Earlier  
Later
New Search
Posted Nick Remark
#openstack-sdks - 2017-11-16
07:59:11 openstackgerrit OpenStack Proposal Bot proposed openstack/python-openstacksdk master: Updated from global requirements https://review.openstack.org/520169
08:36:39 openstackgerrit Rodolfo Alonso Hernandez proposed openstack/python-openstacksdk master: Avoid default mutable values in arguments https://review.openstack.org/510870
10:58:20 openstackgerrit OpenStack Proposal Bot proposed openstack-infra/shade master: Updated from global requirements https://review.openstack.org/520379
11:00:40 openstackgerrit OpenStack Proposal Bot proposed openstack/cliff master: Updated from global requirements https://review.openstack.org/520389
11:07:07 openstackgerrit OpenStack Proposal Bot proposed openstack/keystoneauth master: Updated from global requirements https://review.openstack.org/520418
11:19:46 openstackgerrit OpenStack Proposal Bot proposed openstack/os-api-ref master: Updated from global requirements https://review.openstack.org/520469
11:20:01 openstackgerrit OpenStack Proposal Bot proposed openstack/os-client-config master: Updated from global requirements https://review.openstack.org/520472
11:20:16 openstackgerrit OpenStack Proposal Bot proposed openstack/os-service-types master: Updated from global requirements https://review.openstack.org/520475
11:20:48 openstackgerrit OpenStack Proposal Bot proposed openstack/osc-lib master: Updated from global requirements https://review.openstack.org/520481
11:25:43 openstackgerrit OpenStack Proposal Bot proposed openstack/python-openstackclient master: Updated from global requirements https://review.openstack.org/520518
11:26:07 openstackgerrit OpenStack Proposal Bot proposed openstack/python-openstacksdk master: Updated from global requirements https://review.openstack.org/520169
14:29:59 openstackgerrit Monty Taylor proposed openstack/python-openstacksdk master: Remove ansible functional tests for now https://review.openstack.org/520602
14:57:44 openstackgerrit Merged openstack-infra/shade master: Updated from global requirements https://review.openstack.org/520379
15:04:08 openstackgerrit Merged openstack/python-openstacksdk master: Move task_manager and adapter up a level from cloud https://review.openstack.org/519772
15:04:09 openstackgerrit Merged openstack/python-openstacksdk master: Remove bogus and unneeded role from job definition https://review.openstack.org/520173
15:05:57 openstackgerrit Merged openstack/python-openstacksdk master: Updated from global requirements https://review.openstack.org/520169
15:14:30 openstackgerrit Merged openstack/python-openstacksdk master: Avoid default mutable values in arguments https://review.openstack.org/510870
15:17:04 openstackgerrit Jose Castro Leon proposed openstack/python-openstackclient master: Add support for endpoing filter commands https://review.openstack.org/520614
15:44:53 openstackgerrit Jose Castro Leon proposed openstack/python-openstackclient master: Add support for endpoing group commands https://review.openstack.org/520627
15:50:17 openstackgerrit Jose Castro Leon proposed openstack/python-openstackclient master: Add support for endpoing filter commands https://review.openstack.org/520614
17:01:50 openstackgerrit Monty Taylor proposed openstack-infra/shade master: Cleanup objects that we create on behalf of images https://review.openstack.org/520652
17:01:52 openstackgerrit Monty Taylor proposed openstack-infra/shade master: Add method to cleanup autocreated image objects https://review.openstack.org/520653
17:07:47 openstackgerrit Merged openstack/python-openstackclient master: Updated from global requirements https://review.openstack.org/520518
17:48:48 openstackgerrit Andreas Jaeger proposed openstack-infra/shade master: Remove setting of version/release from releasenotes https://review.openstack.org/520667
18:39:15 openstackgerrit Dean Troyer proposed openstack/python-openstackclient master: Prepare for os-clinet-config to go away https://review.openstack.org/520675
19:07:27 openstackgerrit Monty Taylor proposed openstack/python-openstackclient master: Add logic to handle old and new sdk constructor https://review.openstack.org/520688
19:07:55 mordred dtroyer: ^^ we seem to have both been thinking about the same general area this morning :)
19:11:04 dtroyer Nice, thanks. That'll work for now, eventually the SDK instantiation should move to Clientmanager directly…
19:13:23 openstackgerrit Merged openstack-infra/shade master: Remove setting of version/release from releasenotes https://review.openstack.org/520667
19:59:05 mordred dtroyer: ++
21:05:46 openstackgerrit Andreas Jaeger proposed openstack/os-client-config master: Remove setting of version/release from releasenotes https://review.openstack.org/520790
21:06:26 openstackgerrit Andreas Jaeger proposed openstack/osc-lib master: Remove setting of version/release from releasenotes https://review.openstack.org/520791
21:13:42 dklyle dtroyer, openstackclient question if you have a minute
21:13:56 dtroyer dklyle: sure
21:14:08 dklyle OS_DOMAIN_ID is from the project scoped token?
21:14:25 dklyle not the user's domain, but the project domain?
21:15:22 dklyle Horizon doesn't provide that value, but gophercloud requires it, want to provide the proper value
21:15:31 dklyle in the openrc file
21:16:49 dklyle I'm missing something in the openstackclient code as I couldn't tell where the env is read
21:17:00 dklyle only some documentation
21:17:24 dtroyer it's a ksa thing actually, and the exact context depends on which plugin is in use
21:17:44 dtroyer but this question should be answerable…I'm looking through KSA now
21:18:05 dklyle ksa is the piece I was missing
21:19:03 dtroyer but for gophercloud anting it, I have no idea if that uses things in the same way
21:19:17 dtroyer it should, for all of our sanity
21:19:44 dklyle let me double check gophercloud's usage
21:21:27 dklyle I think they intend project domain. It's used for auth when providing username
21:22:42 dklyle the possibility of domain scoped tokens has domain id very complicated in Horizon as does v3 policy
21:22:43 dtroyer the KSA plugin classes have domain_id and project_domain_id
21:23:17 dklyle so domain_id is the domain gophercloud is auth'ing to
21:24:16 dtroyer now that I think about it, I'm not sure how it works to use different domains for user and project. this is just across the edge of my keystone experience
21:24:42 dklyle you can cross-domain role assignments
21:24:56 dklyle as long as you have a role in domain-x you can log into it
21:25:01 dtroyer ok. I've never actually used a non-default domain cloud
21:25:14 dklyle regardless if your user is from domain-y
21:25:58 dklyle I think will lead me to the right value, I just wanted to make sure the intention of osc for that value
21:26:48 dklyle *I think this
21:26:50 dklyle thanks
21:45:56 openstackgerrit Monty Taylor proposed openstack-infra/shade master: Add method to cleanup autocreated image objects https://review.openstack.org/520653
21:48:03 mordred dklyle, dtroyer: we should actually ping gophercloud and get them to split it- OS_USER_DOMAIN_ID and OS_PROJECT_DOMAIN_ID are important distinctions
21:49:02 mordred dklyle: but yes- OS_DOMAIN_ID is for domain scoped tokens - it's not (as I originally thought it should be) a thing one could use if OS_USER_DOMAIN_ID and OS_PROJECT_DOMAIN_ID were the same thing
21:52:08 dklyle mordred, I need to look at ksa more closely
21:52:22 mordred dklyle, dtroyer: also - while I'm pointificating - in case it's useful information - OS_(USER|PROJECT)_DOMAIN_ID are only needed if USERNAME or PROJECT_NAME are being used. if USER_ID or PROJECT_ID are used, domain information is not required
21:52:22 dklyle but I agree they are very different things
21:52:41 mordred tl;dr - user names and project names are only unique within a domain, but ids are unique globally
21:52:58 dklyle right
21:53:20 dklyle gophercloud only requires OS_DOMAIN_ID if using username
21:54:10 mordred dklyle: speaking of ksa - we've added a bunch of things to it this past cycle that may be useful at your python layer - I've been meaning to go poke through code and check
21:54:54 mordred dklyle: yah - that's a thing I'd like to chat with him about- because for people using envvars or clouds.yaml files that would result in different auth behavior between gophercloud and ksa-based things like python-openstackclient
21:55:07 dklyle I can make a PR to gophercloud, so it better aligns with ksa
21:56:00 mordred at least make sure it supports OS_(USER|PROJECT)_DOMAIN_(ID|NAME) - since those are the settings that are most likely be in people's clouds.yaml or openrc files
21:56:19 dklyle mordred, which python layer?
21:56:22 mordred whether it also has different behavior if OS_DOMAIN_ID is there is likely not as big a deal
21:56:31 dklyle yes Horizon doesn't populate an OS_DOMAIN_ID
21:56:41 dklyle in the openrc, hence the issue
21:56:49 mordred dklyle: ++ horizon is doing the right thing imo
21:57:31 dklyle finally
21:57:48 mordred dklyle: the openstack_dashboard code - but nevermind, that's using python-*client for things, so the discovery bits in ksa will bubble up through those
21:58:09 dklyle we talked at the last PTG about ksa integration
21:58:16 dklyle tighter integration
21:58:26 mordred oh neat! I should have dropped in and lurked on that
21:58:49 dklyle we've implemented duplicate functionality over the years, some probably predating
21:58:59 mordred there's definitely some catalog/version-discovery things that I *think* could simplify some things for you ... but I haven't had a chance to look deeply yet
21:59:02 mordred yah.
21:59:20 dklyle our biggest concern is stability
21:59:23 mordred ksa is awesome, but its existence definitely came a bit late
21:59:37 dklyle and someone that understands enough that has enough time to take it on
21:59:42 mordred that's the hardest part
21:59:45 dklyle the same story
21:59:48 dklyle yeah
22:00:05 mordred dang. there was a horizon bug I was going to mention ... but now I forgot what it was ...
22:00:33 dklyle if I can carve out a largish chunk of time, I might try to take it on, but I haven't yet
22:01:59 mordred dklyle: I'll try to help some - the keystoneauth work I needed to get done last cycle is 95% done - doing some work to help consume it isn't a terrible idea
22:12:07 openstackgerrit Monty Taylor proposed openstack/python-openstacksdk master: Remove ansible functional tests for now https://review.openstack.org/520602
22:12:08 openstackgerrit Monty Taylor proposed openstack/python-openstacksdk master: Add OSC tips unittest jobs https://review.openstack.org/520812
22:12:48 openstackgerrit Monty Taylor proposed openstack/python-openstacksdk master: Remove ansible functional tests for now https://review.openstack.org/520602
22:12:49 openstackgerrit Monty Taylor proposed openstack/python-openstacksdk master: Add OSC tips unittest jobs https://review.openstack.org/520812
22:13:54 mordred dtroyer: ^^ addition of the osc tips job so that we can make sure we don't break osc ...
22:14:02 mordred dtroyer: should we add the functional test job too?
22:14:40 dtroyer the functional-tips job actually does more, the unit tests mock out a lot

Earlier   Later