 08:38:06 #link https://review.openstack.org/#/q/status:open+project:openstack-dev/devstack+branch:master+topic:devstack-functional functional tests 08:38:36 which puts some base functional jobs for libraries into devstack 08:39:02 since they are used by several parties and there is no better place to host them :) 08:39:27 yea, devastack become better place 08:40:34 thanks, andreaf i will try to catch the series but have to travel to india office immediately till new year 08:40:48 ok I think I have been rumbling long enough on this :) 08:41:08 no, good information on those ongoing things 08:42:08 thanks for all your dedicated work on this 08:42:15 anything else on this 08:42:35 not from me 08:42:52 k, let's move next 08:43:00 #topic Critical Reviews 08:43:16 any other critical reviews to get in on priority 08:43:35 gmann: andreaf https://review.openstack.org/#/c/524176/ i am not sure what else i need to improve to make it working 08:45:11 chandankumar: what is issue in that, sorry i did not follow on that much 08:45:26 chandankumar: that does not the POST_FAILURE issue 08:45:51 gmann: andreaf one is post failure issue 08:46:13 chandankumar, gmann: the job uses now the base tox job, which at the end of the run looks for test results 08:46:31 in this patch i have fixed removing installed packages after sanity run 08:46:38 it finds a .testrepository folder, so it tries to extract test results from it but no test runs so it fails 08:47:27 chandankumar: yeah I will review it, it's a nice improvement 08:47:42 chandankumar: but we really need to get the POST_FAILURE solved 08:48:55 andreaf: i am not sure how to fixed that part and also once sanity check passes for a project it is not getting appended there 08:49:03 there i also need help 08:49:11 failed one is working fine 08:49:58 chandankumar: I'm looking at the log but it's very long :D hard to manage 08:50:19 chandankumar: perhaps we should split logs into individual plugin ones 08:50:20 listing all tests 08:50:47 andreaf: you mean project wise? 08:50:51 chandankumar: like log the plugin check to a dedicated file for each plugin (install, list, cleanup) 08:51:13 chandankumar: yeah we could have a subfolder and a log file per plugin in there or so 08:51:20 andreaf: sure, i will do that 08:51:25 chandankumar: and in the main log just one line per plugin 08:51:35 andreaf: i will update the patch 08:51:39 +1 for subfolder otherwise it will be huge number files at top dir 08:51:42 chandankumar: because it's really hard to see which plugin passed and which failed 08:52:34 chandankumar: but otherwise the job seems to do its work 08:52:40 networking-plumgrid plugin is failing now 08:53:01 andreaf: i will take a look on that. 08:53:25 so it's identifying failures which is good 08:53:26 chandankumar: ok thanks 08:53:28 On tempest plugin split goal side: octavia is at risk only 08:53:37 All are in controle 08:53:40 ok, let's jump to open 08:53:42 #topic Open Discussion 08:53:52 ironic guys are working on tempest plugin migration this week 08:53:53 the POST_FAILURE is kind of masking the failure in RUN though so it would be good to fix that 08:53:57 chandankumar: for review only right ? 08:54:08 chandankumar: cool, thanks for the update and work on that 08:54:23 gmann: yes for review and syncing the code 08:54:33 chandankumar: link, i can help there 08:54:49 gmann: https://review.openstack.org/#/q/project:openstack/octavia-tempest-plugin 08:55:28 thanks ll check 08:55:40 gmann: andreaf ironic tempest plugin stuff is tracked here https://etherpad.openstack.org/p/ironic-tempest-plugin-migration 08:55:47 #link https://review.openstack.org/#/q/project:openstack/octavia-tempest-plugin octavia tempest plugin 08:55:50 gmann: thanks for taking care of ec2-api one 08:56:05 #link https://etherpad.openstack.org/p/ironic-tempest-plugin-migration ironic tempest plugin 08:56:10 chandankumar: np, waiting for project config to give the repo 08:56:22 gmann: i need to mail PTL once again then 08:56:29 gmann: he is not fan of irc 08:56:39 ah yea need PTL +1 08:56:44 chandankumar: thanks 08:56:50 2 min left, any other things to discuss 08:57:03 nope from my side. 08:57:39 let's close the meeting 08:57:46 3 ... 08:57:50 2 .. 08:57:51 I will have a kind of long christmas break 08:57:57 1 . 08:58:01 from before xmas till Jan 9th 08:58:09 andreaf: ok have a nice break and enjoy 08:58:20 gmann: thanks :) 08:58:28 thanks all for joining meeting 08:58:31 #endmeeting