16:00:05 #startmeeting policy 16:00:06 Meeting started Wed Aug 23 16:00:05 2017 UTC and is due to finish in 60 minutes. The chair is lbragstad. Information about MeetBot at http://wiki.debian.org/MeetBot. 16:00:08 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 16:00:09 ping raildo, ktychkova, rderose, htruta, hrybacki, atrmr, gagehugo, lamt, thinrichs, edmondsw, ruan_he, ayoung, morgan, raj_singh, johnthetubaguy, knikolla, nhelgeson 16:00:11 The meeting name has been set to 'policy' 16:00:19 o/ 16:00:19 #link https://etherpad.openstack.org/p/keystone-policy-meeting 16:00:22 agenda ^ 16:00:26 o/ 16:00:32 o/ 16:00:48 o/ 16:00:49 o/ 16:01:11 i know we have a couple more folks join - so we'll wait a minute 16:01:18 joining* 16:03:06 alright - let's get started 16:03:10 short agenda today 16:03:19 #topic global roles update 16:03:23 #link #link https://review.openstack.org/#/q/status:open+project:openstack/keystone+branch:master+topic:bp/global-roles 16:03:47 ^ there is the implementation for global role assignments for users and groups 16:03:52 o/ 16:04:10 i plan to get a patch up that allows you to get a globally scoped token by the end of the week 16:04:28 once i get a little more planning done for the PTG i'll start that 16:04:39 but please feel free to start playing with the implementation and reviewing 16:04:53 i'm always a fan of early feedback 16:05:39 more information on what we'll be doing for the PoC in Denver can be found in another etherpad 16:05:40 #link https://etherpad.openstack.org/p/keystone-global-roles-poc 16:06:11 that's about all i had for an update - does anyone have questions? 16:06:33 not atm, thanks for spear heading that lbragstad 16:06:41 yep! happy to 16:06:41 +1 16:06:56 #topic open discussion 16:07:01 floor is open 16:07:33 lbragstad: if no one has anything else, let's talk about global role vision per our earlier convo 16:07:41 hrybacki: go for it 16:08:41 okay, so tl;dr we want to think about where we would be in an ideal world e.g. what are the services fully responsible for vs keystone* 16:08:56 in a world where global roles are already a thing* 16:09:18 1 second, my client is acting up 16:12:12 hrybacki: still having issues? 16:12:35 my browser keeps freezing up, sorry 16:13:27 hrybacki: just with irccloud? 16:13:59 ok - i can pick things up until hrybacki get's things squared away 16:14:37 i guess what he wanted clarification on was what policy definition/maintenance looks like after global roles are in place 16:15:37 and my initial response was that policy at the service should not consist of a scope check in policy, but in code, and the policy just consists of a mapping from the role to the action 16:16:00 yes 16:16:28 +1 16:16:28 is there anything else that should be tacked on to that? 16:16:50 back, thanks lbragstad 16:16:56 hrybacki: get it working? 16:17:18 I think so. Maybe I just need to do some solid tab-closing maintenance 16:17:21 so the service responsibility is to do proper scope checking in code 16:17:35 edmondsw: yeah - i'd agree with that 16:18:26 What if we have a set of standard (Default) global roles 16:18:37 i think that will be easy to build on once projects have defaults in code 16:18:55 What if an operator decides to add a new global role 16:18:59 hrybacki you mean standard roles... it is an assignment that adds scope, and we don't have standard assignments 16:19:14 i.e., standard roles, not standard global roles 16:19:27 edmondsw: I'm thinking down the road. What if were to have standard global roles 16:19:35 standard roles being "project_admin" 16:19:35 we won't 16:19:42 agreed upon by the community e.g. a global observer 16:19:54 that would just be observer, not global observer 16:20:20 and then if you want bob to have that role globally, you give them a global role assignment. If you want julie to have that role on a specific project, you give them a project-specific assignment 16:20:24 yeah - then you can give something the `observer` role globally, to a project, or on a domain 16:20:48 /me nods 16:21:33 that's one of the beautiful things about what we're doing here... we avoid all that nonsense from previous discussions about the role itself having global scope 16:22:03 agree with that. 16:22:04 then when projects move scope checks into code, the scope check enforces things automatically 16:22:32 operation.scope == 'global' but not context.global: 16:22:41 raise Forbidden 16:23:12 okay, thanks for fielding my questions :) 16:23:37 does that clear things up? 16:24:06 for now. I need to re-read the BPs keeping this in mind 16:24:25 anyone have anything else? 16:25:25 looks like we can get some time back 16:25:31 thanks for coming! 16:25:33 #endmeeting