| Posted | Nick | Remark | |
|---|---|---|---|
| #openstack-sdks - 2018-03-05 | |||
| 12:23:23 | dtroyer | I'm going to add an osc functional test to every client we depend on to keep this nonsense from happening again | |
| 12:24:16 | dtroyer | long-term, we kick all of those legacy decrepit old clients to the curb and replace them with the SDK, once it reaches 1.0 | |
| 12:24:31 | dtroyer | I've been holding my breath for that for 3 years now | |
| 12:24:58 | cdent | such stamina! | |
| 12:25:42 | openstackgerrit | Tobias Brox proposed openstack-infra/shade master: Add subnetpool_name_or_id parameter to create_subnet https://review.openstack.org/548722 | |
| 12:30:28 | openstackgerrit | Tobias Brox proposed openstack-infra/shade master: Add subnetpool_name_or_id parameter to create_subnet https://review.openstack.org/548722 | |
| 12:31:39 | dtroyer | cdent: how do you think I got through the snow? Also https://github.com/dtroyer/osc-snowpenstack, although I'm afraid it isn't very far along yet | |
| 12:31:54 | openstackgerrit | Tobias Brox proposed openstack-infra/shade master: Add subnetpool_name_or_id parameter to create_subnet https://review.openstack.org/548722 | |
| 12:32:26 | tobixen | sorry for the noise above | |
| 12:34:26 | cdent | dtroyer: if you keep this up, there will be some kind of crisis at every ptg just so you're forced to make these things | |
| 12:36:52 | dtroyer | naw, I'm liking this new Friday evening PTG tradition, it replaces the Friday evening summit pow-wow with jamielennox and our plotting to take over the world. | |
| 12:41:37 | mordred | dtroyer: ++ to kicking to the curb | |
| 12:55:18 | mordred | dtroyer: https://review.openstack.org/549749 Release 0.12.0 of python-openstacksdk | |
| 12:55:27 | mordred | dtroyer: just pushed that up - as soon as it lands the other issue should go away | |
| 12:58:07 | dtroyer | mordred: cool, thanks! | |
| 13:14:47 | openstackgerrit | Kairat Kushaev proposed openstack/keystoneauth master: use defusedxml for XML parsing https://review.openstack.org/536761 | |
| 14:42:20 | openstackgerrit | Monty Taylor proposed openstack/python-openstackclient master: Stop using novaclient for removed actions https://review.openstack.org/549777 | |
| 14:44:24 | mordred | dtroyer: I've got an appointment I have to get to this morning, so that ^^ is only half-done. however, I think with about an hour-ish more of work we can at least remove osc's use of the novaclient stuff that got removed - that patch also surfaced a place where an improvement could be made to oneof the sdk apis | |
| 14:45:29 | dtroyer | mordred: heh, ok. I was nearly done with the os-host bits doing it the way we did the last removed APIs | |
| 14:45:41 | mordred | dtroyer: heh. :) | |
| 14:47:00 | mordred | dtroyer: well, if there's a different way to deal with os-host, while I think we *can* add objects for it to sdk, it's a pretty terrible api so I wouldn't be sad if we didn't add support to sdk for it at this point either | |
| 14:48:13 | dtroyer | since I have that done almost, you skip it and I'll finish my addition to api.compute_v2 | |
| 14:48:56 | dtroyer | does the SDK have any specific handling for the microversion header? | |
| 14:54:52 | mordred | dtroyer: right now it's just a passthrough to the ksa support - so any of the rest calls can take a microversion= parameter and ksa will pass the headers correctly. | |
| 14:56:23 | mordred | dtroyer: I think we're at the point where it's time to start adding microversion logic to calls in sdk as we hit them though | |
| 14:57:34 | dtroyer | mordred: agreed. lets not hold up 1.0 for that though? | |
| 14:57:42 | mordred | oh golly no | |
| 14:57:53 | mordred | totally agree | |
| 14:58:26 | umbSublime | Hi, I'm trying to find equivalent of compute /os-services/disable in openstack, but I can't find it. I'd like to still use my connection to do the custom call. In the case of instances there was the _action method where I could pass some custom_body. Is there equivalent for hypervisors ? | |
| 15:00:54 | dtroyer | umbSublime: (I just happen to have the Compute api-ref in a foreground window :) https://developer.openstack.org/api-ref/compute/#hypervisors-os-hypervisors says there is not, those are all read-only APIs | |
| 15:01:58 | mordred | dtroyer: three things on my pre-1.0 list - one is finishing plumbing version discovery through (I've got code locally that's 95% done, just need to tie it off) ... I want to finish removing the service_filter objects (which depends on finishing version discovery) - and I'd like to remove the dict/__getitem__ passthrough magic in the os-client-config layer since that's the root of most of the | |
| 15:02:00 | mordred | weird/confusing error messages | |
| 15:03:08 | mordred | dtroyer: it's not unreasonable to get that pretty much done in the next week or two | |
| 15:08:41 | dtroyer | mordred: ok, good to hear. My 2 hours a week on this isn't helping, I know, but damn we need to get something out. I'm about to send a rant to the ML responding to Andrey/Matt that's get me flame roasted I'm sure. I'm back to working on OSC in my spare time but dammit this shit has to stop | |
| 15:13:18 | umbSublime | thanks dtroyer I guess since it's in the compute api, I can just use _action method | |
| 15:14:14 | dtroyer | mordred: also, I saw somewhere a reference to places where the SDK might be used by services. for the love of all the snow in Dublin PLEASE NO! The fundamental assumptions are different and I don't want clients to have anything to do with the crap that services want. | |
| 15:33:40 | TheJulia | The consensus with nova -> ironic was to just migrate to rest calls where we are not using the client library for ironic, so we’re better able to navigate api version issues. Explicitly not use the SDK, fwiw. | |
| 15:34:07 | TheJulia | As the behaviors/needs are different | |
| 15:37:29 | dtroyer | TheJulia: exactly and thank you. | |
| 15:37:50 | dtroyer | I wish I could remember who it was talking about that... | |
| 15:39:52 | TheJulia | There was some other chatter, but I realize my preference to just do rest calls is heavily influenced by experience of pain. | |
| 16:18:46 | openstackgerrit | Tobias Brox proposed openstack-infra/shade master: Allow grant_role to select users in a different domain https://review.openstack.org/549744 | |
| 16:39:12 | openstackgerrit | Dean Troyer proposed openstack/python-openstackclient master: Re-implement more removed novaclient bits https://review.openstack.org/549813 | |
| 16:40:00 | dtroyer | mordred: ^^^ is the os-hosts bit, it still needs tests for the zpi.compute_v2 methods but I'm landing and wanted to see if functional tests pass | |
| 16:54:36 | openstackgerrit | Andrey Kurilin proposed openstack/python-openstackclient master: Rework "openstack server add floating ip" cmd https://review.openstack.org/549820 | |
| 20:19:13 | openstackgerrit | Dean Troyer proposed openstack/python-openstackclient master: WIP: Example of converting floating IP command to nova-net/neutron split https://review.openstack.org/549864 | |
| 20:21:25 | dtroyer | mordred: ^^^^ is an example that leverages the bits we already have done for switching between nova-net and neutron implementations | |
| 22:10:22 | jamielennox | dtroyer: i miss our pow-wows as well, sad i missed it | |
| #openstack-sdks - 2018-03-06 | |||
| 05:30:32 | Qiming | mordred, the API microversioning is broken in SDK? We were passing OpenStack-API-Version headers properly before, but in recent versions, the header is gone ... | |
| 06:54:42 | openstackgerrit | Kairat Kushaev proposed openstack/keystoneauth master: use defusedxml for XML parsing https://review.openstack.org/536761 | |
| 09:03:35 | openstackgerrit | Olivier Bourdon proposed openstack-infra/shade master: Add proper return value for validate_node https://review.openstack.org/549981 | |
| 14:32:33 | openstackgerrit | Corey Bryant proposed openstack/cliff master: Align parsed() call with cmd2 versions >= 0.7.3 https://review.openstack.org/548625 | |
| 14:39:16 | mordred | Qiming: keystoneauth should be passing the version headers for any place where we've requested a microversion ... but to my knowledge we're not currently negotiating any microversions ... can you tell me more about a place where we should be sending a microversion? | |
| 14:40:54 | slaweq | mordred: hi | |
| 14:41:47 | slaweq | mordred: do You know maybe why I don't have possibility to give +2/-2 and W+1 in python-openstacksdk anymore? | |
| 14:42:03 | slaweq | mordred: did You remove me maybe from core group for this project? | |
| 14:42:58 | slaweq | mordred: for example here: https://review.openstack.org/#/c/549474/ I can give only +1 and -1 | |
| 14:44:10 | mordred | slaweq: ah! I put up a project-config change to make an openstacksdk team that was applied to all the correct repos - I'm guessing it landed - I need to go add people to it ... one sec | |
| 14:44:38 | slaweq | mordred: sure, I just though that I wasn't good enough and You removed me :) | |
| 14:45:01 | slaweq | and I was little bit sad because of that :( | |
| 14:47:23 | mordred | slaweq: done! | |
| 14:47:36 | mordred | slaweq: no - defintely not :) | |
| 14:47:46 | slaweq | mordred: thx, now I have +2 again :) | |
| 14:47:50 | mordred | \o/ | |
| 15:01:04 | openstackgerrit | Dean Troyer proposed openstack/python-openstackclient master: WIP: Example of converting floating IP command to nova-net/neutron split https://review.openstack.org/550105 | |
| 15:01:57 | dtroyer | mordred: ^^^ what 12 hours of travel without $DAY_JOB email gets | |
| 15:07:14 | openstackgerrit | Olivier Bourdon proposed openstack-infra/shade master: Add extra failure codes to bad request exception https://review.openstack.org/550107 | |
| 15:23:05 | openstackgerrit | Matt Riedemann proposed openstack/python-openstackclient master: Default --nic to 'auto' if creating a server with >= 2.37 https://review.openstack.org/545946 | |
| 15:33:40 | openstackgerrit | Corey Bryant proposed openstack/cliff master: Align parsed() call with cmd2 versions >= 0.7.3 https://review.openstack.org/548625 | |
| 16:39:50 | openstackgerrit | Olivier Bourdon proposed openstack-infra/shade master: Add extra failure codes to bad request exception https://review.openstack.org/550107 | |
| 16:59:47 | openstackgerrit | Chris Dent proposed openstack/microversion-parse master: Provide a helper method to get headers from environ https://review.openstack.org/495356 | |
| 16:59:48 | openstackgerrit | Chris Dent proposed openstack/microversion-parse master: Import functions from placement https://review.openstack.org/496212 | |
| 17:28:07 | openstackgerrit | Dean Troyer proposed openstack/python-openstackclient master: Re-implement more removed novaclient bits https://review.openstack.org/550176 | |
| 17:28:07 | openstackgerrit | Dean Troyer proposed openstack/python-openstackclient master: Convert add floating IP command to nova-net/neutron split https://review.openstack.org/550105 | |
| 20:49:09 | openstackgerrit | Chris Dent proposed openstack/microversion-parse master: Add MicroversionMiddleware https://review.openstack.org/550252 | |
| 20:49:09 | openstackgerrit | Chris Dent proposed openstack/microversion-parse master: Improve Version.matches() method https://review.openstack.org/550251 | |
| 21:22:40 | openstackgerrit | Chris Dent proposed openstack/microversion-parse master: Allow passing a json_error_formatter to the middleware https://review.openstack.org/550263 | |
| 21:22:41 | openstackgerrit | Chris Dent proposed openstack/microversion-parse master: Clean up package version and python version support https://review.openstack.org/550264 | |
| #openstack-sdks - 2018-03-07 | |||
| 02:44:43 | Qiming | mordred, previously, we did that in SDK using the Profile, client can do profile.set_api_version to specify a microversion for a service individually ... when the request is later sent to to keystoneauth1, we compile the microversion request into a single header by following the api-wg guideline so the server side can know better which api version is requested | |
| 02:46:47 | Qiming | mordred, these code looked to me have been changed quite a lot ... :) I don't recognize them now ... would love to hear your opinion on the way we support API microversioning. | |
| 02:49:27 | Qiming | mordred, imho, we have several things to do wrt API microversioning: 1) re-enable sdk to forward API microversion request to server; 2) better parsing of the API microversion returned (assuming the service is playing a good citizen); 3) negotiation of API microversion to send based on (explicitly) specified version request and the range returned by the server | |
| 13:11:04 | openstackgerrit | Chris Dent proposed openstack/api-wg master: Add guidance on needing cache-control headers https://review.openstack.org/550468 | |
| 14:51:43 | olivierbourdon38 | hello everyone, any potential reviewers for some shade fixes: https://review.openstack.org/549981 https://review.openstack.org/550107 ? thx | |
| 14:57:17 | olivierbourdon38 | yes I think it should be solved now by https://review.openstack.org/550484 | |
| 14:57:24 | TheJulia | yeah | |
| 14:57:43 | olivierbourdon38 | at least from what I digged out so far | |
| 14:58:06 | TheJulia | yeah, that should be it I think | |
| 14:58:28 | TheJulia | olivierbourdon38: reviewed both patches, One I guess just needs a docstring, the other I think needs some testing | |
| 15:00:50 | olivierbourdon38 | TheJulia many thx will update accordingly | |
| 15:41:26 | openstackgerrit | Olivier Bourdon proposed openstack-infra/shade master: Add proper return value for validate_node https://review.openstack.org/549981 | |
| 16:35:34 | openstackgerrit | Olivier Bourdon proposed openstack-infra/shade master: Add proper return value for validate_node https://review.openstack.org/549981 | |
| 16:58:08 | openstackgerrit | Dean Troyer proposed openstack/python-openstackclient master: Re-implement more removed novaclient bits https://review.openstack.org/550176 | |
| 16:58:08 | openstackgerrit | Dean Troyer proposed openstack/python-openstackclient master: Convert add floating IP command to nova-net/neutron split https://review.openstack.org/550105 | |
| 18:21:59 | openstackgerrit | Olivier Bourdon proposed openstack-infra/shade master: Add proper return value for validate_node https://review.openstack.org/549981 | |
| 18:24:03 | openstackgerrit | Olivier Bourdon proposed openstack-infra/shade master: Add proper return value for validate_node https://review.openstack.org/549981 | |
| 18:31:30 | openstackgerrit | James Denton proposed openstack/python-openstacksdk master: Adds prefixlen to the request body when creating subnets https://review.openstack.org/550558 | |
| 19:09:58 | openstackgerrit | James Denton proposed openstack/python-openstacksdk master: Adds prefixlen to the request body when creating subnets https://review.openstack.org/550558 | |
| 19:27:50 | openstackgerrit | Olivier Bourdon proposed openstack-infra/shade master: Add proper return value for validate_node https://review.openstack.org/549981 | |
| 19:55:00 | olivierbourdon38 | TheJulia fixed https://review.openstack.org/#/c/549981 adding proper testing ;-) | |
| 21:11:36 | openstackgerrit | Monty Taylor proposed openstack-infra/shade master: DNM testing building some things on some nodes https://review.openstack.org/550606 | |