16:00:23 <markvoelker> #startmeeting defcore
16:00:24 <openstack> Meeting started Wed Jul 13 16:00:23 2016 UTC and is due to finish in 60 minutes.  The chair is markvoelker. Information about MeetBot at http://wiki.debian.org/MeetBot.
16:00:25 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
16:00:27 <openstack> The meeting name has been set to 'defcore'
16:00:28 <topol> o/
16:00:31 <eglute> o/
16:00:32 <catherineD> o/
16:00:34 <hogepodge> o/
16:00:43 <markvoelker> #chair eglute
16:00:46 <woodster_> o/
16:00:48 <gema> o/
16:00:50 <openstack> Current chairs: eglute markvoelker
16:00:52 <markvoelker> #chair hogepodge
16:00:54 <openstack> Current chairs: eglute hogepodge markvoelker
16:01:09 <markvoelker> #link https://etherpad.openstack.org/p/DefCoreLunar.11 Today's etherpad
16:01:36 <luzC> o/
16:02:05 <markvoelker> Do we have Brad to talk about the IBM interop challenge?
16:02:13 <topol> yes I am here
16:02:25 <markvoelker> #topic IBM Interop Challenge
16:02:29 <markvoelker> Take it away sir
16:02:44 <topol> Thanks markvoelker! Hi everyone
16:02:50 <dwalleck> (late) o/
16:03:03 <topol> At OpenStack Austin, our GM Don Rippert kicked off an OpenStack Interoperability Challenge.
16:03:16 <topol> The goal of this challenge was to prove/highlight that OpenStack is interoperable at a solution level (work builds upon existing DefCore/RefStack efforts).
16:03:46 <topol> We have been defining this challenge effort and reviewing it with many OpenStack sponsor companies. An incredibly large number of folks have agreed to participate in our Interoperability Challenge and have agreed to show interoperability at the Barcelona Summit.
16:04:05 <topol> We have a   kickoff meeting next week and would like participation from DefCore/RefStack
16:04:25 <topol> This work is meant to build on and showcase DefCore/RefStack
16:05:02 <hogepodge> topol: do yo have any links for this?
16:05:04 <markvoelker> topol: When's the meeting and what do you need from us?
16:05:17 <Rockyg> o/
16:05:35 <brunssen> o/
16:05:40 <topol> I have a slide deck I can make available to you all
16:06:38 <gema> that'd be great
16:06:50 <topol> Ideally would like someone from DefCore/RefStak to partcipate and be informed. I believe Rockyg will be there but we are happy to have others
16:06:58 <topol> This is an open invitation
16:07:32 <eglute> topol could a calendar invite be sent to the defcore mailing list?
16:07:42 <Rockyg> Yup.  I'm there.  I think catherineD is, too.
16:07:49 <markvoelker> topol: Sure, we can probably have some folks join in...once we know when it is. =)  Thanks for making us aware.
16:08:18 <topol> Our  first kickoff meeting is this Monday at 2pm EST.  We can send all the details to the defcore mailing list
16:08:32 <markvoelker> #action topol to send out slide deck and meeting info to defcore ML
16:09:08 <markvoelker> topol: cool.  Anything else you want to discuss today?
16:09:39 <topol> Perfect.  we will do that.  Im not sure if the meeting time on monday is 100% confiremd but I will get it confirmed befre I mail the list
16:09:43 <catherineD> yes  I will be at the meeting
16:09:54 <topol> That is all I have. We look forward to working with all of you.
16:10:03 <eglute> thank you topol!
16:10:06 <markvoelker> Thanks Brad.
16:10:12 <markvoelker> Moving on then....
16:10:15 <topol> Thanks
16:10:23 <markvoelker> #topic Flag test that depend o os-networks extension
16:10:35 <markvoelker> #link https://review.openstack.org/#/c/329220/
16:11:00 <markvoelker> dwalleck: want to speak to this one?
16:11:17 <dwalleck> Sure
16:12:15 <dwalleck> There was some good feedback on this last week from mtreinish and andreaf. It's possible that the issue I've been encountering is a bug/feature add for their credentials provider
16:12:49 <dwalleck> I opened a PR with Tempest to address this: https://review.openstack.org/#/c/341592/
16:13:26 <hogepodge> I'm good with flagging based on the information and response from QA
16:13:53 <dwalleck> It might be the case that the Tempest credential provider was trying to be extra helpful, but maybe too helpful in certain cases. I'm waiting for feedback from the QA folks on this change
16:14:05 <markvoelker> Yeah, I think that's fair.  One request: could we actually get that bug added to the text of the flag?  Really helps with tracking.
16:14:23 <dwalleck> markvoelker: Can do
16:14:37 <markvoelker> dwalleck: Thanks.  If you need an example: https://github.com/openstack/defcore/blob/master/2016.01.json#L804
16:15:01 <markvoelker> Cool.  Anything else on this one today?
16:15:19 <dwalleck> That's all I have. Thanks for the patience and discussion on this point
16:15:36 <markvoelker> Thanks dwalleck.
16:15:38 <markvoelker> Ok, moving on
16:16:01 <markvoelker> #topic Flag request for Hard Reboot Capability
16:16:07 <markvoelker> hogepodge: take it away
16:16:11 <hogepodge> This should be pretty simple.
16:16:22 <hogepodge> A vendor was having problems with this test and asked for a flag on it.
16:16:32 <hogepodge> After talking with the qa team, it seems like the fix is fairly simple.
16:16:50 <hogepodge> Their reboots were happening too fast for keys to be synched to the drive, so the test was failing.
16:17:15 <dwalleck> v3rm!llion
16:17:19 <dwalleck> gah
16:17:22 <hogepodge> The fix is adding a sync call in the guest before the hard reboot.
16:18:01 <hogepodge> But, I wanted to make sure everyone was aware of the situation and have a chance to give feedback. They've been really good about filing bugs and engaging with the QA team.
16:18:24 <hogepodge> #link https://bugs.launchpad.net/defcore/+bug/1584019
16:18:24 <openstack> Launchpad bug 1584019 in defcore "Non working hard reboot test listed in spec 2016.01.required" [Medium,New] - Assigned to Chris Hoge (hoge)
16:18:30 <hogepodge> #link https://bugs.launchpad.net/tempest/+bug/1583987
16:18:30 <openstack> Launchpad bug 1583987 in tempest "reboot_server_hard - can trigger loss of pub key data which fails test" [Undecided,In progress] - Assigned to Ken'ichi Ohmichi (oomichi)
16:18:38 <hogepodge> #link https://review.openstack.org/#/c/340723/
16:19:22 <hogepodge> My suggestion is to not flag given a fix is in progress.
16:19:40 <markvoelker> Good to know.  I think I missed notification on that bug being filed somehow (probably because I was mostly internet-less all last week) so thanks for bringing it up.
16:19:56 <catherineD> hogepodge: is that vendor able to pass the test and thus certify without the flag?
16:20:11 <markvoelker> I'll make a note to look it over.  Just for clarity: no actual flag patch has been filed, right?  Just the bug?
16:20:36 <hogepodge> catherineD: they have a workaround that involves setting a delay to 30 seconds. It makes their test run take a long time, but they can get the test to pass.
16:21:47 <hogepodge> markvoelker: right, they asked me for a flag bug haven't filed a patch for it
16:21:56 <markvoelker> hogepodge: got it, thanks
16:22:05 <markvoelker> Ok, anything else on this one?
16:22:16 <hogepodge> not from me
16:22:25 <markvoelker> Moving right along...
16:22:29 <markvoelker> #topic Admin Credential Tests
16:23:05 <markvoelker> Not much to update on this one: basically there are two different problem classes here.  We have some tests that genuinely need admin creds, and a bunch that don't.
16:23:15 <catherineD> for https://review.openstack.org/#/c/299491/
16:23:53 <markvoelker> https://review.openstack.org/#/c/300608/ would get rid of tests that deal with extended port attributes, which really do usually need admin and therefore aren't good candidates for us
16:24:26 <catherineD> I file that to raise the awareness of the testcases involve ...since then we got many patches that eithe rmove or flag the tests ... so the purpose is served ... I will abandon this patch
16:24:45 <markvoelker> hogepodge: eglute: I think you had both +2'd that one on earlier patchsets, so may just need a quick recheck from you?
16:24:50 <markvoelker> https://review.openstack.org/#/c/300608/ that is
16:24:56 <catherineD> https://review.openstack.org/#/c/299491/ will be abandon
16:24:56 <eglute> done
16:25:48 <markvoelker> catherineD: thanks.  To be clear, 300608 does raise some valid concerns.  There's a bunch of neutron tests using admin creds today that don't need to, so thanks for finding them. =)
16:25:59 <hogepodge> markvoelker: does that patch need to include anything for next.json?
16:26:01 <markvoelker> I have submitted https://review.openstack.org/#/c/335558/ to Tempest to deal with those...it just needs reviews
16:26:21 <markvoelker> hogepodge: which patch?  Sorry, too many links. =)
16:26:59 <hogepodge> markvoelker: catherineD: https://review.openstack.org/#/c/300608/5 , does this need to be applied to next.json
16:27:40 <catherineD> I believe those tests had already been removed from next.json ... but I will check
16:27:43 <hogepodge> markvoelker: are there any other tests that need attention?
16:28:06 <markvoelker> hogepodge: I think those are the only neutron ones, so we should be in good shape when/if that tempest patch lands
16:28:42 <luzC> markvoelker I tested your tempest patch and it worked perfectly (no admin credentials) :)
16:28:52 <markvoelker> luzC: thanks. =)
16:28:52 <luzC> already +1 it
16:29:28 <hogepodge> luzC: awesome, thanks!
16:30:14 <markvoelker> If others want to provide feedback on that patch, you're more than welcome.
16:30:23 <markvoelker> ok, anything else we need to talk about on this topic today?
16:30:30 <catherineD> confirm the tests in https://review.openstack.org/#/c/300608/    were already removed from  next.json https://review.openstack.org/#/c/324892/
16:30:42 <markvoelker> thanks catherineD
16:31:16 <markvoelker> Moving on then...
16:31:20 <markvoelker> #topic Additional Properties
16:31:25 <markvoelker> hogepodge: go for it
16:31:47 <hogepodge> I'm not sure I have much more to add, just checking to see if we have consensus and can move on it.
16:32:19 <hogepodge> markvoelker: I saw your review and can address those points
16:32:22 <catherineD> hogepodge: how is the .py file being used?
16:32:23 <hogepodge> #link https://review.openstack.org/#/c/333067/
16:32:47 <hogepodge> catherineD: good catch, I need to update the information on that
16:32:57 <catherineD> hogepodge: thx
16:32:57 <markvoelker> hogepodge: Thanks.  I think looking at how we're communicating these things is going to be a useful exercise to prevent this situation again in the future.
16:33:37 <hogepodge> catherineD: I've been speaking with mtreinish about it, vendors are going to have to run it against their subunit output, and will likely need to rerun the test with a modified tempest-lib to disable tests that have problems
16:34:00 <hogepodge> catherineD: the tests can fail from strict checking before the action is tested.
16:34:11 <hogepodge> (capability that is)
16:34:21 <hogepodge> catherineD: markvoelker: I'll send a new patch clarifying today.
16:35:03 <hogepodge> markvoelker: regarding the communinication, yes, I see these things as growing pains.
16:35:05 <catherineD> hogepodge: thx
16:35:10 <markvoelker> hogepodge: Ewww.  Can we make that easier somehow?  E.g. maybe provide the necessary lib patch somewhere?
16:36:20 <hogepodge> markvoelker: I'm not sure. It's not great, but also the least invasive. The lib patch would have to happen on a case by case basis, since properties are added in different api calls
16:36:32 <hogepodge> markvoelker: but I can have the vendor give me a diff
16:37:07 <catherineD> would it be easier if refstack-client jsut uses this lib patch all the time?
16:37:15 <markvoelker> Hmm...ok.  I'll try to pick through that a little this week to understand it better.  I'd like to make this an easily repeatable process if we can.
16:37:43 <luzC> I can help to modify refstack-client if needed
16:37:47 <dwalleck> I'm glad to be a guinea pig wherever it might be helpful
16:37:48 <markvoelker> catherineD: probably not.  I think we want most people to use the real tests.  It's only a relatively small number that need the waiver program.
16:37:51 <Rockyg> I'm sure hogepodge  would,too
16:37:53 <hogepodge> catherineD: could be a refstack client feature that generates the patch. I think it should be automateable.
16:38:09 <markvoelker> hogepodge: that would be neat
16:38:23 <hogepodge> luzC: want to work together on that?
16:38:37 <dwalleck> hogepodge: ++ to automatable
16:38:53 <catherineD> hogepodge: yea it would be much more repeatable
16:39:00 <luzC> yes totally :)
16:39:29 <catherineD> at least the process is repeatable
16:39:41 <luzC> hogepodge yes we can work together :)
16:39:41 <hogepodge> Is the Foundation ok for working with vendors, with the assumption we're working out the fine details of the patch?
16:40:15 <hogepodge> markvoelker: eglute: (my previous question)
16:40:32 <eglute> Yes,
16:40:50 <markvoelker> Yeah, I think they're generally onboard.
16:41:39 <hogepodge> Thanks
16:41:46 <eglute> i would like to see the waiver merged, so that hogepodge can move forward with certifying companies
16:42:02 <hogepodge> eglute: I'll get the new patch up today.
16:42:09 <eglute> thank you hogepodge
16:42:11 <markvoelker> FYI, the next BoD meeting is August 23 and I'm sure this will make the agenda. =)
16:42:45 <eglute> I am sure it will!
16:42:48 <markvoelker> Ok, anything further on this today?
16:43:02 <hogepodge> no, thanks
16:43:19 <markvoelker> Muy bueno.  Thanks for your work on this hogepodge, and everyone else for their feedback.
16:43:40 <hogepodge> Happy to help. Thanks everyone.
16:43:40 <eglute> indeed, thanks everyone!
16:43:46 <markvoelker> #topic DefCore Sprint/Midcycle
16:43:54 <markvoelker> #link https://etherpad.openstack.org/p/DefCoreSummer2016Sprint Midcycle planning etherpad
16:44:00 <eglute> I sent out calendar invites
16:44:01 <markvoelker> We had a few suggestions for additional topics
16:44:07 <markvoelker> thanks eglute
16:44:22 <markvoelker> hogepodge, do we know when the Foundation folks are coming/would like to present?
16:44:29 <markvoelker> Should be fairly easy to accommodate them.
16:44:49 <markvoelker> Or, alternately: who should I coordinate with to figure out a good timeslot for them?
16:45:03 <hogepodge> markvoelker: I don't know for certain yet. They're in China right now, but I'll chase them down for an answer.
16:45:13 <markvoelker> Ok, thanks
16:45:49 <hogepodge> Likely Tuesday or Thursday, since I think I'll be traveling with them to/from Austin to work on some other items.
16:46:20 <markvoelker> Sounds good, just let me know when you find out and we'll work something out.
16:46:25 <markvoelker> On the additionalAttributes session suggestion: it looks like we're very close to landing the waiver patches.  Still think we need to budget time for it?
16:46:26 <hogepodge> ok
16:46:45 <eglute> markvoelker i would not think so
16:46:52 <hogepodge> markvoelker: maybe? depends on how the second patch goes.
16:47:24 <markvoelker> Ok.  Maybe the thing to do here is leave it for Fudge Time for the moment and we'll budget additional time if we don't end up landing the patches this week.
16:47:28 <markvoelker> Sound reasonable?
16:47:41 <hogepodge> sounds reasonable, put it in pencil :-D
16:47:48 <markvoelker> =)
16:48:51 <markvoelker> The last suggestion we had was for a bug triage/squashing session.  I'm ok with that if we have time for it, but it's also the sort of thing we could do at a weekly meeting or set up a triage day.
16:49:24 <gema> markvoelker: like we have a lot of free time on weekly meetings to do that?
16:49:28 <eglute> markvoelker would those be tempest bugs? or?
16:49:43 <gema> oh, you mean schedule a separate one
16:50:11 <markvoelker> gema: =)  Well, we're doing patch reviews fairly regularly in meetings so I think we could do bug triage one week if necessary.  But yeah, I separate is fine with me too.
16:50:26 <gema> :D
16:50:26 <hogepodge> Could be useful for wrapping up next. It could be good to do pairs of next against clouds we have access to, so we can shake out any final issues before going to the board.
16:50:41 <hogepodge> s/pairs/tests/
16:50:46 <markvoelker> There are only 10 bugs in LP for DefCore right now, so it should be relatively quick.
16:50:55 <dwalleck> hogepodge: ++, very good idea
16:51:06 <eglute> do we have time scheduled to wrap up 2016.08?
16:51:20 <catherineD> hogepodge: I have a full test of next,json against a few of my existing clouds ... looks good after all the next,json related patches are merged
16:51:29 <hogepodge> catherineD: +1
16:51:53 <markvoelker> catherineD: hogepodge: I have a couple of results against some of my lab clouds too
16:51:53 <hogepodge> catherineD: Can you send me the RefStack links? :-D
16:52:16 <hogepodge> I'm setting up a little Mitaka cloud right now to test against also.
16:52:31 <catherineD> oh I did not send up to RefStack ... send to my own server ... but I can do that
16:53:15 <eglute> markvoelker will we need to do anything else to finalize 2016.08?
16:54:05 <markvoelker> eglute: I think we're in pretty good shape once we get the admin credential stuff landed, but I'll look it over again this week. I suspect we won't need to dedicate (much?) time to it at the midcycle.
16:54:38 <eglute> thanks markvoelker. i dont think there is currently any time scheduled for it
16:54:57 <markvoelker> When we get a little closer to the BoD meeting I'll put up a patch finalizing it for the BoD to review prior to the meeting.
16:55:07 <eglute> great, thank you
16:56:27 <markvoelker> One other thing on the midcycle: if you haven't added your name and employer to the etherpad and are planning to attend in-person, please do so.  RAX needs that info a day beforehand to be ready for us.
16:57:09 <markvoelker> Ok, anything else on the midcycle today?  We're down to a few minutes remaining
16:57:15 <eglute> Also, regarding dinners: i would suggest Riverwalk, but it is a bit of a drive from the Castle
16:57:30 <catherineD> eglute: yay!!
16:57:33 <gema> :D
16:57:46 <hogepodge> eglute: I know, I want to stay downtown, but also want to be practical :-D
16:57:53 * markvoelker doesn't mind a drive down to the riverwalk personally
16:58:16 <eglute> well, the Castle is not downtown, but i do not recomend staying very close to the castle
16:58:17 <dwalleck> One does not simply drive to the riverwalk...
16:58:24 <markvoelker> Ok, one final topic with our two remaining minutes
16:58:25 <catherineD> now we have local tour guides :-)
16:58:29 <hogepodge> Before the meeting closes, I want to make one note about the agenda. I've been make a best guess effort as to how to order everything, but if you have an item that you want to appear earlier in the meeting, please feel free to move it. I generally get the agenda out the night before the meeting, and don't regard it as set in stone.
16:58:32 <markvoelker> dwalleck: ain't that the truth. =p
16:58:44 <markvoelker> hogepodge: thanks for that
16:58:52 <markvoelker> #topic Session submissions for Barcelona
16:59:10 <markvoelker> The deadline to submit talks is tonight just before midnight PDT.  Get those talks in!
16:59:17 <markvoelker> And with that we're out of time...
16:59:25 <eglute> hogepodge is yours defcore 101 type?
16:59:29 <markvoelker> Thanks everyone!  Back to #openstack-defcore....
16:59:29 <hogepodge> link to talks you've sumitted
16:59:34 <hogepodge> eglute: yeah
16:59:40 <eglute> cool thanks!
16:59:49 <markvoelker> #endmeeting