Earlier  
Later
New Search
Posted Nick Remark
#openstack-sdks - 2018-08-23
15:06:59 samueldmq http://paste.openstack.org/show/728691/
15:07:14 samueldmq #link https://github.com/openstack/openstacksdk/blob/master/openstack/cloud/openstackcloud.py#L7280
15:07:42 samueldmq Shrews: this is odd. I can't even do container.name because it's a dict, differently from all other objects in openstackcloud
15:07:44 samueldmq is that a bug?
15:09:34 Shrews samueldmq: likely a bug
15:09:59 samueldmq yeah, it does self._container_cache[name] = container.headers
15:10:08 samueldmq and then the container object is just headers lol
15:10:38 samueldmq Shrews: I'll open a bug against that...
15:47:05 openstackgerrit Dmitry Tantsur proposed openstack/openstacksdk master: Run bifrost integration test jobs https://review.openstack.org/586564
15:47:32 openstackgerrit Dmitry Tantsur proposed openstack/openstacksdk master: [DNM] Testing bifrost CI https://review.openstack.org/595572
15:51:59 openstackgerrit Merged openstack/openstackclient master: Add python-qinlingclient to zuul job https://review.openstack.org/562949
16:20:19 openstackgerrit Merged openstack/api-sig master: import zuul job settings from project-config https://review.openstack.org/593943
16:45:29 elmiko dtantsur edleafe i need a clarification, that change we merged is to fix the api-wg reference in the tc repo?
16:47:23 dtantsur elmiko: I approved a change to move CI jobs in-tree
16:47:42 dtantsur this is a purely technical thing, related to a reorganization of project-config due to approaching Python 3 transition
16:48:21 elmiko dtantsur: ok, so not related to the tc stuff
16:50:19 dtantsur nope
16:50:30 elmiko thanks!
16:51:48 mordred dtantsur: oh wow. that is a CRAZY thing to return from a /v1 endpoint
16:52:17 dtantsur mordred: probably :)
16:52:22 mordred but also - yeah - we're going to need to update keystoneauth - as that does not match any of the known or supported payloads
16:52:43 mordred and is especially bad because it contains the key 'version' which is one of the things that keystoneauth looks for to figure out what sort of payload it has found
16:52:58 mordred so it expects the contents of 'version' at the top level to be a dict :)
16:54:19 elmiko cdent dtantsur edleafe https://etherpad.openstack.org/p/api-sig-newsletter ready for review
16:54:24 mordred dtantsur: I must now get lunch (crazypants travel day) - I'll see if I can come up with some options in a bit
16:54:54 dtantsur mordred: cool! I'll be out already, but I'll read your pings in the morning hopefully
16:54:59 elmiko get a good belt for those crazy pants XD
16:55:49 edleafe elmiko: wow, no pedantic grammar corrections needed. Huzzah!
16:55:50 cdent elmiko: I don't know, all that stuff about working group stuff is like "we've got no news, so I'm going to relay this really booooring thing"
16:56:13 elmiko edleafe: whoa... this is a first!
16:56:27 elmiko cdent: i can drop that last para
16:56:35 edleafe But I agree with cdent - that last paragraph should be scrapped
16:56:43 elmiko ++
16:56:51 elmiko done
16:57:01 edleafe Not like anyone reads these things
16:57:05 elmiko haha
16:57:14 cdent often more than once
16:57:15 elmiko i did read the ones for the weeks i missed, that's at least one person
16:57:27 dtantsur the same ^^
16:57:36 edleafe We should probably put in a bit like "if you are reading this, ping cdent and he will send you $10"
16:57:37 elmiko ok, i'm gonna ship this. thanks again!
16:57:42 elmiko LOL
16:57:49 edleafe I meant *outside* of our group
16:57:57 elmiko also, no surprise, i will be out for next week's meeting =(
19:37:01 openstackgerrit Mathieu Gagné proposed openstack/openstacksdk master: Add the ability to extend a volume size https://review.openstack.org/595873
#openstack-sdks - 2018-08-24
01:24:18 openstackgerrit Monty Taylor proposed openstack/openstacksdk master: baremetal: support newer microversions in {get,update,patch}_machine https://review.openstack.org/594242
02:06:20 mordred dtantsur|afk: ok. SO - I dug all the way down the the rabbit hole in to keystoneauth and I don't think we can support versioned endpoints for ironic with the noauth / standalone without some updates to the version response from ironic
02:06:25 mordred we can update to fail more better
02:06:34 mordred but the version response has no microversion data
02:06:40 mordred so we have to assume v1
02:07:01 mordred that said - I think it wouldn't be awful to add the microversion info to the versioned response
02:07:18 mordred and we can update keystoneauth slightly to make it upwards compatible when ironic grows doing that
02:16:44 openstackgerrit Monty Taylor proposed openstack/keystoneauth master: Add support for ironic single-version responses https://review.openstack.org/595980
02:16:51 mordred dtantsur|afk, TheJulia: ^^
02:20:48 mordred dtantsur|afk: for now, in sdk, I think just putting in the try/except around the discovery failure seems to 'work' for endpoint things ... although given the goal is supporting newer microversions I'm not sure it fully achieves all of the goals
02:22:06 mordred dtantsur|afk: we COULD try to add a warning if someone tries to use ironic through sdk with an endpoint_override that ends in /v1 - saying that microversions canot be detected
02:22:08 mordred hrm
02:23:02 mordred the discovery path might do the wrong thing is given an unversioned endpoint in endpoint_override
14:26:38 mnaser mordred: does shade have a way of listing stack resources, i don't see one right off the bat..
14:29:13 Shrews mnaser: i don't see one either
14:29:56 mnaser Shrews: is there a way to get a client via shade or not anymore? would need to go to openstacksdk for that?
14:31:23 Shrews mnaser: there should be a orchestration_client property
14:32:18 mnaser Shrews: that's wonderful, thank you
14:54:53 openstackgerrit Samuel de Medeiros Queiroz proposed openstack/openstacksdk master: WIP: Functional tests for core functionality https://review.openstack.org/596359
14:57:00 pooja_jadhav_ hello all, I am wrirring functional_tests in openstacksdk for masakari. In that, for test_host module. for get_host_test, trying assert host objects. (expected, actual). But its failing because of timestamp in created_at field.
14:58:48 pooja_jadhav_ while creating host its giving correct datetime with timestamp. but while saving host record in DB, its saves only date nd time (not timestamp). so while getting that host, created_at field will append timestamp as 0000. its mismatching reference and actual object.
14:58:58 pooja_jadhav_ please anyone can help me in this?
15:01:51 openstackgerrit Samuel de Medeiros Queiroz proposed openstack/openstacksdk master: WIP: Functional tests for core functionality https://review.openstack.org/596359
15:03:01 cdent Shrews: you have any insight on pooja_jadhav_ 's issue?
15:04:13 Shrews pooja_jadhav_: cdent: my initial thought is to maybe compare individual fields (as test_create() does), or to mock the time function used
15:04:48 Shrews (if i understand correctly)
15:04:58 Shrews sorry, have to AFK for a bit now
15:05:35 samueldmq mordred: cdent: Shrews: would be awesome to have your eyes on https://review.openstack.org/596359
15:05:58 samueldmq I am sure it's not on shade patterns yet, but I wanted to make sure what I'm doing looks sane
15:06:18 samueldmq it's for my masters research and I'd like to run that against several clouds really soon
15:06:46 cdent samueldmq: your masters is apparently more interesting than mine was :)
15:09:07 samueldmq cdent: what was yours about?
15:09:36 samueldmq I'm investigating intra-platform interop in iaas, focusing on openstack
15:10:08 cdent generically info science, but I guess if you had to label it with something it was sort of the intersection of HCI and engelbart's ideas on augmentation. So it wasn't actually boring at all, but it didn't involve assaulting clouds :)
15:13:26 samueldmq heheh yes :) looks interesting
15:13:38 samueldmq assaulting clouds seems interesting in the openstack context
15:42:40 pooja_jadhav_ shrews: Can u help me in the how to mock time function while creating any resource like (host or segment)
15:43:31 pooja_jadhav_ shrews: in the functional tests
15:43:59 pooja_jadhav_ shrews: so that i can assert whole object for remaining fields
15:46:21 pooja_jadhav_ shrews: But i think usually we dont use mock in functional tests
17:44:58 openstackgerrit Samuel de Medeiros Queiroz proposed openstack/openstacksdk master: WIP: Functional tests for core functionality https://review.openstack.org/596359
18:51:13 openstackgerrit Julia Kreger proposed openstack/openstacksdk master: [DNM] Testing bifrost CI https://review.openstack.org/595572
19:45:43 openstackgerrit Julia Kreger proposed openstack/openstacksdk master: [DNM] Testing bifrost CI https://review.openstack.org/595572
20:07:03 TheJulia dtantsur|afk: http://logs.openstack.org/25/596425/2/check/bifrost-integration-tinyipa-ubuntu-xenial/85f6c1b/job-output.txt.gz#_2018-08-24_19_52_46_756788 <-- tl;dr openstacksdk master branch state is broken for bifrost :\ But there is a pretty backtrace.
21:09:25 samueldmq I love comments in openstackcloud
21:09:56 samueldmq looking at image create code when: "# This makes me want to die inside"
21:10:03 samueldmq and others :-)
21:16:20 TheJulia heh
#openstack-sdks - 2018-08-25
06:09:37 mordred dtantsur|afk: re: pulling microversions from headers - because that's literally not how any of the rest of the services do things, and I'd rather fix ironic if we can instead of adding YET ANOTHER completely different way of doing the exact same thing
06:10:56 mordred dtantsur|afk: (yes, we can, and probably will add support to do that because it's our job to make this work - but it's ... frustrating to say the least)
06:11:53 mordred dtantsur|afk: as for the 'remove /v1' - we can't - because the issue here is what to do if the versioned endpoint is passed in directly because of the noauth plugin
06:12:39 mordred when someone passes in a direct endpoint via endpoint_override (which is the mechanism the endpoint/noauth approach is using) - we don't do *any* modifications to the endpoint passed in - it is assumed that the endpoint the user is passing in is the endpoint they want to use
06:12:53 mordred and in fact, it is correct - it IS the enpdoint the user wants to use if they pass in the versioned endpoint
06:13:09 mordred it's just that the endpoint doesn't have a version discover document
06:13:17 mordred *sigh*

Earlier   Later