| *** etoews has joined #openstack-sdks | 01:20 | |
| *** ayoung has quit IRC | 01:23 | |
| *** ayoung has joined #openstack-sdks | 01:23 | |
| *** etoews has quit IRC | 01:25 | |
| *** etoews has joined #openstack-sdks | 01:56 | |
| *** etoews has quit IRC | 02:00 | |
| *** etoews has joined #openstack-sdks | 02:23 | |
| *** etoews has quit IRC | 02:28 | |
| *** etoews has joined #openstack-sdks | 02:41 | |
| *** etoews has quit IRC | 02:46 | |
| *** stevemar has joined #openstack-sdks | 03:38 | |
| *** jamielennox has quit IRC | 04:09 | |
| *** etoews has joined #openstack-sdks | 04:11 | |
| *** etoews has quit IRC | 04:16 | |
| *** etoews has joined #openstack-sdks | 04:19 | |
| *** stevemar has quit IRC | 04:22 | |
| *** jamielennox has joined #openstack-sdks | 04:24 | |
| *** etoews has quit IRC | 04:26 | |
| *** etoews has joined #openstack-sdks | 04:40 | |
| *** jamielennox has quit IRC | 04:40 | |
| *** etoews has quit IRC | 04:44 | |
| *** briancurtin has quit IRC | 07:57 | |
| *** etoews has joined #openstack-sdks | 08:07 | |
| *** etoews has quit IRC | 08:12 | |
| *** etoews has joined #openstack-sdks | 08:26 | |
| *** etoews has quit IRC | 08:31 | |
| *** openstackgerrit has joined #openstack-sdks | 13:01 | |
| *** etoews has joined #openstack-sdks | 14:38 | |
| *** etoews has quit IRC | 14:43 | |
| *** etoews has joined #openstack-sdks | 14:56 | |
| *** briancurtin has joined #openstack-sdks | 14:58 | |
| *** briancurtin has joined #openstack-sdks | 14:58 | |
| *** etoews has quit IRC | 15:01 | |
| *** terrylhowe has left #openstack-sdks | 15:49 | |
| openstackgerrit | OpenStack Proposal Bot proposed a change to openstack/python-openstackclient: Updated from global requirements https://review.openstack.org/126682 | 15:56 |
|---|---|---|
| openstackgerrit | OpenStack Proposal Bot proposed a change to stackforge/python-openstacksdk: Updated from global requirements https://review.openstack.org/127650 | 15:58 |
| openstackgerrit | Monty Taylor proposed a change to stackforge/os-client-config: Add support for command line argument processing https://review.openstack.org/127807 | 17:33 |
| dtroyer | mordred: so dammit, I'm 3/4 through doing just that ^^^ for osc to combine with the auth plugin option handling... | 17:43 |
| dtroyer | mordred: if you're using any of the existing cli parsers, watch out for their default environment variables stomping on your values from the yaml file. The precedence combination is likely not what you expect | 17:50 |
| mordred | dtroyer: oops. we should talk more then :) | 17:52 |
| dtroyer | combining these two is like straddling two boats going northwest and west…I'm going to get wet before it's over | 17:54 |
| mordred | dtroyer: I just got it plumbed through from ansible, but hadn't tried passing in args list from python-*client argparse yet | 17:54 |
| mordred | yah | 17:54 |
| dtroyer | argparse gives you a Namespace object, that takes a bit of care... | 17:54 |
| mordred | oh lovely | 17:54 |
| dtroyer | I was headed in the direction of just adding a method to take the Namespace and merge it on top o everything else | 17:54 |
| dtroyer | on the assumption that command line args should have preference over config file args | 17:55 |
| mordred | yes. I think command line args are the most explicitly request things - totally agree | 17:55 |
| dtroyer | let me get things cleaned up a bit and I'll push up what I have and we can compare notes | 17:55 |
| mordred | basically I'm treating the passed in args list (which is what I get from ansible) same as command line args - it just doesn't come from argparse | 17:56 |
| mordred | cool | 17:56 |
| dtroyer | mordred: ok, so I just realized that what I have is taking a slightly longer way around and the ability to add the options later isn't a requirement… | 18:51 |
| dtroyer | however, I did stop to write up the option/config handling sequence in OSC at https://etherpad.openstack.org/p/os-client-config to lay out the auth plugins options | 18:52 |
| dtroyer | I'm curious to see how different what you have in ansible is | 18:52 |
| dtroyer | I'm going to try it now with 127807 and add the two tweaks I need yet | 18:55 |
| mordred | dtroyer: reading | 19:17 |
| mordred | dtroyer: hrm. it sounds like kdc auth plugins mean we need the os-client-config to re-run? or .. hrm | 19:20 |
| mordred | it sounds like since we'll happily pick up any value you might want to put in a config file | 19:20 |
| mordred | then we should have all of the things that would be needed - other than not having a great way to do required things validation | 19:21 |
| mordred | maybe yeah - the thing you say about REQUIRED_VALUES optionally being skipped is the thing | 19:21 |
| dtroyer | in osc I'm letting that validation be handled by the plugins | 19:21 |
| mordred | yah | 19:21 |
| dtroyer | I don't like not having control over how thigs look | 19:21 |
| dtroyer | but so far it's ok | 19:21 |
| mordred | I think that's right. | 19:21 |
| mordred | also, you already swapped the config/env var precedence, no? | 19:22 |
| dtroyer | not yet | 19:22 |
| dtroyer | so let me paste what I've done to get_one_cloud() to handle the Namespace | 19:22 |
| mordred | okie. cool | 19:22 |
| mordred | dtroyer: yeah - that looks totally sane | 19:23 |
| mordred | dtroyer: devananda and I discovered tons of fun related to project_id and project_name yesterday | 19:23 |
| mordred | apparently HP and Rackspace public clouds both have different behavior than devstack clouds | 19:24 |
| mordred | :) | 19:24 |
| dtroyer | I'm not sure how to handle the Namespace as a single arg, I want cloud_config to do the breakdown work so the argparse= bit does that | 19:24 |
| dtroyer | I'm not surprised about the name/id stuff. it's a mess | 19:24 |
| mordred | in that in both rax and hp, project_id and project_name are interchangable completely | 19:24 |
| mordred | in devsatck, you can pass project_id to project_name, but not vice-versa | 19:24 |
| dtroyer | does current v3 not allow that too? | 19:25 |
| dtroyer | ah, ok | 19:25 |
| mordred | total madness | 19:25 |
| dtroyer | I've only put ID into name I think | 19:25 |
| dtroyer | that's mostly because I've just ignored project-id, why? I want to rename it to project and let the backend figure it out | 19:26 |
| mordred | yes. I kinda want to do that too | 19:26 |
| dtroyer | so if everyone accepts either in project-name, we're good | 19:26 |
| mordred | yup. project-name seems to work across teh board | 19:26 |
| dtroyer | I'm also doing some filtering from the auth plugins…do you think that belongs in os-client-config? I don't care where it lives, at least until I get it into a lowest-level api somewhere | 19:28 |
| dtroyer | you already have some of the same stuff I have in OSC wrt tenant/project | 19:28 |
| mordred | yeah - tenant/project I definitely want to hide from users because it's crazy and nobody programming in openstack should need to know the dance | 19:30 |
| mordred | auth plugins ... maybe? I haven't played with those at all - what do you think? | 19:30 |
| mordred | is there a good way to dtrt there? (and/or would that make os-client-config need to depend on keystoneclient? | 19:31 |
| dtroyer | I want everything in OSC to be consistent and some of the option names from the plugins are odd. It would be nice to have all of that corection in one place. I'm not sure if plugin-specific stuff belongs here though | 19:31 |
| dtroyer | I wouldn't make anything dependent, just do the mappings there | 19:31 |
| mordred | cool. well, I share the same desire | 19:33 |
| mordred | also - the auth plugins are new - why they heck do they have weird option names? | 19:33 |
| dtroyer | same as always, people don't talk | 19:33 |
| mordred | yay! | 19:34 |
| dtroyer | I'd have thought the existing names might have influenced some of that, but , you know the story | 19:34 |
| mordred | my favorite thing we discoverd is that python-keystoneclient, when you pass project_name to it - still sends it over the wire as tenant_name | 19:35 |
| dtroyer | I do have a couple of things to push back into KSC once this is firmed up, I'll add those option names too | 19:35 |
| dtroyer | v2 auth _always_ will do that…we you using v3? | 19:35 |
| mordred | not to my knowledge | 19:36 |
| mordred | but that was deva on devstack - so it's possible it was doing v3 | 19:36 |
| dtroyer | there is a path when ksc does the version discovery where tenant might still show up in a v3 request. I haven't used that yet | 19:38 |
| dtroyer | ok, I'm good on those reviews with the exception of that one bool type check | 19:43 |
| *** briancurtin has quit IRC | 19:52 | |
| *** etoews has joined #openstack-sdks | 19:55 | |
| *** etoews has quit IRC | 20:00 | |
| openstackgerrit | Monty Taylor proposed a change to stackforge/os-client-config: Add support for command line argument processing https://review.openstack.org/127807 | 20:01 |
| mordred | dtroyer: cool. thanks - nice catch | 20:01 |
| dtroyer | You want deva to look at these or shall I +A them? | 20:05 |
| *** etoews has joined #openstack-sdks | 20:13 | |
| mordred | go for it! | 20:14 |
| mordred | in fact, once they land, maybe we should release an 0.3.0? | 20:14 |
| *** etoews has quit IRC | 20:18 | |
| dtroyer | I'd like to get the Namespace bit in first? | 20:18 |
| dtroyer | I'll need a rebase yet... | 20:19 |
| openstackgerrit | A change was merged to stackforge/os-client-config: Remove unused class method get_services https://review.openstack.org/127668 | 20:20 |
| openstackgerrit | A change was merged to stackforge/os-client-config: Handle no vendor clouds config files https://review.openstack.org/124645 | 20:21 |
| openstackgerrit | A change was merged to stackforge/os-client-config: Add cache control settings https://review.openstack.org/124643 | 20:24 |
| openstackgerrit | A change was merged to stackforge/os-client-config: Handle the project/tenant nonesense more cleanly https://review.openstack.org/127664 | 20:25 |
| openstackgerrit | A change was merged to stackforge/os-client-config: Handle lack of username for project_name defaults https://review.openstack.org/127679 | 20:26 |
| openstackgerrit | A change was merged to stackforge/os-client-config: Add support for command line argument processing https://review.openstack.org/127807 | 20:27 |
| openstackgerrit | Dean Troyer proposed a change to stackforge/os-client-config: Add support for argparse Namespace objects https://review.openstack.org/127810 | 20:30 |
| openstackgerrit | Dean Troyer proposed a change to stackforge/os-client-config: Add support for argparse Namespace objects https://review.openstack.org/127810 | 20:41 |
| openstackgerrit | Dean Troyer proposed a change to stackforge/os-client-config: Add support for argparse Namespace objects https://review.openstack.org/127810 | 20:48 |
| *** briancurtin has joined #openstack-sdks | 20:53 | |
| openstackgerrit | Dean Troyer proposed a change to stackforge/os-client-config: Add support for argparse Namespace objects https://review.openstack.org/127810 | 20:56 |
| dtroyer | jeez…apparently I can't py26 on the first or second try… | 20:56 |
| mordred | heh | 20:57 |
| mordred | py26 sucks | 20:57 |
| mordred | dtroyer: so - do you need the cloud param to continue to live in the config dict? | 20:58 |
| *** briancurtin has quit IRC | 20:58 | |
| mordred | dtroyer: removing it made a bunch of pass-through stuff easier for me in shade ... but if you need it I can totally do the other thing | 20:58 |
| mordred | (other than that, that patch looks great to me whenever it works for you) | 20:59 |
| *** briancurtin has joined #openstack-sdks | 20:59 | |
| dtroyer | no I don't need it | 21:00 |
| mordred | excellent | 21:00 |
| mordred | dtroyer: oh - actually, turns out I may not actually need it to be deleted | 21:03 |
| mordred | so never mnd | 21:03 |
| *** briancurtin has quit IRC | 21:04 | |
| mordred | k. your patch works in my stack | 21:04 |
| mordred | I think it may be time to add a bunch of tests ... | 21:04 |
| dtroyer | nice… | 21:04 |
| mordred | dtroyer: so, I know you probably have other things on your tdl | 21:07 |
| mordred | but master osc on my laptop: openstack server list returns nothing - but nova list shows a server | 21:07 |
| mordred | do you want me to dig in and diagnose that? | 21:07 |
| dtroyer | —debug might be interesting | 21:08 |
| mordred | dtroyer: may be az related | 21:09 |
| dtroyer | ah, could be…I don't have that to test with anywhere | 21:09 |
| mordred | it _looks_ like it's making the request against only one az | 21:09 |
| mordred | https://az-1.region-a.geo-1.compute.hpcloudsvc.com/v1.1 | 21:09 |
| dtroyer | dog park is calling…gotta run | 21:12 |
| mordred | Ihave fun | 21:12 |
| *** etoews has joined #openstack-sdks | 21:16 | |
| mordred | dtroyer: ZOMG ZOMG ZOMG | 21:21 |
| mordred | it's in keystoneclient | 21:21 |
| mordred | I'm going to stab something | 21:21 |
| *** etoews has quit IRC | 21:21 | |
| mordred | ah. nope. it's not. one sec ... | 21:22 |
| openstackgerrit | Monty Taylor proposed a change to openstack/python-openstackclient: Fix operation on clouds with availability-zones https://review.openstack.org/127811 | 21:26 |
| openstackgerrit | Monty Taylor proposed a change to openstack/python-openstackclient: Fix operation on clouds with availability-zones https://review.openstack.org/127811 | 21:30 |
| mordred | dtroyer: k. that fixes it in local testing | 21:31 |
| *** etoews has joined #openstack-sdks | 21:34 | |
| *** etoews has quit IRC | 21:40 | |
| *** briancurtin has joined #openstack-sdks | 22:00 | |
| *** ayoung has quit IRC | 22:03 | |
| *** briancurtin has quit IRC | 22:05 | |
| *** bknudson has joined #openstack-sdks | 22:23 | |
| openstackgerrit | OpenStack Proposal Bot proposed a change to openstack/python-openstackclient: Updated from global requirements https://review.openstack.org/126682 | 22:38 |
| *** briancurtin has joined #openstack-sdks | 23:01 | |
| *** briancurtin has quit IRC | 23:05 | |
| *** sigmavirus24_awa is now known as sigmavirus24 | 23:29 | |
| *** stevemar has joined #openstack-sdks | 23:36 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!