16:00:03 <lbragstad> #startmeeting keystone
16:00:04 <openstack> Meeting started Tue May  1 16:00:03 2018 UTC and is due to finish in 60 minutes.  The chair is lbragstad. Information about MeetBot at http://wiki.debian.org/MeetBot.
16:00:05 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
16:00:06 <lbragstad> ping ayoung, breton, cmurphy, dstanek, gagehugo, henrynash, hrybacki, knikolla, lamt, lbragstad, lwanderley, kmalloc, rderose, rodrigods, samueldmq, spilla, aselius, dpar, jdennis, ruan_he, wxy, sonuk
16:00:08 <openstack> The meeting name has been set to 'keystone'
16:00:15 <lbragstad> #link https://etherpad.openstack.org/p/keystone-weekly-meeting
16:00:17 <lbragstad> agenda ^
16:00:18 <knikolla> o/
16:00:34 <lamt> o/
16:01:12 <wxy|> o/
16:01:48 <lbragstad> since we have a pretty light agenda - we'll give people a few minute to join
16:01:51 <knikolla> FYI: Moving forward, Tuesday's I will be dedicating my working day to Keystone.
16:01:57 <knikolla> Tuesdays*
16:02:03 <lbragstad> knikolla: awesome - i'm happy to hear that
16:02:28 <lbragstad> knikolla: we should sync up on the community goal stuff later if you have time
16:03:13 <knikolla> Sounds good. Except lunch time right after the meeting I'll be here.
16:03:50 <gagehugo> o/
16:05:39 <lbragstad> #topic exposing limit model information
16:05:47 <lbragstad> #link https://review.openstack.org/#/q/status:open+project:openstack/keystone+branch:master+topic:bug/1765193
16:06:06 <lbragstad> so - i have a couple patches up for the limit stuff, and curious about feedback
16:06:11 <kmalloc> o/
16:06:26 <lbragstad> i know a few folks have taken a look already, which i appreciate
16:06:48 <lbragstad> curious what people think about the overall approach
16:06:54 <cmurphy> o/
16:06:59 <lbragstad> and the API in general
16:07:56 <lbragstad> other than that, johnthetubaguy and i just wrapped up a discussion on limits
16:08:54 <lbragstad> wxy|: i doubt you've had a chance process it all yet
16:08:57 <knikolla> do we envision other configuration attributes besides "model"?
16:09:35 <lbragstad> so - model is a dictionary that contains the configured enforcement model (so, flat by default)
16:10:08 <kmalloc> i don't really envision much more on our end besides the model.
16:10:29 <knikolla> is/will model be a global setting or per-domain?
16:10:36 <lbragstad> if we ever implement a 'strict' model for example, that would be relayed in the "model" bit
16:10:46 <kmalloc> knikolla: no, it's a global def.
16:10:47 <lbragstad> global initially
16:11:12 <kmalloc> too hard to reliably make it per-domain, too many fiddly bits and issues on that front (changing between models will basically invalidate the limit data)
16:11:20 <kmalloc> this should not be something toggled per-domain.
16:11:28 <lbragstad> i'm pretty sure my brain would fall out of my ears thinking about this on a per domain basis
16:12:10 <lbragstad> the quota model is something that the system administrator, or operator, should be setting
16:12:23 <lbragstad> because it's how they want resources to adhere in their deployment
16:12:44 <lbragstad> or how strict they want to be with those resources and their enforcement
16:13:19 <lbragstad> then they can use limits to make resource allocations for their customers
16:13:31 <knikolla> agree, makes sense
16:14:24 <lbragstad> i think the two big things that came out of the discussion with johnthetubaguy is:
16:14:37 <lbragstad> 1.) we need to restrict the current proposal to only support two hierarchies
16:15:01 <lbragstad> 2.) we need to be clear about how usage is calculated and what that interaction looks like
16:15:53 <lbragstad> i was going to try and write all this up as a subsequent patch to what wxy| has proposed, since he's already got 90% of this written up
16:16:01 <lbragstad> just clarifying those two points
16:17:40 <lbragstad> does anyone have questions on the unified limit stuff?
16:17:42 <wxy|> lbragstad: thanks. Really helpful discussion with johnthetubaguy. I'll read the whole log later.
16:18:13 <lbragstad> wxy|: ++ it's a lot to parse, but i think it's helpful (at least for me anyway)
16:18:19 <lbragstad> hopefully others find it useful
16:19:48 <lbragstad> lot of moving parts, but that's about all i had for limit stuff
16:19:53 <lbragstad> #topic open discussion
16:20:57 <lbragstad> floor is open if anyone has anything they'd like to bring up
16:21:19 <knikolla> do we have an etherpad where we note down keystoners that will be at the summit?
16:21:29 <knikolla> and arrival departure dates
16:21:41 <lbragstad> knikolla: yep
16:21:45 <lbragstad> #link https://etherpad.openstack.org/p/YVR-keystone-forum-sessions
16:22:35 <knikolla> lbragstad: thanks
16:22:40 <lbragstad> yep
16:24:50 <lbragstad> anything else?
16:24:52 <wxy|> about enable sqlite FK for unit test, https://bugs.launchpad.net/keystone/+bug/1744195  My plan is to enable it step by step by files. Then remove the "enable_sqlite_foreign_key" to enable FK by default at last.  An example is https://review.openstack.org/#/c/558029/
16:24:54 <openstack> Launchpad bug 1744195 in OpenStack Identity (keystone) "unit tests don't enable Foreign Key for Sqlite" [Medium,In progress] - Assigned to wangxiyuan (wangxiyuan)
16:25:02 <wxy|> any thought?
16:25:27 <lbragstad> i like the approach
16:26:01 <lbragstad> as opposed to doing it all in a single patch
16:26:58 <wxy|> I'll try. It may be a large patch since it will effect lots of tests I guess
16:27:20 <wxy|> http://logs.openstack.org/85/558185/2/check/openstack-tox-py27/87b4ea9/testr_results.html.gz
16:27:36 <wxy|> The tests which will fail once the FK is enabled.
16:28:13 <lbragstad> the implementation has a toggle for TestCase that enables or disables it, right?
16:28:34 <wxy|> yeah
16:29:09 <wxy|> it's "enable_sqlite_foreign_key", which is False by default now.
16:29:30 <lbragstad> so then each patch enables it for a specific test class then fixes the tests?
16:30:02 <wxy|> may for a file, or some files in each patch
16:30:11 <lbragstad> sure
16:31:11 <lbragstad> i think that seems reasonable
16:33:32 <lbragstad> i know everyone is pretty busy, so if there is nothing else, we can call it early and get some time back before office hours
16:35:15 <lbragstad> alright - thanks for coming everyone
16:35:19 <kmalloc> make sure to update the summit attendance etherpad
16:35:20 <kmalloc> ;)
16:35:41 <lbragstad> ++
16:35:49 <lbragstad> #endmeeting