16:01:26 <markvoelker> #startmeeting defcore
16:01:26 <openstack> Meeting started Wed May 25 16:01:26 2016 UTC and is due to finish in 60 minutes.  The chair is markvoelker. Information about MeetBot at http://wiki.debian.org/MeetBot.
16:01:27 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
16:01:30 <openstack> The meeting name has been set to 'defcore'
16:01:33 <hogepodge> o/
16:01:33 <markvoelker> #chair hogepodge
16:01:34 <openstack> Current chairs: hogepodge markvoelker
16:01:37 <markvoelker> o/
16:01:43 <catherineD> o/
16:01:45 <gema> o/ (can only stay until half past)
16:01:45 <docaedo> o/
16:01:47 <markvoelker> #info agenda: https://etherpad.openstack.org/p/DefCoreLunar.4
16:01:50 <shamail> Hi all.
16:02:06 <markvoelker> #link Today's agenda: https://etherpad.openstack.org/p/DefCoreLunar.4
16:02:10 <dwalleck> o/
16:02:21 <markvoelker> Ok folks, let's dive right in....
16:02:29 <markvoelker> #topic I'm baaaaaaaack
16:02:42 <gema> markvoelker: welcome back!
16:02:48 <markvoelker> Thanks hogepodge for running the meeting while I was traveling the past couple of weeks!
16:02:49 <hogepodge> hi markvoelker!
16:02:57 <Rockyg> o/
16:03:12 <dwalleck> welcome back!
16:03:16 <Rockyg> gema, I haven't gotten to the review yet :-(
16:03:25 <markvoelker> Just FYI, I did a talk on interop for the China Open Source Cloud League while I was in China and they're quite interested in what we're doing. =)
16:03:25 <shamail> Welcome back markvoelker and thanks hogepodge
16:03:31 <gema> Rockyg: don't worry, now it is a slightly different doc, hopefully better :D
16:03:44 <markvoelker> #topic Pending TC Resolutions
16:03:48 <gema> Rockyg: I will be off next week, so next meeting for me is in 2 weeks
16:04:07 <markvoelker> Mostly just a heads-up if folks aren't aware, but there are a couple of DefCore-related TC resolutions winding their way though gerrit
16:04:25 <markvoelker> #link Add resolution explaining which tests we think defcore should use https://review.openstack.org/312718
16:04:33 <Rockyg> mark, next time you go to China, we can probably arrange a talk through Huawei, too.  Sponsor a meetup or some such.  Same offer to the rest of you
16:04:42 <markvoelker> #link Add resolution asking defcore committee to avoid using proxy APIs in tests https://review.openstack.org/312719
16:05:05 <markvoelker> Rockyg: sure.  Probably won't be going back till next spring, but ya never know. =)
16:05:36 <Rockyg> gema, cool.  Gives me some time.  Life iss currently exciting here.
16:06:15 <markvoelker> #topic Glance Import Refactor FYI
16:06:30 <hogepodge> markvoelker: on the previous topic
16:06:38 <markvoelker> hogepodge: go ahead
16:07:01 <hogepodge> markvoelker: would it make sense to flag the proxy tests when 2016.08 lands to encourage the direct APIs, as a response to the pending TC resolution?
16:07:11 <hogepodge> deprecate, etc
16:07:29 <hogepodge> (assuming the resolution passes)
16:07:35 <markvoelker> Well, I think it would make sense to let the TC debate the resolution and see if lands before we take action. =)
16:07:45 <hogepodge> naturally, but thinking ahead
16:07:59 <markvoelker> But assuming it does, we should probably look at rescoring those with future direction set to 0
16:08:05 <hogepodge> the proxies have been sources of problems for us regardless
16:08:23 <Rockyg> Yeah.  The future direction is definitely away from those.
16:08:42 <VanL_> Why not just deprecate the proxy API tests (assuming 1. This passes, 2. We agree)
16:08:46 <hogepodge> a year in, they haven't been fixed (esp the image calls)
16:09:03 <hogepodge> VanL_: deprecation works too
16:09:13 <hogepodge> do both, as a clear signal
16:09:20 <Rockyg> Well, the image stuff has been slow to get changed/accepted....
16:09:55 <catherineD> hogepodge: these tests are currently flagged .  Right?
16:09:58 <VanL_> These are clearly not the right direction, and we don't want to bake in that *every cloud* has to have them if we are trying to get rid of them
16:10:05 <hogepodge> catherineD: not all of the proxy tests are flagged
16:10:10 <shamail> VanL_: +1
16:10:19 <hogepodge> afaik
16:10:27 <Rockyg> +1
16:11:01 <dwalleck> I agree with markvoelker about making sure the TC resolution lands, but when it does the path  VanL_ and hogepodge said makes sense
16:11:05 <shamail> hogepodge: do we need to take an action to flag them?
16:11:15 <catherineD> hogepodge: do we have vendors failing the un-flagged tests ?
16:11:27 <VanL_> Don't just flag them, remove them.
16:11:29 <hogepodge> catherineD: yes
16:11:41 <Rockyg> we are seeing that the prod devs are making decisions based on current features and imposing them on current supported versions (sort of backporting directions)
16:12:09 <markvoelker> Well, flagging is thing we could do that has effect on Board-approved Guidelines.  Deprecation would get them out of future Guidelines.
16:12:15 <markvoelker> So I think the answer isn't one or the other.
16:12:38 <catherineD> hogepodge: so those vendors could not certified for powered logo ... I think we do have to at least flag them ...
16:12:39 <shamail> I guess should they.
16:12:55 <Rockyg> for us, we are still on juno, but will move to Mitaka, so already only v3 on Keystone.  Problem with the 2015.07 tests...
16:13:18 <catherineD> markvoelker: do we need to flag before deprecate?
16:13:22 <shamail> Should we flag for now (pending resolution) and then remove them eventually?
16:13:39 <markvoelker> catherineD: no.  The two are independent.  You can do one, the other, or both.
16:13:40 <shamail> What catherineD said :)
16:13:45 <Rockyg> Well flagging seems our version of deprecating...
16:14:13 <catherineD> markvoelker: so we can go from required to deprecate ..
16:14:16 <hogepodge> Rockyg: we do both. deprecation for one cycle is required to remove capabilities
16:14:20 <markvoelker> catherineD: yes
16:14:28 <Rockyg> companies who move forward won't be caught by old direction tests.
16:14:34 <dwalleck> Well, flagging could be seen to imply that the test could be fixed
16:14:56 <VanL_> I'd vote for deprecation.
16:14:58 <shamail> Good point dwalleck
16:14:58 <hogepodge> dwalleck: flags are supposed to have reasons and resolutions, which range from fixing the test to removing the test
16:14:59 <markvoelker> dwalleck: We could just cite D400 when flagging it
16:15:04 <Rockyg> Or removed if not fixed
16:15:07 <markvoelker> E.g. it does not meet Criteria
16:15:28 <shamail> I agree with VanL_ if the options are independent.
16:16:24 <VanL_> We have an opportunity here before we put forward the next guideline
16:16:29 <hogepodge> shamail: flagging removes problematic tests from the enforced standards, deprecation is a signal that capabilities will be removed
16:16:52 <hogepodge> shamail: so yes, I'm in favor of doing both
16:17:08 <markvoelker> VanL_: Yeah, I imagine the TC will vote on the resolution well before we send the next Guideline to the Board in a few months
16:17:10 <shamail> Thanks for the explanation.
16:17:29 <markvoelker> VanL_: so once they vote on it, I would have no problem with someone proposing deprecation
16:17:50 <markvoelker> VanL_: and proposing flags against current Guidelines, for that matter
16:18:40 <markvoelker> Ok folks, sounds like we have general agreement that once the TC votes on that resolution we want to respond accordingly.  Ready to move to next topic?
16:19:02 <hogepodge> +1
16:19:12 <Rockyg> how aoubt an agreement line for minutes?
16:19:12 <hogepodge> thanks markvoelker
16:19:16 <dwalleck> +1
16:19:25 <Rockyg> or info...
16:19:54 <markvoelker> #agreed Once the TC votes on the aforementioned resolutions, we should respond in current and forthcoming Guidelines
16:19:59 <catherineD> markvoelker: I think  for approved guidelines the test should be flagged ... for future guidelines tests should be deprecate
16:20:25 <markvoelker> catherineD: correct
16:20:29 <markvoelker> Ok, moving on
16:20:37 <markvoelker> #topic Glance Import Refactor FYI
16:21:09 <markvoelker> Rockyg sent a note about this to the list as well, but just thought we'd highlight it here as well since several of you have been involved in that conversation
16:21:20 <markvoelker> #link image import refactor spec http://specs.openstack.org/openstack/glance-specs/specs/mitaka/approved/image-import/image-import-refactor.html
16:21:41 <markvoelker> A meeting has been proposed for May 26 at 15:00 UTC if you'd like to attend.  Don't think the venue was set last time I looked at the thread...
16:21:43 * markvoelker checks
16:22:05 <Rockyg> thanks
16:22:07 <catherineD> seems like only 10 slots for the meeting
16:22:35 <Rockyg> could happen in meeting-cp
16:22:48 <markvoelker> Looks like it'll be at https://plus.google.com/events/cb4acoebucn25vu8f7enprp85j4
16:22:57 <markvoelker> More info here: https://wiki.openstack.org/wiki/VirtualSprints#Image_Import_Refactor_Sync_.231_--_Newton
16:23:13 <markvoelker> #link  Image IMport Refactor Sync Meeting Info https://wiki.openstack.org/wiki/VirtualSprints#Image_Import_Refactor_Sync_.231_--_Newton
16:23:48 <markvoelker> Anything else on this topic before we move on?
16:24:02 <hogepodge> nikhil:do you have more info on how to join the meeting?
16:24:35 <markvoelker> hogepodge: see link above
16:25:44 <markvoelker> Ok, moving on...
16:25:56 <markvoelker> #topic Open Reviews
16:26:24 <markvoelker> We have several reviews open for DefCore and a few DefCore-related reviews open elsewhere (including the TC resolutions above)
16:26:58 <markvoelker> Let's take a few minutes to run down the list and see if we need to talk through anything in the meeting.  If not, that's fine too.
16:27:17 <markvoelker> #link Remove test lists and generators, update procedures https://review.openstack.org/#/c/313669/
16:27:50 <markvoelker> This one looks ready to land and is pretty non-controversial.  Anything to discuss here?
16:27:55 <hogepodge> I'm ready for it to be merged if there's agreement
16:28:19 <dwalleck> Looks reasonable
16:28:25 <shamail> +1
16:28:29 <markvoelker> Ok.  Let's move on then.
16:28:29 <catherineD> I will need to send a patch to flag a few tests in the advisory section that still require 2 users
16:28:43 <catherineD> after that patch is merged
16:28:44 <hogepodge> I've made changes to refstack-client, and interop page to reflect the changes
16:29:23 <markvoelker> catherineD: ok
16:29:25 <markvoelker> hogepodge: thanks
16:29:28 <markvoelker> #link Added formal 1.5 json schema for gating against https://review.openstack.org/#/c/311265/
16:30:04 <markvoelker> hogepodge: I started looking at this one this morning (still digging out from under my travel backlog) and will finish up this afternoon
16:30:22 <Rockyg> +1
16:30:30 <catherineD> markvoelker: for this one please refstack will need to update the website before DefCore merge this one
16:30:36 <markvoelker> catherineD: we'll make sure not to merge until refstack is ready
16:30:47 <catherineD> markvoelker: thx
16:31:08 * gema goes out of the room without making too much noise, see you guys in two weeks (please keep sending comments on the spec, there is a new version to review/comment on)
16:31:12 <hogepodge> there's an associated patch to make it gating
16:31:25 <Rockyg> bye, gema!
16:31:36 <catherineD> please review this one on RefStack for 1.5 schema support https://review.openstack.org/#/c/319057/
16:31:42 <hogepodge> #link https://review.openstack.org/#/c/317800/
16:31:45 <shamail> See you gema
16:32:06 <hogepodge> thanks gema, have a nice vacation
16:32:29 <shamail> Newbie question: does DefCore actively review these changes during the meeting or just compile a list so we can review before end of the week?
16:32:30 <markvoelker> catherineD: hogepodge: thanks, added both to the etherpad
16:33:02 <markvoelker> shamail: most of the reviewing happens offline in gerrit, but sometimes it's useful to discuss individual items while we're together.
16:33:07 <Rockyg> hghlight the list needing review, shamail
16:33:16 <markvoelker> Anything else on this change?
16:33:21 <hogepodge> my main concern with the patch is if we need to do linting since the schema kind of already does that
16:33:24 <shamail> Thanks, seemed like a lot to read/absorb in the meeting itself. :)
16:34:23 <markvoelker> hogepodge: you mean the gate check?
16:34:36 <hogepodge> yeah. I do a schema check and a lint check
16:35:00 <hogepodge> I wasn't seeing other projects using the linting library
16:35:12 <hogepodge> that I pull in with that patch. Just something to think about
16:35:33 <markvoelker> hogepodge: Hmm.  Ok, I'll look that over this afternoon too then.
16:35:59 <markvoelker> Next up, another schema change...
16:36:07 <markvoelker> #link Introduce target_programs definition https://review.openstack.org/#/c/310621/
16:36:33 <markvoelker> This one seems to have gone a bit cold, so could use some reviews
16:37:27 <hogepodge> I'll add a review that it needs a formal schema
16:37:44 <dwalleck> I'll take a look at it this afternoon
16:37:46 <markvoelker> hogepodge: ++, I don't think we'd started that work when this was initially proposed
16:38:11 <markvoelker> But would definitely be good to have going forward
16:38:17 <catherineD> More important is to discuss whether DefCore support non-defcore guidelines at RefStack for poublic view
16:39:28 <hogepodge> My feeling is that refstack needs to finish up work to support DefCore, which is the project mandate. There's usefulness in allowing projects to define guidelines to help with future defcore direction. Heat immediately comes to mind
16:40:49 <markvoelker> catherineD: Agree that's an important conversation, but I think this change is a smaller scope.  Is there a refstack review on that somewhere we should be contributing to?
16:41:31 <catherineD> hogepodge: agree.  I think the intention to prepare for introducing none defcore or openstack target programs
16:42:17 <Rockyg> hogepodge, the mandate is DefCore, but the founders and original core all had the greater interop test sets as the goal.
16:42:33 <catherineD> markvoelker: not yet for this schema version ... we only have a ptach for hogepodge: 's version 1.5
16:42:38 <Rockyg> But, yes, inital focus was defcore
16:42:46 <markvoelker> catherineD: ok thanks
16:43:04 * markvoelker glances at clock
16:43:11 <Rockyg> is defcore, but we need to keep the greater mission in mind for the design
16:43:38 <shamail> Rockyg: +1
16:44:01 <markvoelker> Anything else on this particular patch?
16:45:14 <hogepodge> Rockyg: https://github.com/openstack/governance/blob/26bed1be455737c91f30de24f42518ebb36b7665/reference/projects.yaml#L3714
16:46:14 <hogepodge> It's the project mission, and it's currently a work in progress fulfilling that mission.
16:46:30 <Rockyg> Yeah.  The yaml file does not capture the mission that was put together before the yaml files were uesed.
16:47:00 <hogepodge> Rockyg: that was that statement that was conditional for the project to be accepted into the big tent. It is deliberate and meaningful
16:47:02 <markvoelker> I suggest that we take mission-statement-of-refstack discussions up with the refstack folks elsewhere in the interest of time. =)
16:47:17 <Rockyg> ++
16:47:21 <markvoelker> #link compute-image flags were not carried to 2015.07 https://review.openstack.org/#/c/310582/
16:47:25 <dwalleck> markvoelker: +1, good discussion to have though
16:48:13 <markvoelker> hogepodge: do we have bugs filed on the implicit deps here yet?
16:48:47 <markvoelker> hogepodge: also, I guess since we're getting rid of the guideline directories and lists we can probably drop those from the patchset...
16:48:55 <hogepodge> markvoelker: no, I can do it this afternoon
16:49:04 <markvoelker> hogepodge: awesome, thanks
16:49:14 <markvoelker> Any other discussion on this one?
16:49:15 <hogepodge> markvoelker: sure, it'll need a rebase
16:49:57 <markvoelker> #link Add simple test for Neutron GET / https://review.openstack.org/#/c/311747/
16:50:16 <markvoelker> This one's actually in Tempest, but it's a test being proposed because we wanted to include a Capability and found it didn't have a test
16:50:40 <markvoelker> Seems like the general logic has been ok'd, so I can probably work up a similar thing for Glance this week
16:51:04 <hogepodge> +1
16:51:42 <markvoelker> Ok, next couple have to do with admin creds in the networking tests...
16:51:52 <markvoelker> #link Remove network port tests that requires admin credentials https://review.openstack.org/#/c/300608/
16:51:57 <hogepodge> markvoelker: more of my backlog work items
16:52:08 <markvoelker> #link Remove capabilities due to tests required admin credentials https://review.openstack.org/#/c/299491/
16:52:12 <hogepodge> hopefully I can get to looking over the network tests tomorrow and Friday
16:52:23 <catherineD> I think we should merge this one since those tests already been flagged in the next.json
16:52:40 <dwalleck> hogepodge: Do we have the go ahead from the Neutron team to rework these tests?
16:52:54 <markvoelker> hogepodge: Ok, awesome.  I was going to start poking at those probably next week-ish if nobody else had that on their plate as I think we definitely want to get those fixed up
16:53:18 <catherineD> it would take time to fix .... meanwhile we should flag them so the users do not have to spend time debug
16:53:36 <hogepodge> dwalleck: I don't have enough information to make an ask
16:54:13 <dwalleck> hogepodge: Good point, neither do I :-) I've looked at them but I don't fully understand if those scenarios can be done without admin creds
16:54:48 <catherineD> dwalleck: hogepodge: it would take time from investoigating for solution and actually fix them ..
16:54:50 <hogepodge> catherineD: I'll +1 with recommendations by the end of the day. Is that reasonable
16:54:53 <markvoelker> catherineD: these tests currently aren't required anyway, right?
16:55:08 <markvoelker> E.g. they were advisory in 2016.01
16:55:28 <catherineD> markvoelker: right but some users looks forward for next required test also look at advisory tests
16:55:49 <catherineD> they are flagged in the required section of the next.json
16:56:31 <markvoelker> catherineD: I'm not sure removing them is a good idea if we think they can be fixed up to not require admin credentials.  By removing them completely we'd have to start the advisory cycle all over again.
16:57:09 <catherineD> markvoelker: not removing --> flag them I mean this patch https://review.openstack.org/#/c/300608/
16:57:34 <catherineD> sorry my mistake
16:57:50 <hogepodge> catherineD: sometimes admin tests are really easy to fix. I just need to check the endpoint to see if admin is required for the operation.
16:57:52 <catherineD> I would rework the patch to flag them
16:57:57 <hogepodge> I should have done it last week
16:58:07 <markvoelker> Doesn't that patch remove them though?
16:58:21 <hogepodge> it's usually a matter of changing the client type, the test remain the same
16:58:22 <markvoelker> Oh, you're saying remove the tests, not the capability
16:58:24 <catherineD> markvoelker: yes it does ... my mistake
16:58:53 <catherineD> let me rework both patches to flag the tests
16:58:59 <markvoelker> catherineD: cool
16:59:07 <markvoelker> Ok, last couple of minutes here folks
16:59:32 <markvoelker> #link Fix Criteria defs in next.json https://review.openstack.org/#/c/319010/
16:59:44 <markvoelker> Nothing much to say on that one, just a sync with the definitions we already agreed on
17:00:12 <markvoelker> We're about out of time...thanks folks!  See you in #openstack-defcore
17:00:30 <markvoelker> #endmeeting