15:03:27 <markvoelker> #startmeeting DefCore
15:03:27 <openstack> Meeting started Wed Aug 26 15:03:27 2015 UTC and is due to finish in 60 minutes.  The chair is markvoelker. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:03:28 <zehicle2> o/
15:03:29 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
15:03:32 <openstack> The meeting name has been set to 'defcore'
15:03:34 <markvoelker> chair zehicle2
15:03:42 <barrett1> o/
15:03:42 <markvoelker> #chair zehicle2
15:03:43 <openstack> Current chairs: markvoelker zehicle2
15:03:52 <markvoelker> #chair eglute
15:03:53 <zehicle2> :) thanks Markl
15:03:53 <openstack> Current chairs: eglute markvoelker zehicle2
15:03:55 <markvoelker> =)
15:04:02 <eglute> thanks markvoelker!
15:04:32 <eglute> #topic agenda
15:04:35 <eglute> #link https://etherpad.openstack.org/p/DefCoreFlag.12
15:05:34 <zehicle2> other items for the agenda?
15:05:50 <eglute> i been out all week, so i don't have anything
15:07:44 <zehicle2> I was thinking we'd focus on new capabilities
15:08:06 * markvoelker added links to the current capability patches to the etherpad
15:08:12 <eglute> ok, do you want to start with that or swift tests (as per agenda)
15:08:22 <markvoelker> Is hogepodge here?
15:08:36 <markvoelker> We deferred that one from last week b/c he was at the PAO ops meetup...
15:09:01 <eglute> is there OpenStack SV going on?
15:09:06 <eglute> he might be there
15:09:22 <markvoelker> Yes, and he probably is.  Defer again?
15:10:04 <eglute> yes
15:10:06 <zehicle2> +1
15:10:14 <zehicle2> I'm not ready to talk about it in his pace
15:10:15 <markvoelker> Ok, on to capabilities
15:10:38 <markvoelker> Oh wait...actually there's another item in the agenda first
15:10:44 <markvoelker> The vendor data collection patch
15:10:44 <catherineD|2> markvoelker: I added the Heat patch still working on scoring ...
15:11:17 <markvoelker> Not sure we really have much to talk about there though....I started an ML thread on it which so far has generated one additional -1.
15:11:36 <markvoelker> Might just need to let further d
15:11:46 <markvoelker> iscussion on the ML happen.
15:11:48 <scotticus> markvoelker i'm curious about the ML thread you posted.
15:11:55 <scotticus> would discussion about that need to take place there?
15:12:26 <markvoelker> scotticus: if folks want to discuss it here I'm fine with that, but might be nice to keep the discussion in one place....
15:12:54 <eglute> scotticus +1 on ML thread, but if you have questions might be faster here
15:13:26 <scotticus> i guess the main question is, what are we going to be required to publish?
15:13:48 <scotticus> the configuration of the testbed?
15:13:52 <scotticus> or production systems?
15:13:58 <markvoelker> scotticus: Take a look at the patch: https://review.openstack.org/#/c/207209/6/HACKING.rst
15:14:40 <markvoelker> The way it's been discussed is that it's the testbed you're running refstack against
15:15:09 <scotticus> got it.  i think thats fair as long as it isn't production systems.
15:15:32 <zehicle2> and I've been moving towards allowing vendors to report the system they stay will pass the same tests
15:16:42 <zehicle2> originally, my concern was that users would be able to replicate vendor results
15:18:08 <markvoelker> So, I've registered why I think that's not very good info given the stated purpose of the patch in the comments and on the ML.  I won't get into that here again in the interest of time....
15:18:25 <scotticus> k
15:19:04 <markvoelker> But I would be interested to hear how folks things that data baout one testbed run would allow "users to understand the requirements for deployable products and to know the resources available on hosted products."
15:19:40 <markvoelker> I'd suggest we take that up on the ML though, so as not to fragment the conversation further.
15:19:49 <eglute> +1
15:19:59 <eglute> on to capabilities?
15:20:07 <markvoelker> We're already spanning a gerrit review, an ML thread, an in-person meetup, and a couple of IRC logs. =)
15:20:33 <markvoelker> eglute: +1
15:20:45 <eglute> #topic capabilities
15:21:19 <markvoelker> #link https://review.openstack.org/#/c/210080/ network capabilities
15:21:34 <markvoelker> #link https://review.openstack.org/#/c/213353/ Glance capabilities
15:21:46 <markvoelker> #link https://review.openstack.org/#/c/213330/ Keystone capabilities
15:21:57 <markvoelker> #link https://review.openstack.org/#/c/216983/ Heat capabilties
15:22:21 <markvoelker> Those are the four we have now.  I'll probably post one more small one for Nova today as well.
15:22:43 <eglute> thanks markvoelker
15:22:53 <markvoelker> OF the four above, the network one is probably closest to landing and has had some good discussion.  Would be good if we can start closing in on a final patch.
15:23:14 <eglute> i will review today, thanks markvoelker
15:23:20 <markvoelker> I have a todo to fill out the designated sections bits, and there's a little discussion about which tests should go with what capabilties I believe.
15:23:27 <markvoelker> I'll have a closer look at the latter this afternoon.
15:23:36 <eglute> so for " Aug 27 (S-2): ID new Capabilities [Lead by Community] <=== NEXT MILESTONE (T-minus 1 days from now)" milestone
15:24:02 <markvoelker> eglute: Yep.  If there are any other candidates for scoring, we need patches posted by tomorrow.
15:24:08 <markvoelker> We then have a month to complete the scoring.
15:24:12 <eglute> cool
15:24:35 <zehicle2> neutron looks close
15:24:37 <markvoelker> Thus far I haven't seen anything for Cinder (rockyg), Swift (hurricanerix), or other stuff.
15:24:43 <eglute> i think we are on track, assuming you dont want the patches landed by tomorrow i think
15:24:58 <catherineD|2> Will we have meetings to go over the scoring of the other just like what we did at defcore f2f for Neutron?
15:24:58 <markvoelker> eglute: nope, no need to land them by tomorrow.
15:25:24 <zehicle2> beside neutron, are we expecting other caps for Tokyo?
15:25:32 <markvoelker> catherineD|2: I am totally open to doing that, though folks should start reviewing them in Gerrit immediately.
15:25:53 <markvoelker> I will be unable to host any scoring meetings next week on account of VMworld though.
15:26:34 <markvoelker> zehicle2: I guess that depends on the scoring. =)  The patches posted indicate that there's a case for more capabilties to be added, but we need to see how the score.
15:26:39 <zehicle2> IMHO, shared screen meetings for scoring are productive - good to have them setup
15:26:52 <zehicle2> markvoelker: ok
15:27:20 <zehicle2> FWIW - I like having them broken into individual files
15:27:40 <catherineD|2> zehicle2: +1
15:27:46 <zehicle2> I did not get a chance to check your scoring code
15:27:55 <eglute> +1 on individual files
15:28:11 <markvoelker> zehicle2: no worries, it was merged last week
15:28:28 <hurricanerix> markvoelker: I am in another meeting right now, so I have missed most of the discussion, but if you need anything from Swift, can I talk with you about it after the meeting?
15:29:11 <markvoelker> hurricanerix: sure, I'll be around.  Basically we just wanted to remind you that if you plan to propose any additional Capabilities for Swift in the next Guideline, tomorrow is the deadline to get a patch posted.
15:30:16 <zehicle2> on quick review, it looks like the glace v1 caps scores are high enough too
15:30:20 <hurricanerix> Ok, thanks.  I'll check with the PTL to see if anything needs to be done.
15:30:55 <markvoelker> zehicle2: which reminds me
15:31:13 <markvoelker> zehicle2: We may want to have a discussion about Criteria weighting at some point.
15:31:32 <zehicle2> you have reason to think we should adjust?
15:31:38 <markvoelker> zehicle2: right now, all Criteria are weighted approximately equally and I'm starting to think that maybe shouldn't be the case
15:31:49 <markvoelker> For example, with Glance v1....
15:31:50 <zehicle2> happy to review/discuss - it's always a good idea to check and reflect
15:32:21 <zehicle2> markvoelker: I was waiting to get some data - would be happy to review it
15:32:42 <zehicle2> FWIW, we'd likely want to have all the existing caps setup in the current format to play w/ different weights
15:32:43 <markvoelker> It scored 0 on TC future direction, and has already been replaced by v2 for some time.  It seems weird to make it mandatory at this stage.
15:33:12 <markvoelker> I haven't thought about this a whole lot yet, so no need to discuss it today, but I have some ideas percolating.
15:33:13 <zehicle2> the rules are setup so that we can overrule scores
15:33:37 <zehicle2> in the past, we did not want TC direct to trump everything else.
15:33:38 <markvoelker> Oh, and FWIW: the scoring script just reads the scoring weights right out of the JSON file, so it's fairly trivial to play with different weights locally.
15:34:17 <zehicle2> ok, I made a bad assumption that you used the scoring pages
15:35:13 <markvoelker> zehicle2: Nope, it grabs them striaght out of the JSON.  E.g. https://github.com/openstack/defcore/blob/master/2015.next.json#L1671-L1731 (or whichever JSON file you tell it to use; .next is just the default)
15:37:00 <markvoelker> So anyway, back to more immediately actionable things. =)  Scoring meetings: do we want them? I'm happy to set one up for week after next if so.
15:37:22 <markvoelker> Next week I think both Egle and I are out at VMworld, but we can use that time to have folks do preliminary reviews in Gerrit.
15:37:49 <eglute> yes, I will be at VMworld
15:38:52 <zehicle2> markvoelker: yes, I think we need them
15:39:35 <markvoelker> Ok.  Our old timeslot for that was Wednesday at 2pm EDT = 1pm CDT = 11am PDT.  That still ok with folks?
15:39:38 <zehicle2> I'd also be interested to talk about weights
15:40:06 <zehicle2> could we use this time and just add a voice/screen?
15:40:29 <zehicle2> I don't know that we have so much business - we could dedicate meetings to scoring
15:40:57 <zehicle2> the other slot is OK too - this one is already blocked out for me
15:41:21 * markvoelker could do either and will defer to group consensus
15:41:48 <markvoelker> eglute: catherineD|2: your thoughts?
15:41:53 <eglute> i am ok with either as well
15:42:08 <catherineD|2> I prefer this time slot ..
15:42:38 <markvoelker> OK, so we'll plan on not having a separate scoring meeting but instead using this time, starting week after next
15:42:42 <markvoelker> Sound good?
15:43:02 <markvoelker> And if we need to continue discussion, try to keep the other slot clear on your calendars and we'll meet again if necessary.
15:43:13 <eglute> +1
15:43:17 <catherineD|2> markvoelker: +1
15:43:30 <markvoelker> In the meantime:
15:43:36 <catherineD|2> just a time check ... could we allow sometime for RefStack update here ...
15:43:39 <markvoelker> #action everyone review the four existing scoring patches
15:44:27 <markvoelker> I think that's all we need to discuss today for capabilities, so....chairs?  Skip to RefStack?
15:44:35 <zehicle2> I'd ask people to +0 on existing patches to show we've got activity
15:44:35 <eglute> yes, refstack
15:44:43 <catherineD|2> thank you!!
15:44:44 <eglute> #topic refstack
15:45:42 <catherineD|2> RefStack can now upload data with signed user ... with that we now support delete uploaded data by the user ...
15:46:03 <zehicle2> to an extent, we could merge the working documents
15:46:16 <catherineD|2> Once uploaded, user can decide to share the data with community anonymously ....
15:46:20 * zehicle2 would like to see the working docs landed
15:46:32 <zehicle2> that would let us discuss individual utems
15:47:16 <zehicle2> markvoelker: eglute do you object if we land the working files?
15:47:34 <eglute> +1
15:47:40 <markvoelker> catherineD|2: cool.  Any docs on how to use those features?
15:47:42 <catherineD|2> I would like to disable the feature to allow upload data anonymously .... since data can now share anonymously ...
15:47:43 <zehicle2> sorry catherineD|2 .... I'll hold that.  please keep goin in RefStack
15:48:07 <zehicle2> catherineD|2: +1 on auth'ed uploads
15:48:46 <markvoelker> catherineD|2: I presume by default uploads are not shared?
15:48:50 <catherineD|2> zehicle2: yep so it would be more responsible data upload ...
15:48:52 * zehicle2 thinking about that more deeply
15:48:59 <markvoelker> Err...are shared anonymously?
15:49:09 <catherineD|2> markvoelker: correct default is not shared
15:49:12 <markvoelker> ok
15:49:13 <markvoelker> neat
15:49:35 <zehicle2> the uploaded data is postive results only?
15:49:39 <catherineD|2> markvoelker: yes share is anonymously until we implement vendor registration process ..
15:49:47 <markvoelker> I have some folks who will be running tests this week or next, so would be happy to try that out.
15:49:55 <catherineD|2> zehicle2: yes only pass data
15:50:37 <catherineD|2> passed test names ..
15:50:51 <zehicle2> the biggest risk would be that incomplete vendor results would be shared
15:51:03 <zehicle2> individual uploads are not authorative
15:51:16 <catherineD|2> with that, we are ready to implement vendor registration ... which I need spec fron DefCore and the foundation ..
15:51:49 <eglute> catherineD|2 what do you have in mind for that
15:51:53 <zehicle2> I'm OK as long as vendors get to pick/approve which results are attributed to them
15:51:53 <catherineD|2> zehicle2: data can only be shared if user (vendor) decide to share them
15:52:10 <zehicle2> perfect
15:52:43 <catherineD|2> with allowing  anonymously upload we have no way to clean up the data ...
15:53:29 <zehicle2> I don't see any downsides - the people uploading data expect it to be public at some point.
15:53:31 <catherineD|2> eglute: What I am not clear about vendor registration is who will be authorized to associate vendor to user ...
15:53:54 <eglute> catherineD|2 i think foundation should do that manually, at least for now
15:54:03 <eglute> and by foundation, i mean hogepodge
15:54:04 <eglute> :)
15:54:20 <catherineD|2> eglute: ++1  I think it should be foundation
15:54:29 <eglute> right
15:54:47 <catherineD|2> because user can change company ... and how do we control that?
15:55:05 * markvoelker starts pondering about results signed using keypairs so vendors can just control who they give keys to within their orgs, and can rotate keys when employees change
15:55:12 <eglute> right, it will have to be manual
15:55:36 <eglute> or what markvoelker said. but that assumes companies are that well organized
15:56:38 <catherineD|2> markvoelker: actually data is control via keypairs ...
15:57:14 <catherineD|2> RefStack would like to have a spec before any implementation on this ...
15:57:42 <catherineD|2> this is 1 of the 2 remianing action item for Liberty for RefStack ..
15:57:46 <markvoelker> catherineD|2: Ok, can you draft up an email to the ML explaining what you need from us and/or the Foundation?
15:57:56 <markvoelker> (we're down to 3 minutes here)
15:58:03 <catherineD|2> markvoelker: OK
15:58:04 <markvoelker> I'll be happy to help draw something up
15:58:15 <markvoelker> Though, per earlier, next week's pretty much a wash for me. =/
15:58:16 <zehicle2> I'd like to jump back to merging the working cap docs
15:58:37 <zehicle2> they are working materials
15:58:52 <zehicle2> and I think they are in a state that we can take patches on details instead of the whole doc
15:59:00 <markvoelker> zehicle2: I'd prefer not to merge the network one since there's already good discussion going on in that review
15:59:09 <zehicle2> ok
15:59:14 <markvoelker> Line item comments are working well there
15:59:40 <markvoelker> One last thing before time expires on us: could we land this? https://review.openstack.org/#/c/215263/2
15:59:58 <markvoelker> Until we do, nobody can certify against .07 unless they're running Xen.
16:00:10 <markvoelker> ....aaaand we're out of time. =)
16:00:23 <zehicle2> markvoelker: yes, I think so
16:00:27 <markvoelker> #endmeeting