13:00:01 <alex_xu> #startmeeting nova api
13:00:02 <openstack> Meeting started Wed Aug 23 13:00:01 2017 UTC and is due to finish in 60 minutes.  The chair is alex_xu. Information about MeetBot at http://wiki.debian.org/MeetBot.
13:00:03 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
13:00:06 <openstack> The meeting name has been set to 'nova_api'
13:00:11 <alex_xu> who is here today?
13:00:21 <gmann_> o/
13:00:41 <stephenfin> o/
13:00:44 <cdent> o/
13:01:19 <mriedem> o/
13:01:26 * mriedem leaves in 15 minutes
13:01:56 <alex_xu> ok I think we can start the meeting
13:02:07 <alex_xu> #topic PTG in denver
13:02:26 <alex_xu> I think we can begin to think about the thing for next release
13:02:40 <alex_xu> #link https://etherpad.openstack.org/p/nova-ptg-queens
13:03:04 <alex_xu> mriedem already add some API deprecation in the etherpad
13:03:21 <alex_xu> gmann_ also adds no-more-extension-in-queens in the etherpad
13:04:47 <alex_xu> except those, only thing left in my mind is the api concept doc, I remember there still have some todo at there
13:05:18 <gmann_> alex_xu: ah yea, we did not give time for those
13:05:58 <alex_xu> yea, those are not easy
13:06:12 <mriedem> i don't know if we'll do separate microversions for those proposed deprecations or not
13:06:29 <mriedem> it's kind of nice to not have so many microversions, but they are free
13:07:04 <alex_xu> yea, and we call them 'micro'version
13:07:32 <gmann_> but yea we can do all together  also
13:08:07 <mriedem> alex_xu: another thing,
13:08:23 <mriedem> can you write a spec for the stuff you're doing with filtering allocation candidates by traits?
13:08:28 <mriedem> since that's going to be an api change
13:08:37 <alex_xu> mriedem: yea, sure
13:08:40 <mriedem> thanks
13:08:44 <alex_xu> np
13:09:26 <gmann_> i remember one more item we discussed sometime long back
13:09:36 <gmann_> to cleanup the hide address things.
13:09:42 <alex_xu> ah yea
13:09:43 <gmann_> i need to add spec for that
13:10:22 <alex_xu> this morning, we also talk about how to test the bug like this https://bugs.launchpad.net/nova/+bug/1658571
13:10:23 <openstack> Launchpad bug 1658571 in OpenStack Compute (nova) "Microversion 2.37 break 2.32 usage" [High,Fix released] - Assigned to Artom Lifshitz (notartom)
13:11:07 <mriedem> the only way i know is to run existing tests against the latest microversion, like subclass the tests
13:11:15 <mriedem> and override as needed for new behavior
13:11:36 <mriedem> i did some of that with 2.53
13:11:43 <alex_xu> yea
13:11:52 <alex_xu> #link https://review.openstack.org/#/c/430352/
13:11:54 <gmann_> yea, we need more refactoring and capping of microversion for that
13:11:56 <alex_xu> there is a patch
13:12:32 <mriedem> with that, i have to go and take my daughter to school, ttyl
13:12:41 <alex_xu> gmann_: what kind of refactoring, I guess there should be a lot of work on capping microversion for the tests won't work for lastest version
13:12:51 <alex_xu> mriedem: yea, ttyl
13:13:43 <alex_xu> yea, subclass is also a way
13:13:45 <gmann_> alex_xu: like if request schema changed in x mocroversion, then that tests needs some common way to select the request sample etc
13:13:58 <gmann_> same for response change things also.
13:14:51 <gmann_> but that will depend case by case
13:15:03 <alex_xu> gmann_: you mean seelct the x version sample file when running under lastest version?
13:15:34 <gmann_> yea
13:16:05 <gmann_> but that can be handle more with version cap
13:16:22 <alex_xu> gmann_: or is it worth we add some tools like cap microversion, then use that in the new tests
13:17:04 <alex_xu> I remember I reviewed on of mriedem patch which sublcass the oldversion test class. But there are a lot of pain on the review
13:17:16 <gmann_> alex_xu: that is much on same line with tempest microversion framework
13:17:56 <gmann_> alex_xu: ok, i ll check that, have not looked
13:18:56 <alex_xu> #link https://review.openstack.org/#/c/471899/25/nova/tests/unit/api/openstack/compute/test_hypervisors.py
13:19:06 <alex_xu> gmann_: ^ I remember this is hard to review
13:19:13 <gmann_> alex_xu: thanks
13:20:32 <alex_xu> gmann_: that subclass rewrite a lot of tests.
13:20:39 <gmann_> alex_xu: that is nice but might not cover all things like everything working with latest version
13:21:06 <cdent> it is a bit hard to tell what’s going on in some of those
13:22:37 <alex_xu> gmann_: any specific case which not cover?
13:23:06 <alex_xu> gmann_: any specific case which not cover?
13:23:30 * alex_xu resend the message, since he saw gmann_ drop the network...
13:23:57 <alex_xu> ah, there are gmann_ and gmann__ and gmann  :)
13:24:20 <gmann__> alex_xu: say version 2.x changing network API and we run all exiting tests of 2.x-1
13:24:21 <alex_xu> cdent: yea, super pain when review that patch
13:24:22 <gmann__> :)
13:25:00 <gmann__> but if 2.x accidentally change the other API schema (with changing common stuff in parameter_type)
13:25:18 <alex_xu> gmann__: ah, right
13:26:31 <gmann__> alex_xu: it is very hard to achieve 100% coverage for all version (one of pain point of microversion)
13:26:41 <gmann__> but with running all tests with latest version is maximum we can cover but hard to achieve that may be
13:26:54 <alex_xu> yea
13:27:38 <alex_xu> maybe it is worth to look at if the lastest version test scenario + cap version in the new microversion
13:27:47 <gmann__> but we can think other ideas also like you had of generating json schema sample etc
13:28:45 <alex_xu> I'm not sure we have enough bandwidth to convert all the tests in shortterm, I guess that will spend us many release again...
13:28:58 <gmann__> yea :(
13:29:04 <alex_xu> gmann__: in the talk this morning, I feel that doesn't help with that bug
13:29:06 <cdent> yeah, better would be something we can do just going forward
13:29:51 <gmann__> i will spend more time in next week (hope i finish my PoC this week) and let's see if soemthing comes up
13:30:02 <gmann__> cdent: +1 something is better than nothing
13:30:43 <cdent> I’ve tried to think if there is some automagic solution that would involve gabbi and changing environment variables
13:30:53 <cdent> but I haven’t had a chance to think about it properly
13:31:23 <alex_xu> gmann__: cool, thanks
13:31:37 <alex_xu> cdent: +1 also
13:34:16 <alex_xu> ok, I guess that is all I have
13:34:26 <cdent> I’ve got nothing
13:34:32 <gmann__> me too.
13:35:09 <alex_xu> ok, cool, let us end the meeting, talk about those things next week again, hope we get better idea
13:35:14 <alex_xu> thanks all!
13:35:15 <gmann__> alex_xu: thanks.
13:35:23 <alex_xu> #endmeeting