Earlier  
Later
New Search
Posted Nick Remark
#openstack-sdks - 2017-04-03
18:08:32 sdague elmiko: I do kind of wonder if zaneb's ideas around application tokens might get us back headed towards that path
18:08:55 elmiko sdague: i have not seen that written up, but it sounds interesting
18:09:07 cdent elmiko: https://review.openstack.org/#/c/447031/
18:09:13 elmiko cdent++
18:09:29 sdague edleafe: sure, I applaud that effort. However, I also feel like we've got a set of services where this matches well, and it would be nice to get that flag in the ground.
18:09:32 cdent (in the links in the bottom)
18:10:20 dtroyer ZZelle: hey
18:11:02 elmiko i totally agree with the thought that for certain "core" services, mandating microversions is a much stronger idea. but that takes us away from the big tent idea
18:12:01 cdent (I'm going to need to go at any moment now, I think, if there's a concrete outcome to this conversation that is germane to the doc, can somebody leave a new comment there so I can act on it?)
18:12:12 elmiko cdent: ack
18:12:17 cdent thanks
18:12:39 ZZelle dtroyer, about your comment in https://review.openstack.org/#/c/452328/1/doc/source/command-objects/server.rst
18:12:50 ZZelle dtroyer, i am not sure to understand what you mean
18:13:33 cdent I think being more specific about the scope of the doc and the rigor of the standard being set is probably a good way to go, especially if we leave open the option (we always do) for more docs later
18:13:34 cdent That moves the doc forward
18:13:45 cdent but I don't know that that helps us with what might be called the social problem.
18:13:58 cdent but it may be we don't need to do anything about that
18:14:06 dtroyer its just the form of the help line, we put the '(name or ID)' test at the end pretty much everywhere, except some differences have snuck in. Fix yours, I'll get the other ones
18:14:26 dtroyer there's an RST error in there too that stevemar noted but +2'd anyway
18:14:53 ZZelle dtroyer, so something like 'Server ... (name or ID).' instead of 'Server (name or ID) ...'?
18:15:04 stevemar ZZelle: yep
18:15:13 ZZelle stevemar, dtroyer good for me
18:15:17 dtroyer yes. look at nearly every other command in that file than the add commands near yours
18:15:18 elmiko cdent: imo, for a guideline that will define a tag, it should be as explicit as possible in terms of the hoops needed to jump through
18:15:37 elmiko so, that may speak to having 2 separate docs
18:17:23 cdent I remain a bit unclear on what the second doc is
18:18:48 elmiko i guess that would be needed if we couldn't make the first one specific enough?
18:19:02 elmiko or just having 2 tags or something, one for interop and one for ?
18:19:32 cdent yes, that blank is what I'm blank on
18:19:48 openstackgerrit Cedric Brandily proposed openstack/python-openstackclient master: Enable to add/remove port to/from a server https://review.openstack.org/452328
18:19:54 ZZelle stevemar, dtroyer ^^
18:21:38 elmiko cdent: yeah, sadly, i have no better answers here =(
18:22:07 dtroyer ZZelle: thanks
18:22:42 cdent elmiko: no worries. if/when edleafe comes back he may have something, I gotta run to dinner, thanks for the input, we'll figure something out.
18:22:53 elmiko cdent: cool, enjoy o/
18:23:35 stevemar ZZelle: ohhh are you cedric?
18:27:01 sshank dtroyer, I think https://review.openstack.org/#/c/403907/ (flavor associate, dissociate) is ready for reviews.
18:28:29 openstackgerrit Ankur proposed openstack/python-openstacksdk master: Introduce Base for Octavia (load balancing) https://review.openstack.org/428414
18:32:49 openstackgerrit Ankur proposed openstack/python-openstackclient master: Network L3 Router Commands for OSC https://review.openstack.org/385729
18:39:15 dtroyer ankur-gupta-f1: re https://review.openstack.org/#/c/449757/, why are you filtering the show output? we've removed specific fields before, I don't understand why you are treating this one like a list command
18:55:31 openstackgerrit Merged openstack/python-openstackclient master: Add help commands withouth auth in functional https://review.openstack.org/452407
19:01:51 ankur-gupta-f4 dtroyer: based on the api response. Since the advanced commands hit different endpoints that the normal list commands
19:02:06 ankur-gupta-f4 I can go back thru them do verify if you want
19:03:20 dtroyer what happened to advanced commands going into plugins?
19:04:24 ankur-gupta-f4 Because its not advanced in the sense that certain features have to be enabled to use the commands. Its just a complex endpoint
19:05:06 dtroyer right, that's what I thought made it different. 'advanced services' has a specific meaning in neutron
19:05:39 ankur-gupta-f4 Yea. Will reword it. Sorry for confusion
19:06:00 ZZelle stevemar, yes i am
19:06:15 dtroyer but that doesn't change a) different endpoint and b) the same resource isn't represented by the same fields?
19:06:25 dtroyer what endpoint is it?
19:07:11 ankur-gupta-f4 i.e. for routers for the list command it hits /v2.0/routers for the list routers --agent <agent-id> it hits /v2.0/routers/<router-id>/l3-agents i believe
19:07:29 dtroyer ok, again, endpoint has a specific meaning: service catalog thing
19:07:42 dtroyer that's a route in wsgi terms
19:07:53 dtroyer so much overloading of words!
19:07:53 ankur-gupta-f4 okay. route.
19:08:12 dtroyer so the server returns a different object in that case?
19:08:27 ankur-gupta-f4 correct
19:08:40 dtroyer and the SDK doesn't fill it out? this is one place I'd expect the SDK's high-level-ness to step in and make things look right
19:08:41 ankur-gupta-f4 https://github.com/openstack/python-openstacksdk/blob/master/openstack/network/v2/router.py#L134 is how the route is determined in the SDK
19:09:37 dtroyer we should just start issuing random 3 letter names for all of these different things
19:10:53 ankur-gupta-f4 route, complex command, idiotic API.. should suffice
19:11:08 dtroyer so, back to the issue at hand, is —router a filter or a type specifier?
19:13:03 ankur-gupta-f4 not a filter in the sense that it filters the resulting data from hitting the common list route. rather a flag to indicate that instead of hitting the usual list route, to hit the specific route that would return a different routers list object from the API.
19:15:20 dtroyer ok, the problem is that the columns are different
19:16:20 dtroyer in your examples the router ID is the same so it is that the server returns a different object, except it is still the same router at each endpoint?
19:17:06 ankur-gupta-f4 yes
19:17:36 ankur-gupta-f4 well..
19:19:10 ankur-gupta-f4 the example I posted, the paste.openstack.org link, was just to show that the commands are still listing the correct items
19:20:23 dtroyer so I clearly still don't understand wtf neutron is doing here. If I were to say: make a router list command always show the same columns, eept for -c and —long, what would you suggest?
19:21:49 ankur-gupta-f4 I just ran the commands in --debug to look at the dicts.
19:21:54 openstackgerrit Nakul Dahiwade proposed openstack/python-openstacksdk master: Introduce Listener for Octavia (load balancing) https://review.openstack.org/451574
19:22:31 ankur-gupta-f4 It should be possible to have all the list commands return the same values.
19:23:02 ankur-gupta-f4 what we can do is have all the list commands have the same columns.
19:23:10 ankur-gupta-f4 then if the certain agent flag is given
19:23:15 ankur-gupta-f4 append the additional columns
19:24:01 dtroyer that would be good. if we need additional options to select specic column similar to —long, we can talk about that… say for a common use case
19:26:28 ankur-gupta-f4 dtroyer: thanks. Headed to airport now. Will deal with all that shit Wednesday
19:26:40 dtroyer ok, thanks
19:27:08 ankur-gupta-f4 sshank: here? you will need to update your DHCP-network commands to reflect ^^^^^^
19:27:40 sshank ankur-gupta-f4, Yes. I'll need to read through the logs.
19:30:52 ankur-gupta-f4 basically have the list commands use the common columns from the base list case. and do like a columns = columns + (whatever additional your specific list command adds on (like ha)) column_headers = column_headers + (whatever)
19:33:26 openstackgerrit Ankur proposed openstack/python-openstacksdk master: Introduce Base for Octavia (load balancing) https://review.openstack.org/428414
19:47:16 openstackgerrit Shashank Kumar Shankar proposed openstack/python-openstacksdk master: Introduce Pool for Octavia (load balancing) https://review.openstack.org/449264
19:49:43 openstackgerrit Merged openstack/python-openstackclient master: Fix block-device-mapping when volume_size is empty https://review.openstack.org/451432
20:05:33 dtroyer stevemar: if you can spare a minute have a look at https://review.openstack.org/#/c/450453/, it's finally passing everything and fixes our -tips jobs
20:06:52 stevemar ++
20:10:55 dtroyer thanks
20:28:58 stevemar dtroyer: want me to push it through or wait til osc-lib is released?
20:30:26 dtroyer this is all for master, go ahead
20:30:54 dtroyer I'm working on releases, hopefully this week as I'm limited availability next week
20:48:30 openstackgerrit Nakul Dahiwade proposed openstack/python-openstacksdk master: Introduce Base for Octavia (load balancing) https://review.openstack.org/428414
20:51:20 openstackgerrit Merged openstack/python-openstackclient master: Enable to add/remove port to/from a server https://review.openstack.org/452328
21:09:52 openstackgerrit Nakul Dahiwade proposed openstack/python-openstacksdk master: Introduce Listener for Octavia (load balancing) https://review.openstack.org/451574
21:10:54 openstackgerrit Nakul Dahiwade proposed openstack/python-openstacksdk master: Introduce Listener for Octavia (load balancing) https://review.openstack.org/451574
21:11:39 openstackgerrit Nakul Dahiwade proposed openstack/python-openstacksdk master: Introduce Listener for Octavia (load balancing) https://review.openstack.org/451574
21:23:39 openstackgerrit Dean Troyer proposed openstack/python-openstackclient master: Help/docs cleanups: marker, limit, ip-address metavars https://review.openstack.org/452961
21:32:28 openstackgerrit Dean Troyer proposed openstack/python-openstackclient master: Release notes cleanup for 3.10.0 release https://review.openstack.org/452965
21:38:40 dtroyer jamielennox: when you get a sec have a look at https://review.openstack.org/#/c/452711/ and tell me "I told you that long ago" if this is what I think it is :)
21:50:39 openstackgerrit Merged openstack/python-openstackclient master: Change noauth strategy for plugin loading https://review.openstack.org/450453
22:16:13 openstackgerrit Nakul Dahiwade proposed openstack/python-openstacksdk master: Introduce Base for Octavia (load balancing) https://review.openstack.org/428414

Earlier   Later