Earlier  
Later
New Search
Posted Nick Remark
#openstack-nova - 2017-06-14
20:21:34 openstackgerrit melanie witt proposed openstack/nova master: Add online migration to move quotas to API database https://review.openstack.org/410946
20:21:37 openstackgerrit melanie witt proposed openstack/nova master: Count floating ips to check quota https://review.openstack.org/457862
20:24:51 mriedem oomichi: do you remember why we need to apply the validation.query_schema on all api versioned methods even if some microversions don't have query parameters?
20:25:06 mriedem oomichi: for example, GET /os-keypairs/{id} on 2.1 has https://github.com/openstack/nova/blob/master/nova/api/openstack/compute/keypairs.py#L205
20:25:13 mriedem which is an empty schema
20:25:29 mriedem but then 2.10 has query parameter schema for passing a user_id https://github.com/openstack/nova/blob/master/nova/api/openstack/compute/keypairs.py#L191
20:25:48 mriedem so i understand why we have that for the 2.10 method, but not for 2.1 or 2.2
20:39:19 mriedem oy, i suppose a compute_nodes.hypervisor_hostname field can be an IP?
20:40:13 mriedem or a uuid in the case of ironic
20:40:39 mriedem that makes GET /os-hypervisors/{hypervisor_hostname_pattern}/search fun if you're using ironic
20:41:54 mriedem and this looks wrong https://developer.openstack.org/api-ref/compute/?expanded=search-hypervisor-detail#search-hypervisor
20:41:59 mriedem "The hypervisor host name provided by the Nova virt driver. For the Ironic driver, it is the Ironic node name."
20:42:24 mriedem https://github.com/openstack/nova/blob/master/nova/virt/ironic/driver.py#L332
20:56:46 cfriesen does nova-conductor need to be running in order to run "nova db online_data_migrations"?
20:56:53 melwitt cells meeting in 4 minutes
21:18:30 eandersson Anyone that got some time to help me look into this :D
21:18:30 eandersson https://review.openstack.org/#/c/390984
21:19:33 eandersson I am trying to figure out why the Server Group Tests are failing because the resource usage isn't always updated.
21:19:56 eandersson https://github.com/openstack/nova/blob/master/nova/compute/resource_tracker.py#L168
21:20:27 eandersson Need to figure out if it is the test that is broken, or if there is a race condition happening here.
21:20:45 eandersson jaypipes maybe?
21:30:44 mriedem sdague: any idea why we'd cache compute nodes in the request here https://github.com/openstack/nova/blob/master/nova/api/openstack/compute/hypervisors.py#L110 ?
21:30:53 mriedem nothing is looking up a cached compute node record from what i can tell
21:30:55 mriedem anywhere else in the api
21:31:01 mriedem do we just do it for fun and games?
21:38:13 oomichi mriedem: I guessed that was a reason why unknown queries should be blocked, but now additionalProperties is True on https://github.com/openstack/nova/blob/master/nova/api/openstack/compute/schemas/keypairs.py#L94
21:38:30 mriedem yeah i looked up the original change that added that and it wasn't clear to me
21:38:45 oomichi mriedem: so I can not find the reason now. alex_xu can answer anyways
21:38:55 mriedem yeah i have a lot of questions for alex_xu
21:39:12 mriedem 5:40am for him, i'm not sure why he's not up and working yet :)
21:39:13 oomichi oh, "a lot"..
21:39:50 mriedem see john asked about the same sort of thing here https://review.openstack.org/#/c/398755/18/nova/api/openstack/compute/schemas/keypairs.py
21:40:03 mriedem btw, i removed those decorators from the methods that didn't need them and no tests failed
21:40:08 mriedem so i'm not really sure why we need them
21:40:11 mriedem but that's a question for alex
21:40:16 mriedem one of many
21:40:58 oomichi hehe, I saw the patch on the history. and I found I +2d on that before without this question
21:41:14 mriedem oomichi: did you see my question to sean above about https://github.com/openstack/nova/blob/master/nova/api/openstack/compute/hypervisors.py#L110 ?
21:41:24 mriedem as far as i can tell, nothing uses those cached compute nodes in the request
21:41:32 mriedem i think the cache is more or less for extensions isn't it?
21:43:39 melwitt that's my understanding, that those request caches are just to avoid doing DB lookup every time for each extension that receives the request
21:44:30 melwitt since multiple extensions can operate on the same request
21:44:59 mriedem i don't see anything extending the "hypervisors" resource extension
21:45:09 mriedem therefore i don't think anything would be using the cache
21:45:18 mriedem maybe something used to and we collapsed it?
21:45:26 mriedem unless the view does somehow
21:45:37 melwitt they don't have to extend it to be run, I don't think
21:45:52 mriedem https://review.openstack.org/#/c/53785/
21:46:02 mriedem aha
21:46:04 mriedem os-pci
21:46:05 mriedem mfer
21:46:10 melwitt can't remember for sure though. I just know a lot of it is like onion layers, running each extension on the request object
21:46:20 mriedem time to get out my delete knife
22:04:50 melwitt eandersson: it looks like there might be something wrong with the test. in the log trace there's "Compute_service record created for compute:fake-mini" but then later on there's "Received an update from an unknown host 'compute'. Re-created its InstanceList." which looks like it's not matching it to the earlier created compute
22:06:05 melwitt like I wonder if the first compute service in the test needs to be started more similarly as the second compute service with the host and nodes set for it
22:08:20 eandersson Yea - found that as well
22:08:37 jaypipes eandersson: agree with melwitt. I don't really see anything in that patch that should be producing that error. and it seems the failure doesn't always occur, thus the rechecks.
22:09:08 jaypipes eandersson: the whole disabled check should definitely not be coming into play here.
22:09:15 eandersson Testing it locally it occurs maybe 80% of the time though
22:09:21 jaypipes hmm
22:09:31 jaypipes and only those affinity tests?
22:09:41 jaypipes eandersson: ^
22:10:10 eandersson I am just doing a single test at the moment (test_boot_servers_with_affinity_no_valid_host)
22:10:18 eandersson Adding a sleep at the right place and it succeeds 100%
22:11:48 melwitt eandersson: you might try patching like this and see if it improves anything locally http://paste.openstack.org/show/612599/
22:12:02 openstackgerrit Matt Riedemann proposed openstack/nova master: Handle uuid in HostAPI.compute_node_get https://review.openstack.org/461532
22:12:03 openstackgerrit Matt Riedemann proposed openstack/nova master: WIP: Handle uuids in os-hypervisors API https://review.openstack.org/471899
22:12:04 openstackgerrit Matt Riedemann proposed openstack/nova master: Use uuid for id in os-services API https://review.openstack.org/464280
22:12:19 mriedem melwitt: dansmith: oomichi: check out https://review.openstack.org/471899 if you want to see some terrible things
22:17:01 cfriesen mriedem: does nova-conductor need to be running in order to run "nova db online_data_migrations"?
22:17:47 oomichi mriedem: I am not sure why we want to add empty schema again for this api v2.0, is that WIP reason?
22:18:08 oomichi oh, these schemas are not used on https://review.openstack.org/#/c/471899/8/nova/api/openstack/compute/hypervisors.py ..
22:18:12 eandersson melwitt, was looking into something similar, still something holding it up
22:18:37 eandersson *but
22:18:59 openstackgerrit Takashi NATSUME proposed openstack/python-novaclient master: Microversion 2.48: Standardization of VM diagnostics https://review.openstack.org/473228
22:19:08 eandersson got 100% on that one test now, lets see if the others are going to be as happy :D
22:19:34 mriedem oomichi: it's one of my questions to alex_xu
22:19:40 mriedem i don't want to use empty schema if it can be avoided
22:19:45 mriedem but yeah it's all WIP right now
22:19:59 mriedem cfriesen: that's a dansmith question
22:20:04 mriedem i know some of them go through the objects,
22:20:13 mriedem but i figure they should be local to the db
22:30:33 eandersson https://github.com/openstack/nova/blob/master/nova/tests/functional/regressions/test_bug_1671648.py#L97
22:32:49 eandersson Looks like a very similar issue
22:35:43 melwitt huh, I added that. barely remember it
22:37:02 openstackgerrit Matt Riedemann proposed openstack/nova master: Handle uuid in HostAPI.compute_node_get https://review.openstack.org/461532
22:37:03 openstackgerrit Matt Riedemann proposed openstack/nova master: WIP: Handle uuids in os-hypervisors API https://review.openstack.org/471899
22:37:03 openstackgerrit Matt Riedemann proposed openstack/nova master: Use uuid for id in os-services API https://review.openstack.org/464280
22:37:04 openstackgerrit Matt Riedemann proposed openstack/nova master: Centralize compute_node_search_by_hypervisor in os-hypervisors https://review.openstack.org/474381
22:37:04 openstackgerrit Matt Riedemann proposed openstack/nova master: Stop caching compute nodes in the request https://review.openstack.org/474382
22:37:11 mriedem easy pickins on those bottom two changes
22:40:14 eandersson mriedem, so your suggestion fixed most issues, but it's still hitting this in 4 tests > not self.driver.node_is_available(nodename)
22:40:25 openstackgerrit Merged openstack/nova master: Count server group members to check quota https://review.openstack.org/457859
22:42:18 eandersson I see - FAKE_NODES need to have both hosts
22:48:28 melwitt mriedem: just curious, how was the compute node request cache messing up what you're working on?
22:49:16 openstackgerrit Merged openstack/nova master: Make compute_node_statistics() work across cells https://review.openstack.org/473931
22:53:23 openstackgerrit Erik Olof Gunnar Andersson proposed openstack/nova master: Re-worked ram filter and weights to support overcommit https://review.openstack.org/390984
22:53:40 eandersson That was effort, but I think I got it figured out
22:54:29 eandersson I could probably have left fake_network where it was though
22:59:22 eandersson thanks melwitt

Earlier   Later