Earlier  
Later
New Search
Posted Nick Remark
#openstack-nova - 2017-01-03
19:02:45 jaypipes _gryf: and doing that with a microversion to the placement REST API of course.
19:03:12 _gryf jaypipes, does that requires new blueprint?
19:04:18 jaypipes _gryf: yup... anything that touches the REST API needs a blueprint.
19:05:01 _gryf I can cook it up really quickly if that makes sense for ocata
19:13:01 jaypipes _gryf: sure thing, please do!
19:13:30 _gryf ok :)
19:13:39 mriedem jaypipes: on https://review.openstack.org/#/c/414214/
19:13:55 mriedem http://logs.openstack.org/14/414214/10/check/gate-tempest-dsvm-ironic-ipa-wholedisk-agent_ipmitool-tinyipa-ubuntu-xenial-nv/4188484/logs/screen-n-cpu.txt.gz?level=TRACE doesn't show the conflicts because the placement-api isn't enabled in the ironic CI job
19:14:14 mriedem that's what https://review.openstack.org/#/c/415523/ is for
19:15:18 jaypipes mriedem: ah sorry my mistake.
19:15:25 mriedem np
19:15:31 jaypipes mriedem: looking through the placement api logs on that patch now.
19:19:01 jaypipes mriedem: I see the problem...
19:19:52 mriedem whew
19:41:04 openstackgerrit Dan Smith proposed openstack/nova: Make servers api view load instance fault from proper cell https://review.openstack.org/407238
19:41:04 openstackgerrit Dan Smith proposed openstack/nova: Multicell support for instance listing https://review.openstack.org/396775
19:41:05 openstackgerrit Dan Smith proposed openstack/nova: Fix non-parameterized service id in hypervisors sample tests https://review.openstack.org/410471
19:41:05 openstackgerrit Dan Smith proposed openstack/nova: Make the SingleCellSimple fixture a little more comprehensive https://review.openstack.org/410472
19:41:06 openstackgerrit Dan Smith proposed openstack/nova: WIP: Make api_samples tests use simple cell environment https://review.openstack.org/410473
19:41:07 openstackgerrit Dan Smith proposed openstack/nova: WIP Move instance creation to conductor https://review.openstack.org/319379
19:46:28 jaypipes mriedem: the default detail parameter value is True...
19:47:17 openstackgerrit Jay Pipes proposed openstack/nova: placement: Do not save 0-valued inventory https://review.openstack.org/414214
19:47:31 jaypipes mriedem: awrighy, let's see if ^ does the trick.
19:47:31 mriedem jaypipes: yeah i know,
19:47:46 mriedem jaypipes: maybe all of the tests are passing detail=False :)
19:47:51 mriedem otherwise it's a crazy mystery
19:48:18 jaypipes mriedem: I looked through the _wait_for_server_status() stuff and get_servers() never passed detail=False :(
19:48:40 jaypipes mriedem: but I agree it's a crazy mystery thus my original review comment to dansmith :)
19:51:39 mriedem jaypipes: well, are we sure that fill_faults() isn't called, or just that there are no faults for any of the servers being tested?
19:52:15 mriedem as noted in the patch i had pulled it down and raised some exceptions from that code (in _get_fault() in the view class) and the functional tests hit failures in there
19:52:24 mriedem most of this code is tested in the functional tests, not the unit tests
19:57:22 openstackgerrit Julian proposed openstack/nova: Create script that prints JSON Schema of Notification Payload Classes https://review.openstack.org/411836
20:23:46 openstackgerrit Sujitha proposed openstack/nova: Do not merge: Add post hook script to run tempest with old n-api https://review.openstack.org/401395
20:24:48 openstackgerrit STEW TY proposed openstack/nova: Transform instance.unrescue notifications (WIP) https://review.openstack.org/388275
20:26:52 openstackgerrit OpenStack Proposal Bot proposed openstack/nova: Updated from global requirements https://review.openstack.org/415978
20:28:51 openstackgerrit Matt Riedemann proposed openstack/nova: Make test_attach_interfaces work with use_neutron=True by default https://review.openstack.org/416001
20:28:52 openstackgerrit Matt Riedemann proposed openstack/nova: Make test_nova_manage pass with CONF.use_neutron=True by default https://review.openstack.org/416003
20:28:52 openstackgerrit Matt Riedemann proposed openstack/nova: Make test_user_data and test_multiple_create work with use_neutron=True https://review.openstack.org/415729
20:28:53 openstackgerrit Matt Riedemann proposed openstack/nova: Make test_floating_ip* pass with CONF.use_neutron=True by default https://review.openstack.org/415999
20:28:54 openstackgerrit Matt Riedemann proposed openstack/nova: Stub out os_vif.unplug in libvirt instance destroy tests https://review.openstack.org/416002
20:28:54 openstackgerrit Matt Riedemann proposed openstack/nova: Make test_metadata pass with CONF.use_neutron=True by default https://review.openstack.org/416004
20:28:54 openstackgerrit Matt Riedemann proposed openstack/nova: Make test_security_group_default_rules work with use_neutron=True by default https://review.openstack.org/415767
20:28:55 openstackgerrit Matt Riedemann proposed openstack/nova: Make last remaining unit tests work with Neutron by default https://review.openstack.org/416006
20:28:55 openstackgerrit Matt Riedemann proposed openstack/nova: Make test_virtual_interfaces work with CONF.use_neutron=True by default https://review.openstack.org/415732
20:28:56 openstackgerrit Matt Riedemann proposed openstack/nova: Make test_server_usage work with CONF.use_neutron=True by default https://review.openstack.org/415770
20:28:56 openstackgerrit Matt Riedemann proposed openstack/nova: Make several API unit tests pass with CONF.use_neutron=True by default https://review.openstack.org/415771
20:28:57 openstackgerrit Matt Riedemann proposed openstack/nova: Make test_quota work with CONF.use_neutron=True by default https://review.openstack.org/415707
20:28:57 openstackgerrit Matt Riedemann proposed openstack/nova: Make test_security_groups work with CONF.use_neutron=True by default https://review.openstack.org/415747
20:28:58 openstackgerrit Matt Riedemann proposed openstack/nova: Enable Neutron by default https://review.openstack.org/416083
20:28:58 openstackgerrit Matt Riedemann proposed openstack/nova: Make test_tenant_networks pass with CONF.use_neutron=True by default https://review.openstack.org/415755
20:28:59 openstackgerrit Matt Riedemann proposed openstack/nova: Make test_compute pass with CONF.use_neutron=True by default https://review.openstack.org/399978
20:30:04 mriedem dansmith: sdague: are you guys OK with this? https://review.openstack.org/#/c/415511/
20:30:21 mriedem if so i can remove the -2 from the code change and start working on the tempest API tests
20:31:18 sdague mriedem: I can't really look in detail until tomorrow
20:31:23 sdague the top line seems fine
20:36:02 dansmith mriedem: yeah.. haven't read the actual spec, but seems like something we'll need
20:36:27 mriedem the spec is pretty basic,
20:36:33 mriedem add a microversion to return aggregate uuid
20:36:45 mriedem more or less just hte paperwork, the actual api code is done by jay
20:36:47 mriedem and reviewed by me
20:37:23 dansmith yep
20:37:45 mriedem it's what i like to call synergy
20:44:21 openstackgerrit Matt Riedemann proposed openstack/nova: Enable Neutron by default https://review.openstack.org/416083
20:52:57 _gryf jaypipes, given that all of the REST API are described in generic-resource-pools bp, shouldn't it be amended instead of creating another BP?
20:53:25 jaypipes _gryf: depends on what mriedem thinks :)
20:53:45 _gryf oh…
20:53:51 _gryf :)
20:57:08 efried slunkad, thanks for your patience. I've reviewed https://review.openstack.org/#/c/408148/
21:05:56 mriedem _gryf: to do what?
21:06:03 _gryf mriedem, so what is the preferred approach regarding updating bp? if BP was already implemented in previous cycle, should w amend it in place, or move it to the ocata directory?
21:06:10 mriedem the generic resource pools spec isn't a carte blanche
21:06:38 mriedem _gryf: we already have generic-resource-pools-ocata,
21:06:47 mriedem but that was mostly for tying up loose ends with aggregates
21:07:21 cdent mriedem, _gryf: there's this in progress too, with regard to updating that spec: https://review.openstack.org/#/c/407562/
21:08:26 _gryf cdent, oh, so maybe I'll do the patch on top of yours
21:08:32 mriedem cdent: ok but those are clarifications on what actually got implemented, right? not new things
21:08:44 cdent mriedem: correct, just mentioning it for context
21:08:45 mriedem _gryf: amending a spec from newton to add new things isn't appropriate
21:09:15 mriedem _gryf: for example, we have a separate spec for returning uuids from os-aggregates which is something we'll need for generic-resource-pools, but wasn't actually called out in the original spec,
21:09:20 mriedem and was a separate API change so we created a separate spec
21:09:32 _gryf right
21:09:47 mriedem i believe jay also created a separate bp for getting notifications out of the placement API on resource provider CRUD ops
21:09:58 _gryf ok, so I'll do the separate bp than
21:10:36 mriedem https://blueprints.launchpad.net/nova/+spec/placement-notifications
21:10:48 mriedem _gryf: it would be helpful if you told me what you're looking to do
21:10:51 mriedem code-wise
21:11:27 _gryf mriedem, here is jaypipes patch: https://review.openstack.org/#/c/404990/5
21:11:43 mriedem you want to return specific codes from the placement API yes?
21:11:50 mriedem in a header or something?
21:11:53 _gryf yeah, that's the brief idea
21:12:04 _gryf I was thinking on json response rather
21:12:47 mriedem either way i think it's out of scope for ocata, and it's an api change, so probably requires a spec queue up for pike
21:12:55 mriedem *queued
21:13:26 mriedem a spec would be good for describing how that's going to look to the consumer, because i'm sure the api-wg will have input on this
21:14:07 cdent _gryf: codes meaning what?
21:14:24 _gryf so, I can cook it up and place on the pike directory
21:14:40 _gryf cdent, for distinguish between different 409 cases
21:15:09 cdent ah, I see. yeah the errors guideline from the api-wg has ideas on that
21:15:12 mriedem e.g. resource provider generation conflict vs inventory in use conflict
21:15:30 _gryf yeah

Earlier   Later