| Posted | Nick | Remark | |
|---|---|---|---|
| #openstack-sdks - 2017-09-04 | |||
| 07:14:29 | ldl | okay, I will try | |
| 07:14:34 | ldl | thank you | |
| 07:15:54 | Dinesh_Bhor | ldl: Np | |
| 07:16:52 | ldl | do you mean use pip to install them on Mac? | |
| 07:17:58 | Dinesh_Bhor | ldl: not sure about Mac. Sorry | |
| 07:18:35 | ldl | and does the virtualenv helps? | |
| 07:19:31 | ldl | and windows ok? | |
| #openstack-sdks - 2017-09-05 | |||
| 02:55:23 | openstackgerrit | Reedip proposed openstack/python-openstacksdk master: Add support for dns-domain in Openstack SDK https://review.openstack.org/500660 | |
| 02:56:52 | reedip | stevemar: ^^ | |
| 03:00:54 | openstackgerrit | Reedip proposed openstack/python-openstacksdk master: Add support for dns-domain in Openstack SDK https://review.openstack.org/500660 | |
| 07:10:07 | openstackgerrit | Reedip proposed openstack/python-openstackclient master: Introduce quota unset command https://review.openstack.org/376311 | |
| 08:21:49 | openstackgerrit | Tytus Kurek proposed openstack/python-openstackclient master: Add support for "--dns-domain" argument https://review.openstack.org/500450 | |
| 09:47:44 | openstackgerrit | Tytus Kurek proposed openstack/python-openstackclient master: Add support for "--dns-domain" argument https://review.openstack.org/500450 | |
| 12:54:31 | openstackgerrit | Brian Curtin proposed openstack/python-openstacksdk master: Add support for dns-domain https://review.openstack.org/500660 | |
| 13:29:53 | openstackgerrit | Merged openstack/service-types-authority master: Generate standard api_reference https://review.openstack.org/495887 | |
| 14:19:17 | efried | mordred What's the story with the automatic sta publisher? | |
| 15:00:30 | efried | mordred Seems to have kicked in now. | |
| 15:16:00 | mordred | efried: hey guess what? | |
| 15:16:25 | efried | mordred Tell me | |
| 15:16:28 | mordred | efried: this morning I have discovered that ironic version discovery returns something different and exciting on its versioned endpoint | |
| 15:16:59 | mordred | efried: please enjoy this: https://developer.openstack.org/api-ref/baremetal/#show-v1-api | |
| 15:17:22 | efried | ... | |
| 15:19:04 | efried | mordred I don't think I have an ironic service anywhere. Paste me? | |
| 15:19:22 | efried | oh | |
| 15:19:25 | mordred | efried: yah | |
| 15:19:43 | efried | well holy incompatibility, Batman. | |
| 15:20:01 | mordred | efried: I mean, the issue will never get hit in normal usage because people put the unversioned endpoint in the catalog | |
| 15:20:14 | efried | we hope | |
| 15:20:17 | mordred | and the unversioned document is compatible, so that's what ksa discovery will read and use | |
| 15:20:33 | mordred | efried: they have to - ironicclient hard-code appends a /v1 to what it gets from the catalog :) | |
| 15:20:55 | mordred | so any deployment with ironic not having unversioned endpoint in the catalog would not work with ironicclient | |
| 15:21:06 | efried | So.... what action needs to be taken here? And by whom? | |
| 15:21:41 | mordred | HOWEVER - we have code in shade to simulate what ironicclient does by doing get_endpoint() on our baremetal adapter, appending a '/v1' and then setting endpoint_override | |
| 15:22:00 | mordred | by doing that, if we then called get_endpoint_data things would not work | |
| 15:22:11 | mordred | (that's easy to fix in shade and I've got that fix coming) | |
| 15:22:38 | efried | mordred Does this have impact on https://review.openstack.org/#/c/488137/13/nova/virt/ironic/client_wrapper.py ? | |
| 15:22:51 | mordred | which is all to say - I think, when we get around to it -it's not urgent - we should add support for that version of a versioned discovery document - just so that users don't do a thing and then get confused | |
| 15:23:22 | mordred | efried: no - I don't think it does | |
| 15:23:29 | efried | okay. | |
| 15:23:52 | mordred | efried: we might want to document that if people want to pass an ironic.endpoint_override in their config that it should be the unversioned, not the versioned endpoint | |
| 15:23:56 | mordred | I think that's what they'd do anyway | |
| 15:24:18 | mordred | but while other services would work with either, ksa will in fact blow up in the ironic case if a versioned endpoint is given | |
| 15:25:04 | efried | mordred So like here https://review.openstack.org/#/c/488137/13/nova/conf/ironic.py@43 the deprecation warning could mention that, but once we get rid of api_endpoint, how would we document it just for ironic? | |
| 15:26:25 | efried | I suppose we could update the document for the imported ksa opts - because in any case I imagine you would recommend using an unversioned endpoint_override and specifying the version via the `version` conf opt. | |
| 15:27:34 | openstackgerrit | Merged openstack/keystoneauth master: Make discover._version_between more consistent https://review.openstack.org/485827 | |
| 15:30:56 | mordred | efried: yes - I mean, that's what I'm actually going to do in shade now - just do the darned discovery | |
| 15:31:37 | mordred | efried: however, there is only one major version of ironic, so some folks might desire to put in an override in their config of the versioned endpoint to cause the discovery step to be skipped | |
| 15:31:54 | mordred | and if they did that, they'd get a sad, even though doing that would work for every other service in ksa | |
| 15:32:08 | mordred | efried: oh - wait - let me actually respond to whatyou said :) | |
| 15:32:34 | mordred | efried: YES - I think for the imported ksa opts we should totally update those docs and recommend that practice 100% | |
| 15:33:41 | openstackgerrit | Sean Dague proposed openstack/api-wg master: WIP: microversion architecture archival doc https://review.openstack.org/444892 | |
| 15:35:34 | openstackgerrit | Eric Fried proposed openstack/keystoneauth master: Recommend unversioned for endpoint-override https://review.openstack.org/500892 | |
| 15:35:44 | efried | mordred Something like this? ^^ | |
| 15:38:04 | mordred | efried: yes | |
| 15:38:49 | mordred | sdague: aww. you didn't call "the Multi-Cloud Integrator" Monty. I guess I'm fine being called Sophia | |
| 15:40:15 | sdague | mordred: yeh, well the naming predates | |
| 15:40:44 | sdague | I think I ended up just choosing off of the top of most popular baby names for the year I was writing it | |
| 15:40:53 | sdague | flipping back and forth on genders | |
| 15:49:44 | mordred | sdague: :) | |
| 15:49:56 | mordred | sdague: well - I just left you some inline comments | |
| 15:53:17 | sdague | mordred: cool | |
| 15:53:59 | sdague | the biggest chunk I really need to collect from people is best practices on the implementation side. I can do the services bit, but the API / SDK ones need help from people that have been in those trenches | |
| 16:13:08 | mordred | sdague: yup. I'm happy to help with that | |
| 16:28:16 | mordred | efried: also - one more thing - the new NoAuth plugin in the new ksa release - doesn't work with version discovery | |
| 16:28:27 | mordred | efried: actually - why don't I move that to the keystone channel | |
| 16:28:55 | rabel | is server create --block-device-mapping ignoring the device name with nova libvirt? i'm just trying it out on devstack and it seems to always map the xtra volume to /dev/vdb | |
| 16:30:37 | sdague | rabel: yes, device name can never be guarunteed in the guest with libvirt | |
| 16:30:52 | rabel | sdague: thx | |
| 16:31:49 | sdague | rabel: https://developer.openstack.org/api-ref/compute/#create-server - see block_device_mapping_v2.device_name | |
| 16:50:45 | openstackgerrit | Merged openstack/keystoneauth master: Nits in using-sessions.rst https://review.openstack.org/500211 | |
| 18:46:06 | openstackgerrit | Monty Taylor proposed openstack/keystoneauth master: Move version discovery support to BaseAuthPlugin https://review.openstack.org/500956 | |
| 18:50:37 | openstackgerrit | Monty Taylor proposed openstack/keystoneauth master: Add version discovery support to BaseAuthPlugin https://review.openstack.org/500956 | |
| 20:11:26 | openstackgerrit | Merged openstack/keystoneauth master: Add tests for mutually exclusive [min|max]version https://review.openstack.org/500198 | |
| 21:49:04 | cdent | elmiko, edleafe: Can we cook a schedule out of this? https://etherpad.openstack.org/p/api-ptg-queens Do we need to? elmiko have there been any nibbles on the guided reviews? | |
| 22:12:56 | elmiko | cdent: only 2 nibbles so far that i'm aware of. dtantsur|afk (ironic) and tellesnobrega (sahara) | |
| 22:13:13 | elmiko | nothing more than a nibble though | |
| 22:13:14 | cdent | so effectively some friends of the family | |
| 22:13:20 | elmiko | yeah | |
| 22:13:36 | elmiko | i do have a suggestion to help fill our time though, and i'm happy to add it to the schedule | |
| 22:15:22 | elmiko | added | |
| 22:16:27 | elmiko | cdent: also, in answer to your previous question about meeting last week. we just hung out for like 10-15 minutes, and didn't send a newsletter | |
| 22:16:30 | cdent | elmiko: cool. that also reminded me to add something (hackathon) | |
| 22:16:40 | elmiko | nice +1 | |
| 22:18:11 | elmiko | gotta jet, bbl o/ | |
| 22:20:22 | cdent | o/ | |
| 23:04:05 | openstackgerrit | Monty Taylor proposed openstack/keystoneauth master: Add version discovery support to BaseAuthPlugin https://review.openstack.org/500956 | |
| 23:15:06 | edleafe | cdent: Yeah, a schedule would be desireable. Maybe add a time estimate to each of the topics? | |
| 23:18:27 | cdent | edleafe: many of those topics are just placeholders. as in no one has expressed a desire to talk about them, they are just things “we could” talk about. thus my struggle | |
| 23:20:26 | edleafe | Well, let's schedule them. If no one shows up, we can bring a deck of cards and play rummy for the allotted time | |
| 23:25:12 | cdent | we need a 4th for euchre | |
| #openstack-sdks - 2017-09-06 | |||
| 02:49:59 | openstackgerrit | yanghuichan proposed openstack/os-api-ref master: Fix wrong links https://review.openstack.org/501079 | |
| 03:01:26 | openstackgerrit | ShangXiao proposed openstack/os-api-ref master: Update and replace http with https for doc links https://review.openstack.org/501087 | |
| 03:15:37 | openstackgerrit | OpenStack Proposal Bot proposed openstack/python-openstackclient master: Updated from global requirements https://review.openstack.org/500282 | |
| 03:27:54 | openstackgerrit | Rui Chen proposed openstack/python-openstackclient feature/osc4: [Compute]Make column content readable for both human and machine https://review.openstack.org/493814 | |
| 03:41:51 | openstackgerrit | dangogh proposed openstack/cliff master: add actual column names to error msg Closes-Bug: 1712876 https://review.openstack.org/498277 | |
| 03:41:52 | openstack | bug 1712876 in cliff "openstack cli - show available columns" [Undecided,In progress] https://launchpad.net/bugs/1712876 - Assigned to Dan Kirkwood (dangogh) | |
| 13:08:31 | edleafe | cdent: elmiko: I've updated the etherpad with comments. https://etherpad.openstack.org/p/api-ptg-queens | |
| 13:15:33 | elmiko | edleafe: ack, /me looks | |
| 13:16:55 | elmiko | edleafe: +1, nice additions | |
| 13:17:34 | edleafe | elmiko: I'd like to be able to estimate how long we should allocate for each topic | |