09:00:25 <gmann_> #startmeeting qa
09:00:26 <openstack> Meeting started Thu Mar 22 09:00:25 2018 UTC and is due to finish in 60 minutes.  The chair is gmann_. Information about MeetBot at http://wiki.debian.org/MeetBot.
09:00:27 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
09:00:29 <openstack> The meeting name has been set to 'qa'
09:00:43 <gmann_> who all here today?
09:02:16 <gmann_> seems no one, ll wait for couple of min
09:02:23 <gmann_> andreaf: chandankumar ping
09:03:28 <wznoinsk> hi there
09:03:40 <gmann_> wznoinsk: hi
09:04:37 <wznoinsk> gmann_, I have one item but not sure whether we have enough ppl to discuss
09:05:08 <gmann_> wznoinsk: sure
09:05:33 <gmann_> let's wait for 1 more min if more people join otherwise we will start
09:05:43 <wznoinsk> gmann_, ack
09:07:31 <gmann_> let's start
09:07:45 <gmann_> #link https://wiki.openstack.org/wiki/Meetings/QATeamMeeting
09:07:51 <gmann_> agenda for today ^^
09:08:13 <gmann_> i will try to cover minimum as not much people today
09:08:15 <gmann_> #topic Announcement and Action Item (Optional)
09:08:47 <gmann_> i had 1 action item from last meeting
09:08:54 <gmann_> gmann_ to have a priority tracking review topic link also somewhere.
09:09:23 <gmann_> i created the etherpad and we can start adding the gerrit link there for review
09:09:24 <gmann_> #link https://etherpad.openstack.org/p/qa-rocky-priorities-tracking
09:10:10 <gmann_> I will skip other agenda and directly jump to open
09:10:20 <gmann_> #topic Open Discussion
09:10:41 <gmann_> wznoinsk: you have something , please go ahead
09:11:15 <wznoinsk> gmann_, right, I found this feature to be very useful in debugging SDN controller environments https://review.openstack.org/#/c/553896/
09:11:24 <wznoinsk> gmann_, hence I'm proposing it
09:12:16 <wznoinsk> gmann_, I see a few typos to fix there but I hope the idea is clear
09:13:22 <wznoinsk> gmann_, I coupled it with a script to dump ovs flows etc as well as sosreport, result: https://s2dzmds67jf70fjqjwhpjw-on.drv.tw/failing_tempest_28.html
09:13:42 <wznoinsk> ups, ^ link is not working anymore
09:14:38 <gmann_> wznoinsk: i got the idea
09:14:49 <gmann_> wznoinsk: you want on teardown also?
09:15:23 <gmann_> wznoinsk: i got it for failure where you want to dump the state of cloud or resources of cloud etc
09:15:41 <wznoinsk> gmann_, originally it was only on_failure but I realized, when I had to troubleshoot resource leaks, the test were actually passing but leaving a lot of rubbish behind, hence run_on_teardwon was useful
09:17:33 <wznoinsk> gmann_, I'm struggling to share the html results in public, give me a sec
09:17:36 <gmann_> wznoinsk: we have cleanup CLI to cleanup the leaked resources which can be run as dry run before deleting test leaked resources
09:17:42 <gmann_> wznoinsk: sure
09:18:08 <gmann_> wznoinsk:  its office hour so take your time :)
09:27:57 <gmann_> wznoinsk: while you get the results link, does cleanup CLI fulfill the teardown case ?
09:29:16 <gmann_> wznoinsk: because i was thinking, after teardown either we have leaked resource or whatever cloud state is so we do have  things for both, cleanup CLI can clean leaked resources and cloud state can be checked by openstack CLI
09:29:52 <gmann_> wznoinsk: but anyways those kind of discussion we can do on patch also, it does not need to be here
09:29:54 <wznoinsk> gmann_, not in 100% as one of the reasons I have the leftover resources is there were 2 or 3 same routers created and I wouldn't see that straight away if I'd do it at cleanup CLI only
09:30:37 <wznoinsk> gmann_, the results are https://drive.google.com/file/d/1cnsbolA4n0uhBXZH_pvDcEDCSg26rdGI/view?usp=sharing but you have to download and open it from local machine , I'll try to figure out a better place
09:30:43 <gmann_> wznoinsk: you can run cleanup with dry run and check all resources created by test
09:31:32 <gmann_> wznoinsk: sure thanks, failure case looks good for me. till script is ready only and not destructive :)
09:31:32 <wznoinsk> gmann_, ok, I need to get famliar with the cleanup then but it sounds like yes it would be good for my leftovers case
09:32:28 <gmann_> wznoinsk: ok, and further discussion we can do on gerrit. idea looks good to me
09:32:43 <wznoinsk> gmann_, yes, I feel like I'm crossing these tempest boundaries by proposing something like that, although proven to be very useful - especially with tests like east_west where you have 4 combinations of them running in one go (there's no way to --pdb them)
09:33:47 <wznoinsk> gmann_, good to hear!... i'l put more use cases in https://blueprints.launchpad.net/tempest/+spec/run-on-teardown-failure then
09:33:55 <gmann_> i understand that, as long as it is configurable it does not harm and people setting script know what they are doing
09:34:02 <gmann_> wznoinsk: cool
09:34:30 <wznoinsk> gmann_, good discussion, thanks
09:34:44 <gmann_> wznoinsk: we do have process of spec if there is BP
09:35:19 <gmann_> wznoinsk: but i do not think we need in this case until it need more discussion on gerrit patch.
09:36:13 <wznoinsk> gmann_, sure, do I fill all pros/cons in the BP ^ tho?... I have a feeling it would convince ppl a bit faster ;-)
09:36:44 <gmann_> wznoinsk: ok then you can add spec for that and all details can go there
09:37:20 <gmann_> wznoinsk: here #link https://github.com/openstack/qa-specs/tree/master/specs/tempest
09:38:06 <gmann_> and this is the template which is small one i think #link https://github.com/openstack/qa-specs/blob/master/template.rst
09:38:17 <wznoinsk> gmann_, could I leave the 'more info' in a form of LP blueprint and spawn a spec only if there's really more discussion?
09:38:51 <tosky> (or better: http://git.openstack.org/cgit/openstack/qa-specs/tree/template.rst )
09:39:13 <gmann_> wznoinsk: ok, i am ok with that. if we need more design discussion then we can propose spec
09:39:36 <wznoinsk> gmann_, thanks
09:39:40 <tosky> gmann_: wouldn't it be better to limit LP blueprint? Because they will lost after (because it will happen ^_^) the storyboard migration
09:39:47 <gmann_> tosky: thanks. my browser typo of git* give me that but yea git. better
09:40:23 <gmann_> tosky: yea but it is just specless BP till now nothing much
09:40:26 <openstackgerrit> Waldemar Znoinski proposed openstack/tempest master: run a shell command on a test failure (at teardown stage)  https://review.openstack.org/553896
09:40:49 <gmann_> tosky: we will have lot other history data lost for that migration
09:41:28 <gmann_> tosky: but yea we mostly need spec if need discussions or debates
09:42:28 <slaweq> andreaf: hi, I have a question about https://review.openstack.org/#/c/552846/13/.zuul.yaml
09:42:45 <gmann_> anything else for discussion ?
09:43:03 <slaweq> andreaf: I switched from "all-plugin" to "all" as You adviced but now tests fails: http://logs.openstack.org/46/552846/14/check/neutron-tempest-plugin-scenario-linuxbridge/060407d/job-output.txt.gz#_2018-03-22_09_05_39_855262
09:43:25 <slaweq> did I something wrong there? should I maybe use different regex with "all"?
09:44:22 <tosky> gmann_: I have few questions about zuulv3 and multinode, but I guess I would need andreaf too
09:44:34 <tosky> my multinode job is working but I see few potential issues
09:44:50 <gmann_> tosky: sure, we can discuss later too once he is up.
09:45:12 <gmann_> tosky: please put link, i will check if get time tomorrow
09:47:25 <tosky> gmann_: I explained the issues on Monday, so I have the chat logs!
09:47:28 <tosky> it starts here http://eavesdrop.openstack.org/irclogs/%23openstack-qa/%23openstack-qa.2018-03-19.log.html#t2018-03-19T11:25:26
09:47:37 <tosky> and then http://eavesdrop.openstack.org/irclogs/%23openstack-qa/%23openstack-qa.2018-03-19.log.html#t2018-03-19T15:18:43
09:47:45 <gmann_> tosky: thanks, ll check
09:48:23 <gmann_> anything else or we close the office hour.
09:49:05 <gmann_> #endmeeting