13:30:59 #startmeeting qa 13:30:59 Meeting started Tue Jan 14 13:30:59 2020 UTC and is due to finish in 60 minutes. The chair is gmann. Information about MeetBot at http://wiki.debian.org/MeetBot. 13:31:00 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 13:31:02 The meeting name has been set to 'qa' 13:31:14 who all here today ? 13:31:30 Hello Doug is here 13:31:42 o/ 13:31:42 hi o/ 13:32:53 hello everyone. let's start 13:32:56 #link https://wiki.openstack.org/wiki/BugTriage 13:33:05 #link https://wiki.openstack.org/wiki/Meetings/QATeamMeeting#Agenda_for_next_Office_hours 13:33:09 this is agenda ^^ 13:33:19 #topic Announcement and Action Item (Optional) 13:34:02 we have two AI from previous meeting. 13:34:07 1. kopecmartin need to check and report the unit test coverage for cleanup CLI. 13:34:16 2. gmann to add the community goal tracking for QA in office hour agenda. 13:34:41 i have not done that at least. I will do that today so re-adding it. 13:34:53 #action gmann to add the community goal tracking for QA in office hour agenda. 13:35:04 kopecmartin: how about you? 13:35:12 gmann: i did checked the numbers and tempest/cmd/cleanup.py has unit test coverage == 25% and cleanup_service.py 90% 13:35:41 ok. 13:36:24 cleanup_service.py coverage can be improved, but I'm afraid that unit tests for cleanup.py would be too complicated , the code contains basically for loops and overall controll of the whole cleanup process, it would be hard to test it by unit tests i think 13:36:46 cleanup.py should be tested by functional tests in my opinion 13:37:07 we have a zuul role for that up fo review, so we can start by that :) 13:37:20 that's all to my action item 13:38:07 kopecmartin: i agree but we can at least add unit test for its arg etc so that changing them or their behavior can be caught 13:38:25 i remember we broke those with very silly changes. 13:38:32 gmann: ok, that sounds reasonable 13:39:37 kopecmartin: we have BP to track the CLI unit tests. 13:39:39 #link https://blueprints.launchpad.net/tempest/+spec/tempest-cli-unit-test-coverage 13:40:15 I will assign this to you now and also feel free to add notes in that. 13:40:38 done. 13:40:56 #topic Ussuri Priority Items progress 13:41:11 #link https://etherpad.openstack.org/p/qa-ussuri-priority 13:41:58 first two items from masayukig who is not here. 13:42:04 RBAC testing strategy (reader/member/admin X project/domain/system) 13:42:20 i have not reviewed this yet. This is priority item for us. 13:42:44 I should be spending some time in this week for this. 13:42:45 we are trying to get patrole running but it has resource leaks 13:42:57 patrole reviews https://review.opendev.org/#/q/owner:rick.bartra%2540att.com+status:open+project:openstack/patrole 13:43:03 to fix some of them 13:43:37 bigdogstl: yeah, i have reviewed most of them i think 13:44:00 bigdogstl: gate should be fin now as Rick fix for admin job is merged. 13:44:18 we will come into that during patrole section 13:44:20 yes thank you for helping him out on that 13:44:26 ok 13:44:30 next item is- py2.7 drop plan for QA 13:44:50 this is in progress for devstack and not started for Tempest. 13:45:02 Tempest will be during start of Feb as per TripleO request. 13:45:07 CI 13:45:43 skipping few as no progress on them. 13:45:45 Improving Tempest cleanup 13:45:56 kopecmartin: is working on that. 13:46:15 - Bug triage status report before office hour 13:46:30 kopecmartin: did I send you the etherpad links for this ? 13:46:35 i think no.. 13:46:49 no, at least i haven't receive them 13:47:09 ok. 13:47:17 I will do right after meeting 13:47:44 #topic OpenStack Events Updates and Planning 13:48:17 next event is Vancouver summit/PTG. I have not started the planning. I will do during April end or May. 13:48:40 #topic Sub Teams highlights (Sub Teams means individual projects under QA program) 13:48:44 Tempest 13:49:07 #link https://review.opendev.org/#/q/project:openstack/tempest+status:open 13:49:12 these are open review. 13:50:06 is this the moment when we try to highlight some reviews, or is it for the open discussion part, as the agenda seems to hint? 13:50:31 efried fix is merged to explicit telling from tests to create the isolated network for create server request - https://review.opendev.org/#/c/701792/ 13:50:42 tosky: you can if it is Tempest 13:51:01 just a reminder about a new scenario (mostly volume) test: https://review.opendev.org/#/c/695967/ 13:51:24 i'd like to bring this up https://review.opendev.org/#/c/692375/, it's very important 13:51:33 Skip adding host to aggregate with az if host already in other zone https://review.opendev.org/#/c/673342/ 13:52:24 tosky: ok, i think cinder tempest plugin test patch is abandon now so i can review on Tempest one. I was waiting as same tests was added in both place 13:52:43 gmann: yes, the other patch was abandoned some time ago 13:52:51 kopecmartin: sorry about that. I will check that today for sure. 13:53:05 gmann: thanks 13:54:10 bigdogstl: is host with AZ detected automatically or with config ? 13:54:25 i have not looked into review comments in detail. seems masayukig left some comment 13:55:03 I will have Rick address them 13:55:55 ok. adding that in review list 13:56:09 anything else on Tempest ? 13:56:38 Not at this time most of our effort is in patrole 13:56:57 Patrole. 13:57:01 bigdogstl: go ahead. 13:57:08 patrole reviews https://review.opendev.org/#/q/owner:rick.bartra%2540att.com+status:open+project:openstack/patrole 13:57:16 agenda is old for Patrole. I will update for next meeting 13:57:56 bigdogstl: i tried to review almost all open and zuul happy patches. 13:58:08 one i need to investigate more is - #link https://review.opendev.org/#/c/676979/ 13:58:52 i suspect keeping tenant_id in cred during cleanup can have risk of stale tenant_id 13:58:55 Ok I will have Rick review the comments and keep pushing them 13:59:53 he added the investigation on bug and review and the workaround also. 14:00:49 which I need to check in details and find some mechanism if we can restore the user_id and tenant_id during self._initial. 14:01:20 not cleaning them during clean_cred() is something does not sounds good. 14:01:39 anyways, i will add comment on bug and review. 14:02:30 ok 14:02:31 other item in patrole is. I missed to add the Train job for Patrole. 14:02:34 #link https://review.opendev.org/#/c/701271/ 14:02:49 https://review.opendev.org/#/q/topic:negative_multiattach_swap_volume_tests is ready for review if anyone has time, not sure if this is the correct place in the meeting to bring this up. 14:02:52 this is somehow failing which means some policies might have changed. 14:03:17 if you get a chance, can you please take a look at https://review.opendev.org/#/c/692233/ again. It is for the volumes and volume snapshots, I replied back to @gmann comments 14:03:37 lyarwood: thanks. this is correct time and place to bring. I am adding those in list. 14:04:03 correction: volumes and volume snapshots resource leaks* 14:04:46 gmann: wonderful thanks 14:04:56 rick_bartra: create_test_server does not delete the volume. 14:05:18 gmann: can I also bring up https://review.opendev.org/#/c/702089/ for devstack-gate, final bit of n-cauth clean up for stable/rocky 14:06:40 @gmann if it doesn't I am not exactly sure where the volume is deleted. I will need to dig further, but do know the volume is deleted when the volume snapshot is cleaned up in the test 14:07:12 lyarwood: +2. cellv1 is legacy job which i did not check. all other job since rocky use devstack service things 14:08:18 rick_bartra: ok. because once you fix is merged I am thinking to move the volume cleanup part in create_test_server becasue we can easily detect the created volume there for volume backed server request 14:08:51 and we can do that during class level so we do not need to worry about if tests need those volume and should be cleanup after detach etc. 14:09:09 gmann, sounds good. I will find where it is happening and reply back to the change review 14:09:53 rick_bartra: thanks. on L308 comment. did you find that after running single tests or set of tempest set. 14:11:04 gmann, I found it with both. We found resource leaks running downstream, then when I ran this test by itself the same leaks happened 14:12:16 ok. kopecmartin does your patch of adding cleanup CLI call in gate job log the leaked resource somewhere ? 14:12:57 rick_bartra: kopecmartin is adding the cleanup CLI in upstream job which should be able to detect all these resource leak. 14:13:19 #link https://review.opendev.org/#/c/698610/ 14:13:33 gmann: yes, it should be printed out to the console 14:14:00 kopecmartin: ok and what about init saved file ? 14:14:04 gmann, thanks will check it out 14:14:09 Should leaks cause the gate to fail? 14:14:24 or at least a gaol to head that way? 14:14:42 init_saved state is created before tempest execution https://review.opendev.org/#/c/698610/8/playbooks/devstack-tempest.yaml and after tempest execution tempest cleanup --dry-run is called 14:14:50 bigdogstl: not initially till we fix all and it pass everything. but later we should fail and detect the leakage during addition of tests itself 14:15:00 but the role allows to change dry run to actual run, it depends on the needs 14:15:04 bigdogstl: yeah. that is overall goal. 14:15:05 sounds great thanks 14:16:18 kopecmartin: yeah, can we log init_saved state also as part of that role. it should be few resource in that which devstack create. 14:16:56 kopecmartin: let me check the review in detail and then i can comment exact logging i am thinking. 14:17:08 rick_bartra: for that volume things. let me check locally also. 14:17:17 thanks for fixing all those rick_bartra 14:17:27 if we moved away from Patrole, there is a Python 3 fix for subunit2html: https://review.opendev.org/#/c/700778/ - which goes back to the discussion about moving subunit2html somewhere else 14:17:30 gmann, thank you for all the reviews 14:17:38 moving next... 14:18:31 tosky: somewhere else ? 14:18:57 i did not see comment about moving 14:19:15 I mean, from the topics; first Tempest, then Patrole, then... 14:19:16 :) 14:19:38 ohk :) 14:20:10 there was discussion of moving subunit2html somewhere else during PTG - L20 - https://etherpad.openstack.org/p/qa-ussuri-priority 14:20:20 the sentence matches that topic :) 14:20:55 grenade 14:21:11 this to check for grenade zuulv3 native base jobs 14:21:13 erm, I'm guilty; I planned to look and the failures yesterday 14:21:34 but then the world feel apart and I moved to other things; I still plan to check and fix the jobs this week 14:21:41 tosky: you mentioned plugin things is failing for that job ? 14:21:43 so status as last week 14:21:49 phk 14:21:51 phk 14:21:52 ohk 14:22:15 tosky: thanks again for working on those. 14:22:16 yes; both testing jobs which involves plugin are failing, so I guess there is a more general issue somewhere; something that happened in the last months 14:22:38 I will keep everyone updated when I manage to make them working again 14:22:56 cool. 14:23:10 #topic Bug Triage 14:23:36 i have not prepared the etherpad for kopecmartin and sonia who volunteer for bug triage. 14:24:06 but we have 49 New bug in Tempest. #link https://bugs.launchpad.net/tempest/+bugs?search=Search&field.status=New 14:24:53 which is too much i think and we need to triage them on priority. I will try to keep 30 min for bug triage from next office hour. 14:25:32 #topic Critical Reviews 14:25:45 any other critical review than we already discussed ? 14:26:21 #topic Open Discussion 14:26:27 4 min left if any other topic 14:27:53 if nothing. let's close it. 14:27:59 thanks everyone for joining. 14:28:05 #endmeeting