17:01:44 #startmeeting mercadorproject 17:01:45 Meeting started Fri Jul 17 17:01:44 2015 UTC and is due to finish in 60 minutes. The chair is geoffarnold. Information about MeetBot at http://wiki.debian.org/MeetBot. 17:01:46 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 17:01:49 The meeting name has been set to 'mercadorproject' 17:02:13 good day all 17:02:18 Rolle call? 17:02:46 * shaleh waves 17:02:47 Rolle? Where did THAT spelling correction come from? 17:03:16 gyee and the other Keystone cores are getting food 17:03:38 do we have an agenda today? 17:03:43 We're at the Keystone midcycle; we're on a lunch break until 2PM EDT 17:04:16 Review some of the findings from the Keystone midcycle 17:04:18 o/ 17:04:27 sounds good 17:04:29 #topic Keystone midcycle fallout 17:05:17 Biggest issue is that there are no immediate plans to address domain name uniqueness 17:05:27 ugh 17:06:24 So either subscribers will need to limit subdomain name choices, or we add some business logic on top (name mangling) 17:06:36 well, we can limp along and bring a stronger argument for why it is needed. Right? Because mercador is sure to expose more issues. 17:06:46 Agreed. 17:07:10 Nothing like ugly names and bad UX to bring change :-) 17:07:19 Everybody recognizes that 40 years of hierarchical namespaces has set expectations 17:07:35 and we're violating Principle of Least Surprise... 17:07:40 but c'est la vie 17:07:57 Wait for M to get a decent model in place 17:08:14 we can code as if it was right yes? It just pushes the issue on to the user. 17:08:26 Pretty much, yes 17:08:32 The second issue... 17:09:21 is that while K2K will keep the IdP in the subscriber (where it belongs), the policy will stay in the publisher's Keystone 17:10:02 So if a reseller decides to create a role that provides a unique combination of rights, 17:10:10 it's not easy to push it down 17:10:34 Today, that's not too bad, because nobody uses aggressive RBAC 17:11:10 but again, we will expose a few pain points 17:11:19 But based on discussions here on wednesday, we're going to try and encourage much more fine-grained RBAC 17:11:39 Yes, Mercador will drive requirements based on what breaks for us 17:11:47 for POC level and early testing we should be OK. Much further and it gets much more painful 17:11:55 Agreed 17:12:09 One of the other topics here.... 17:12:38 ... is testing of capabilities that require multiple clouds to do the functional tests 17:13:03 We should be able to leverage what Keystone K2K is spinning up 17:13:33 Overall it's been a good midcycle 17:13:39 good to hear 17:14:03 gyee and I will do some whiteboarding - delete that, chalk-boarding - later 17:15:14 remember pictures please 17:15:26 try to write a little bigger than you think you should 17:15:27 I've added some material to the Wiki at https://wiki.openstack.org/wiki/Mercador 17:15:37 Good advice 17:15:56 Oh, one thing that popped up yesterday (unexpectedly)... 17:16:42 I'd been relying on the Keystone API doc description; it turns out the doc and the spec have diverged seriously 17:16:55 yay.... 17:17:06 See this bug: https://bugs.launchpad.net/openstack-api-site/+bug/1448602 17:17:06 Launchpad bug 1448602 in openstack-api-site "Policy related operations of Identity v3 API in API Complete Reference need modification." [High,Triaged] 17:17:56 sigh, usual frustration of not tying docs to code 17:18:02 I also met with the Mass Open Cloud group on Monday 17:18:32 To explore the relationship between Mercador and their fine-grained cross-region Mix & Match project 17:19:55 We've made two submissions for Tokyo - one just Mercador (myself and gyee) and one Mercador + Mix&Match (myself and Orran Krieger) 17:20:24 geoffarnold: can you post a link to Mix&Match please 17:20:47 Hang on.... 17:21:40 ... don't have a link yet - getting one 17:22:50 The Keystone etherpad is here: https://etherpad.openstack.org/p/keystone-liberty-midcycle-meetup 17:23:16 The link will be on the Etherpad soon 17:23:57 geoffarnold: thanks 17:25:47 MOC Mix-and-Match-Federation demo https://github.com/CCI-MOC/moc-public/wiki/Mix-and-Match-Federation 17:27:10 Do we have anything else? 17:27:18 Nobody else is back from lunch 17:27:28 I suspect not 17:27:41 you guys need to absorb the conversations of the week 17:28:21 OK, let's wrap early. Please take a look at (and add to) the wiki page. It's mostly section headings at this point 17:29:53 #endmeeting