Earlier  
Later
New Search
Posted Nick Remark
#openstack-sdks - 2019-08-02
08:16:59 kmalloc keep in mind the blueprint/resource class (we use flask restful) doesn't exist until it is called, so sometimes self.XXX may not work since you're in a class context.
08:28:36 dtantsur okay, cool!
08:28:52 dtantsur now my biggest problem is how to gradually replace wsme without doing everything at once
08:54:09 dtantsur mordred: hi! opinion needed on https://github.com/dtantsur/rust-osauth/pull/7
10:08:20 cdent death to wsme!
10:22:20 dtantsur \o/
11:46:51 mordred dtantsur: responded
11:50:46 dtantsur thx
11:56:10 mordred dtantsur: also - it makes me sad that there are deployments out that which don't expose those urls properly. it's largely because python-novaclient and friends never used them (hardcoding url manipulations instead) - so nobody noticed
12:01:33 dtantsur yep, pretty sad
12:01:55 dtantsur see, my small and young project is already getting ugly hacks because of nova!
12:01:57 dtantsur :D
12:49:26 openstackgerrit Dmitry Tantsur proposed openstack/openstacksdk master: baremetal.configdrive: tolerate user_data as a string https://review.opendev.org/674298
13:39:42 frickler mordred: seems you commented as opendevadmin instead your personal account, I'm assuming that wasn't intended?
14:44:30 openstackgerrit Bence Romsics proposed openstack/openstacksdk master: Add router add/remove route operations https://review.opendev.org/674324
14:47:36 openstackgerrit Bence Romsics proposed openstack/python-openstackclient master: Add command: router add/remove route --route https://review.opendev.org/674325
15:05:31 gtema mordred: Houston - we have a problem. Ansible is not capable in resolving resource objects. Latest merged released change for Keypairs results in Ansible not properly getting keypair attributes
15:05:41 openstackgerrit Riccardo Pittau proposed openstack/openstacksdk master: Sort output of functional test trait https://review.opendev.org/674328
15:17:17 openstackgerrit Matt Riedemann proposed openstack/python-openstackclient master: Add openstack server create --boot-from-volume option https://review.opendev.org/674111
15:28:39 openstackgerrit Riccardo Pittau proposed openstack/openstacksdk master: Sort output of functional test trait https://review.opendev.org/674328
15:43:52 openstackgerrit Artem Goncharov proposed openstack/openstacksdk master: Enable ansible module test for keypair to check return data https://review.opendev.org/674333
16:00:18 elmiko edleafe, dtantsur, following on from thierry's pr, i put up a new patch to remove the early morning hours
16:00:27 dtantsur thx!
16:00:30 elmiko i tagged you all on it
16:01:06 elmiko i know we didn't have a formal vote about it or anything, but i guess the review can serve as that?
16:01:57 dtantsur yep
16:02:16 dtantsur actually, my vote is enough ;) if I stop attending, the slot won't be attended any more
16:02:24 edleafe No!! I want a formal vote, in a grand hall with trumpets and such!
16:04:21 dtantsur accepted!
16:12:01 openstackgerrit Luigi Toscano proposed openstack/openstacksdk master: DNM Test for zuul-jobs https://review.opendev.org/674344
16:18:17 elmiko haha, i think Thierry already workflowed it. dtantsur's +1 was enough XD
16:43:40 mordred frickler: whoops! thanks - I forgot to switch logins after setting up the opendevadmin account yesterday
16:44:39 mordred gtema: that's not awesome. when you say "is not capable in resolving resource objects" - that's because ansible is expecting to be able to return them in the json and that's not working? I thought we'd made resoures subclasses of dict so that this would work :(
16:45:22 gtema yes, I am deeply in debuging Ansible jsonifying Resource
16:45:45 gtema so basically call exit_json(key=SDK_RESOURCE) is returning nulls
16:46:13 gtema it get the keys, but somehow not the values
16:46:40 gtema http://paste.openstack.org/show/755341/
16:47:23 gtema and with the 0.32 including keypair "fix" it becomes broken
16:50:41 gtema in Ansible json.dumps likely will get exception, since it uses "encoding" and switch to "container_to_text" method
16:50:50 gtema somewhere here something is missing
16:53:49 mordred that seems really off - and like there's something missing down deep somewhere
16:54:12 gtema exact
16:54:33 gtema and it's terrible to debug it, since no propel logging is possible
17:35:58 gtema that's some mistery. When I am copying all those ansible functions into unit tests to be able to reproduce - it works, under ansible it fails
17:38:32 gtema the iteritems (iter(d.items(**kw)) return empty values from ansible, but correct from sdk unit tests
18:16:53 openstackgerrit Matt Riedemann proposed openstack/openstacksdk master: Fix DeprecationWarning for using logger.warn https://review.opendev.org/674401
18:16:58 mriedem easy fix for something that's now blowing up the nova functional test job logs ^
18:17:03 mriedem https://logs.opendev.org/61/673961/1/gate/nova-tox-functional-py36/a4cb2af/job-output.txt.gz#_2019-08-01_17_51_24_070219
20:02:06 efried mriedem: That UnsupportedServiceVersion thing - is that showing up in devstack tests or only functional?
20:02:41 mriedem let me ask the magic 8 ball
20:02:44 mriedem i mean logstash
20:04:20 mriedem both
20:04:27 mriedem http://logstash.openstack.org/#dashboard/file/logstash.json?query=message%3A%5C%22UnsupportedServiceVersion%3A%20Service%20placement%20has%20no%20discoverable%20version.%5C%22&from=7d
20:05:29 efried mordred: Can we make placement a known, supportes service type at v1.x with an empty Proxy?
20:09:11 mordred efried: absolutely. let me make you a patch
20:09:22 efried thanks!
20:09:42 mordred efried: does placement need to do version discovery on its root endpoint and/or microversions?
20:09:44 efried mordred: that'll get rid of those warnings, yah?
20:09:48 mordred yeah
20:10:16 efried mordred: There's only one major version rn.
20:10:29 efried By microversion discovery, you mean "I need an endpoint that can handle at least 1.XX"?
20:11:01 mordred well - I mean getting the min/max microversion range from the endpoint - so yeah. I guess the question is - does placement do microversions
20:11:18 efried Placement definitely "does" microversions, yeah.
20:11:22 mordred awesome.
20:11:45 mordred is there a placement.example.com/v1 ?
20:11:53 mordred you know - how about I go read the docs and then ask you follow up questions
20:12:15 efried mordred: And especially if you discover holes in the docs that don't answer these questions, let us know (or open stories or whatever)
20:12:50 efried mordred: cdent has been in charge of this kind of stuff, so I'm pretty sure we're following all the rules wrt discovery documents and endpoint construction and whatnot.
20:13:03 mordred cool
20:13:19 efried like, I'm 95% sure you won't find snowflake behavior
20:15:18 efried efried@canuc:~$ curl http://192.168.218.27/placement
20:15:18 efried {"versions": [{"status": "CURRENT", "min_version": "1.0", "max_version": "1.36", "id": "v1.0", "links": [{"href": "", "rel": "self"}]}]}
20:15:31 efried that href="" is suspicious
20:22:11 efried mordred: Poking at my devstack, it is looking like we do not have versioned endpoints for placement.
20:22:26 efried So forget what I said about cdent, and let's burn him at the stake.
21:54:25 mordred efried: that's fine - I've already got sdk provisions for this sort of thing :)
#openstack-sdks - 2019-08-03
07:09:00 openstackgerrit Artem Goncharov proposed openstack/openstacksdk master: Enable ansible module test for keypair to check return data https://review.opendev.org/674333
18:59:53 openstackgerrit Artem Goncharov proposed openstack/openstacksdk master: Enable ansible module test for keypair to check return data https://review.opendev.org/674333
19:18:40 openstackgerrit Merged openstack/openstacksdk master: Add CloudRegion helper method for arbitrary endpoints https://review.opendev.org/674130
19:20:05 openstackgerrit Merged openstack/openstacksdk master: Fix README example for cloud layer https://review.opendev.org/674041
19:20:06 openstackgerrit Merged openstack/openstacksdk master: Move the history lesson into the docs https://review.opendev.org/674045
20:22:38 openstackgerrit Merged openstack/os-client-config master: Add Python 3 Train unit tests https://review.opendev.org/669450
#openstack-sdks - 2019-08-04
12:56:44 openstackgerrit norman shen proposed openstack/python-openstackclient master: support create domain project https://review.opendev.org/674487
#openstack-sdks - 2019-08-05
07:57:13 openstackgerrit Merged openstack/openstacksdk master: Fix DeprecationWarning for using logger.warn https://review.opendev.org/674401
08:58:08 openstackgerrit Bence Romsics proposed openstack/python-openstackclient master: Add command: router add/remove route --route https://review.opendev.org/674325
10:18:34 openstackgerrit Vishakha Agarwal proposed openstack/openstacksdk master: Add application credential CRUD support https://review.opendev.org/669331
12:10:25 gtema dtantsur, mordred, Shrews: https://review.opendev.org/#/c/674333/ is really important (event as a hotfix), since without it Ansible can't get/create keypairs
12:10:44 gtema dtantsur, can you please have a look, whether ironic failure in the tests is related or not
12:11:20 gtema this change also includes switching Ansible tests to Py3, since the problem is visible only there and more and more Platforms are delivering py3 only ansible
13:21:40 mordred dtantsur|afk, Shrews: yeah - I recommend looking at that patch from gtema - and when it lands let's cut a bugfix release with it
13:22:01 mordred being broken for ansible on python3 is a bad thing
13:22:17 mordred should we add a second run of the ansible tests to cover python2?
13:26:59 efried mordred: should I nack https://review.opendev.org/#/c/674515/ on the theory that that placement placeholder will be in place soon?
14:09:08 gtema mordred: yeah, adding ansible py2 would be likely good
14:15:26 mordred efried: yeah - I'll get you one today
14:22:17 openstackgerrit Monty Taylor proposed openstack/openstacksdk master: Add empty placement proxy https://review.opendev.org/674620
14:22:18 mordred efried: there's the first stab at it
14:23:36 openstackgerrit Monty Taylor proposed openstack/openstacksdk master: Add empty placement proxy https://review.opendev.org/674620
14:31:45 efried \o/ Thanks mordred
15:38:12 Shrews dtantsur|afk: FYI, looks like the asserts in openstack.tests.functional.baremetal.test_baremetal_node compare lists (node.traits) in unsorted order and randomly fail (see https://logs.opendev.org/33/674333/3/check/openstacksdk-functional-devstack-ironic/0e8be8d/testr_results.html.gz)

Earlier   Later