openstackgerrit | Chris Robinson proposed a change to stackforge/golang-client: Add the ability to create, delete, list, and get networks, ports and subnets https://review.openstack.org/131607 | 00:08 |
---|---|---|
*** rmcall_ has joined #openstack-sdks | 00:10 | |
*** rmcall has quit IRC | 00:12 | |
*** rmcall_ is now known as rmcall | 00:12 | |
*** bknudson has joined #openstack-sdks | 00:38 | |
*** rmcall_ has joined #openstack-sdks | 00:43 | |
*** mattfarina has joined #openstack-sdks | 00:44 | |
*** rmcall has quit IRC | 00:44 | |
*** rmcall_ is now known as rmcall | 00:44 | |
*** mattfarina has quit IRC | 00:45 | |
*** terrylhowe has quit IRC | 00:51 | |
*** bknudson has left #openstack-sdks | 00:56 | |
*** ayoung has joined #openstack-sdks | 01:16 | |
*** sigmavirus24_awa is now known as sigmavirus24 | 01:42 | |
*** sigmavirus24 is now known as sigmavirus24_awa | 02:17 | |
*** tellesnobrega_ has joined #openstack-sdks | 02:36 | |
*** tellesnobrega_ has quit IRC | 02:49 | |
*** tellesnobrega_ has joined #openstack-sdks | 02:51 | |
*** ayoung is now known as ayoung-ZZZzz | 03:13 | |
*** tellesnobrega_ has quit IRC | 03:51 | |
*** tellesnobrega_ has joined #openstack-sdks | 03:52 | |
*** tellesnobrega_ has quit IRC | 04:22 | |
*** rmcall has quit IRC | 04:51 | |
*** rmcall has joined #openstack-sdks | 04:51 | |
*** briancurtin has quit IRC | 04:52 | |
*** rmcall has quit IRC | 05:06 | |
*** rmcall has joined #openstack-sdks | 05:08 | |
*** rmcall has quit IRC | 05:39 | |
*** rmcall has joined #openstack-sdks | 05:39 | |
*** rmcall has quit IRC | 06:11 | |
*** rmcall_ has joined #openstack-sdks | 06:11 | |
*** rmcall_ has quit IRC | 06:15 | |
*** rmcall has joined #openstack-sdks | 06:16 | |
*** rmcall has quit IRC | 06:20 | |
*** rmcall has joined #openstack-sdks | 06:23 | |
*** rmcall has quit IRC | 06:28 | |
*** rmcall has joined #openstack-sdks | 06:31 | |
*** rmcall has quit IRC | 07:22 | |
*** rmcall has joined #openstack-sdks | 07:40 | |
*** rmcall has quit IRC | 07:45 | |
*** notstevemar has quit IRC | 08:14 | |
*** k4n0 has joined #openstack-sdks | 08:28 | |
*** terrylhowe has joined #openstack-sdks | 10:49 | |
jamielennox | dtroyer: when you get in let me know, i'm looking through OSC auth plugins | 11:12 |
terrylhowe | Dean was probably up late last night watching the world series or something, might be a while | 11:15 |
*** tellesnobrega_ has joined #openstack-sdks | 11:25 | |
*** tellesnobrega_ has quit IRC | 12:03 | |
jamielennox | terrylhowe: did you have anything to do with the plugins in OSC? | 12:09 |
terrylhowe | no jamielennox I’ve been watching a little though | 12:09 |
jamielennox | terrylhowe: so has it been in a release yet? | 12:10 |
terrylhowe | I don’t think so, but not sure | 12:10 |
jamielennox | i'm trying to figure out what compatibility i need to maintain | 12:10 |
jamielennox | but i'd like to rip most of this out before it gets released | 12:10 |
jamielennox | most/some | 12:10 |
terrylhowe | rip out of ksc or osc? | 12:11 |
jamielennox | osc, a bunch of the stuff that was done custom for osc has an equivalent already in ksc | 12:12 |
jamielennox | i would like OSC to use it in the same way as all the other CLIs | 12:12 |
jamielennox | standardize it | 12:12 |
terrylhowe | ah | 12:13 |
*** ayoung-ZZZzz is now known as ayoung | 12:33 | |
*** jxxxxx has joined #openstack-sdks | 12:46 | |
*** Krast has joined #openstack-sdks | 12:48 | |
*** terrylhowe has quit IRC | 12:50 | |
*** terrylhowe has joined #openstack-sdks | 12:58 | |
*** mattfarina has joined #openstack-sdks | 13:03 | |
*** tellesnobrega_ has joined #openstack-sdks | 13:18 | |
*** terrylhowe has quit IRC | 13:23 | |
*** bknudson has joined #openstack-sdks | 13:24 | |
*** terrylhowe has joined #openstack-sdks | 13:26 | |
*** briancurtin has joined #openstack-sdks | 13:28 | |
*** Krast has quit IRC | 13:43 | |
*** sigmavirus24_awa is now known as sigmavirus24 | 13:53 | |
openstackgerrit | Terry Howe proposed a change to stackforge/python-openstacksdk: Working script to create a jenkins server https://review.openstack.org/121660 | 14:00 |
*** terrylhowe has quit IRC | 14:00 | |
*** stevemar has joined #openstack-sdks | 14:28 | |
*** terrylhowe has joined #openstack-sdks | 14:36 | |
*** terrylhowe has left #openstack-sdks | 14:38 | |
*** tellesnobrega_ has quit IRC | 15:10 | |
*** rmcall has joined #openstack-sdks | 15:24 | |
openstackgerrit | Jamie Lennox proposed a change to openstack/python-openstackclient: Redo auth plugins https://review.openstack.org/131804 | 15:49 |
*** rmcall has quit IRC | 16:03 | |
*** k4n0 has quit IRC | 16:03 | |
*** rmcall has joined #openstack-sdks | 16:04 | |
*** terrylhowe has joined #openstack-sdks | 16:56 | |
dtroyer | jamielennox: Am enroute to Paris today, so will be on and off for a while. | 17:05 |
jamielennox | dtroyer: ok, i uploaded a review that's WIP and will fail all kinds of tests | 17:05 |
dtroyer | jamielennox: mhu did the bulk of the auth plugin work, I (tried to) streamline the integration into ClientManager. It hasn't been released yet so changes are still possible. | 17:05 |
jamielennox | dtroyer: but it's much more along how i would hope the plugins would work | 17:06 |
jamielennox | i'm not really sure how to get from a to b though | 17:06 |
dtroyer | jamielennox: also have a look at the WIP change for os-client-config as it changes how I want to do the config options. but that'll come after whatever plugin changes we need to do | 17:06 |
jamielennox | if you get a chance to have a look at the review i think it would be useful to even understand what i mean by the difference | 17:06 |
jamielennox | dtroyer: ok - i don't know that one but i'll look it up | 17:07 |
dtroyer | I'll have a look. CloudConfig is a new lib from mordred that I've glommed on to save multiple cloud configs in a yaml file so I can do something like: openstack —os-cloud rax server list | 17:08 |
jamielennox | cloud-config or client-config | 17:08 |
jamielennox | gerrit has a project:openstack/os-cloud-config | 17:08 |
dtroyer | CloudConfig is the object name, the project is os-client-config | 17:09 |
dtroyer | blame mordred, he was 'in a mood' when he started that one ;) | 17:09 |
mordred | what's that? | 17:09 |
mordred | jamielennox: yes - os-client-config - not cloud-config | 17:10 |
mordred | totally different thing | 17:10 |
mordred | and I cannot WAIT to have openstack --os-cloud rax server list working | 17:10 |
jamielennox | ok, so it's a library for reading config opts from files | 17:11 |
jamielennox | i was reading the description and got a little concerned it was another client | 17:11 |
jamielennox | mordred: any thoughts on how that will support different plugins or leaving that one up in the air? | 17:12 |
dtroyer | jamielennox: to keep back-compat, I'm planning to 'guess' when the old set of options are provided, everything else will require —os-auth-type to specify | 17:13 |
jamielennox | dtroyer: don't guess - write a plugin that takes the options you need and proxies to the plugin you actually choose | 17:13 |
dtroyer | ie, token+url == old token auth, username+password == Password, token+auth_url == Token | 17:13 |
jamielennox | dtroyer: have a look at https://review.openstack.org/#/c/131804/1/openstackclient/api/auth.py | 17:14 |
jamielennox | my intention was that you could specify --os-auth-plugin (i used -plugin, i don't know where -type came from, i don't care but i would like it standard) and if you didn't give one you'd get a default | 17:15 |
dtroyer | I changes -plugin to -type because plugin is meaningless to most users | 17:16 |
dtroyer | and osc already uses 'plugin' to refer to the command extensions | 17:16 |
jamielennox | yea, it's a better word, would prefer we use the standard mechanism though and deprecate/rename it in ksc | 17:17 |
dtroyer | I'll look at that review closer, I think I see the approach you are taking, I hadn't considered doing that… | 17:17 |
jamielennox | the way it's written at the moment if you do openstack --help you'll get a lot of options and growing | 17:18 |
jamielennox | i think we definitely want to limit that to openstack --os-auth-type XXX --help | 17:18 |
dtroyer | right, that I want to get a handle on, was one reason I was trying to take over the option handling…CLoudconfig changes that again too | 17:19 |
dtroyer | exactly | 17:19 |
jamielennox | ok, well we can talk about it next week. i'll see if i can make those tests pass but i doubt it without some serious ripping out. | 17:20 |
jamielennox | i heard a rumour though you were looking for an OSC release soon and i wanted to suggest this before there are compat issue | 17:20 |
jamielennox | s | 17:20 |
dtroyer | I was, actually wanted to do it last week…but I wantd to get —os-cloud included and it wasn't ready… | 17:21 |
jamielennox | dtroyer, mordred: do we see os-client-config as a place to cache auth between requests or just a way to read options? | 17:24 |
dtroyer | I'm using it to supply the Namespace-like object to the ClientManager, which then goes into the auth-plugin business | 17:25 |
*** terrylhowe has quit IRC | 17:25 | |
jamielennox | dtroyer: right, but there's no reason that os-client-config couldn't just return you the plugin to use for the requests, and in which case it would be the logical place to cache as it will know when options change | 17:26 |
dtroyer | eventually I see one ClientManager per CloudConfig object, with Cloudconfig caching the creds in its auth_ref (or whatever) | 17:26 |
dtroyer | I think mordred has other (non python-*client) use cases so I'm not sure if that fits in well there | 17:27 |
jamielennox | if we can move all the clients to a sessions based approach i actually think that clientmanager can just fade away | 17:27 |
mordred | jamielennox: I'll defer to dtroyer on that - I don't really have any clouds that use different auth, nor do I use auth cred caching anywhere, so I'm not a great person to know much about the use cases | 17:27 |
dtroyer | ah, ok, cool | 17:27 |
mordred | my main thing is that I want to easily be able to configure and use client things | 17:27 |
mordred | and I want it to be VERY simple - beacuse right now it's one of the most complex things ever | 17:27 |
jamielennox | mordred: you want to work this into all the CLIs or just move the other CLIs to using OSC? | 17:28 |
mordred | so, essentially, os-cloud-config is feature complete for me - but if there is something else that should go in, I'm 100% game for that | 17:28 |
mordred | jamielennox: I'd _like_ to move the current clients to it too - just because they exist now and OSC still hasn't fully caught on and I still use python-*client | 17:29 |
mordred | but I hadn't gotten as far as writing up a spec for that yet :) | 17:29 |
mordred | I could see reasonable people disagreeing | 17:29 |
jamielennox | it'd be a lot of work, we haven't had much luck standardizing anything with clients | 17:30 |
mordred | it's also possible that once it goes in to OSC that I might just ditch using other things, start using osc and never look back :) | 17:30 |
jamielennox | mordred: ++ that would be the goal i think | 17:30 |
dtroyer | ClientManager is now pretty self-contained and makes a decent single-point-of-contact for the clients it knows about so you could even ditch the shell… | 17:31 |
openstackgerrit | Ryan Brown proposed a change to openstack/python-openstackclient: Add Heat stack CRUD https://review.openstack.org/111786 | 17:32 |
mordred | dtroyer: I was actually going to play with what it would look like for shade to consume the osc interface you had that example of using as an API up for | 17:33 |
mordred | I ahave not gotten to that yet | 17:33 |
jamielennox | dtroyer: client manager is still useful as a point of contact, but its initial purpose was to cache/standardize all the auth. If we go session base then actual client creation is cheap | 17:33 |
jamielennox | shade? | 17:33 |
dtroyer | jamielennox: it's already totally Session-based for the object-store API ;) | 17:34 |
mordred | jamielennox: shade is a library that came from refactoring a bunch of work I had inside of ansible into a library | 17:34 |
jamielennox | dtroyer: swift? i haven't looked at python-swiftclient for ages but that is surprising | 17:34 |
mordred | jamielennox: I dither on whether it wants to exist or whether it wants to be patches to something | 17:35 |
mordred | jamielennox: at the moment it exists because patches takea while and I need it now :) | 17:35 |
dtroyer | we don't use swiftclient, at the time I started it wasn't really suitable | 17:35 |
* jamielennox is learning all sorts of new words today | 17:35 | |
mordred | but I would not be sad if it went away | 17:35 |
jamielennox | dtroyer: ah - right, well it's kind of easy if you just cheat like that | 17:35 |
mordred | jamielennox: it handles things like "please give me a server with a working IP" and hides the deployer diferences like rackspace giving you a public ip automatically or HP needing you to ask for a floating ip | 17:36 |
dtroyer | jamielennox: that may not be the last one either…reducing dependencies…general sanity, etc | 17:36 |
*** rmcall_ has joined #openstack-sdks | 17:36 | |
jamielennox | mordred: makes sense, possible sdk targets but i wouldn't hold my breath | 17:37 |
mordred | yah | 17:37 |
mordred | I think that's a medium-long term goal | 17:37 |
*** rmcall has quit IRC | 17:37 | |
*** rmcall_ is now known as rmcall | 17:37 | |
jamielennox | dtroyer: i can't tell if that would be a shame to have to rewrite all that or just logical | 17:38 |
mordred | current very naive thinking is step 1) port shade to osc 2) let osc port to sdk 3) move shade into to osc OR sdk as a "simple" api | 17:38 |
jamielennox | true, ansible is native python so it can do sdk | 17:39 |
jamielennox | anyway we can talk about this next week - just wanted to put a possible pin in the OSC release | 17:39 |
dtroyer | jamielennox: I choose logical…it's basically what I wanted for the low-level API in the SDK that didn't get any traction | 17:39 |
jamielennox | gotta run for now | 17:39 |
dtroyer | ditto…later y'all | 17:39 |
*** jamielennox is now known as jamielennox|away | 17:40 | |
*** fifieldt has quit IRC | 17:43 | |
*** tellesnobrega has joined #openstack-sdks | 17:50 | |
*** rmcall has quit IRC | 17:54 | |
*** rmcall has joined #openstack-sdks | 17:55 | |
*** tellesnobrega has quit IRC | 17:55 | |
*** rmcall has quit IRC | 18:19 | |
*** rmcall has joined #openstack-sdks | 18:20 | |
openstackgerrit | Chris Robinson proposed a change to stackforge/golang-client: Add ability to create, delete, list, and get keypairs https://review.openstack.org/131849 | 18:23 |
*** jxxxxx has quit IRC | 18:44 | |
*** rmcall has quit IRC | 19:05 | |
*** devx_ is now known as devx | 20:26 | |
*** mattfarina has quit IRC | 21:06 | |
*** terrylhowe has joined #openstack-sdks | 21:07 | |
openstackgerrit | Terry Howe proposed a change to stackforge/python-openstacksdk: Working script to create a jenkins server https://review.openstack.org/121660 | 21:07 |
*** stevemar has quit IRC | 21:20 | |
*** terrylhowe has quit IRC | 21:21 | |
*** devx is now known as XveD | 21:41 | |
*** openstackgerrit has quit IRC | 21:50 | |
*** briancurtin has quit IRC | 21:58 | |
*** bknudson has quit IRC | 22:26 | |
*** briancurtin has joined #openstack-sdks | 22:38 | |
*** ycombinator_ has joined #openstack-sdks | 23:39 | |
*** ycombinator_ has quit IRC | 23:41 | |
*** glenc has quit IRC | 23:42 | |
*** ycombinator has joined #openstack-sdks | 23:42 | |
*** glenc has joined #openstack-sdks | 23:43 | |
*** glenc has quit IRC | 23:46 | |
*** glenc has joined #openstack-sdks | 23:47 | |
*** glenc has quit IRC | 23:48 | |
*** ycombinator has quit IRC | 23:49 | |
*** glenc has joined #openstack-sdks | 23:50 | |
*** bknudson has joined #openstack-sdks | 23:56 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!