Earlier  
Later
New Search
Posted Nick Remark
#openstack-nova - 2018-01-25
19:31:46 TheJulia mriedem: thanks for the recheck
19:32:16 TheJulia mriedem: yes, waiting for the client changes we need to merge before cutting the release :(
19:33:27 rloo mriedem: the traits work will land for sure; just one client patch in gate. the rescue work has been approved but we need 2 ironic patches to land before client can land. and gate is super not-fast today as you probably know :-(
19:34:19 mriedem there hasn't been any review on the nova patches for rescue, so that's going to be deferred
19:35:15 mriedem also https://review.openstack.org/#/c/508116/ - that doesn't bump the minimum required ironic api version to 1.37 before trying to patch the node with traits
19:35:19 mriedem but that's required isn't it?
19:35:22 mriedem jroll: ^
19:35:44 mriedem we'll likely need someone to work the nova patches at this point since i'm guessing mgoddard and johnthetubaguy are eating boiled meat at this point
19:36:59 rloo mriedem: boiled meat. yummmy. which is better, doing that or reviewing? i'll take a look at the nova patches in a few min.
19:37:08 jroll I can help too
19:37:35 mriedem well i'm assuming the mv needs to change to 1.37 since that's the ironic api version that added traits support
19:37:41 jroll it doesn't
19:37:44 mriedem i could be wrong, there are a lot of +1s on there
19:37:44 jroll explaining in gerrit now
19:37:46 mriedem ok
19:38:38 jroll comment is there mriedem
19:38:46 jroll and this is the one that needs the api version bump https://review.openstack.org/#/c/532288/9
19:39:04 jroll since it's grabbing the traits from the node
19:41:19 izombie Trying to list servers and I'm hitting 404 and I can't understand why
19:41:24 izombie Any help?
19:42:31 TheJulia on a plus side jroll, that is the correct version :)
19:44:14 jroll yep!
19:45:17 mriedem izombie: check the nova-api logs
19:45:37 mriedem jroll: ah ok, thanks
19:46:44 izombie :) I did. Even debugged! I'm lost the route is not matched. It reaches till route middleware and then throws HTTPNotFound
19:48:58 artom izombie, technically I'm supposed to send you to #openstack (see /topic)
19:49:30 melwitt mriedem: any particular reviews I can help with today?
19:49:30 artom izombie, but, what request are you making?
19:49:56 izombie Simple nova list. And it breaks here https://github.com/openstack/nova/blob/master/nova/wsgi.py#L458
19:50:18 izombie artom: thanks for not sending :P
19:50:32 artom izombie, well, I did, in a hint-y sort of way ;)
19:50:39 artom izombie, can you run the client in --debug?
19:50:53 artom And what are you running it against? devstack?
19:51:06 izombie I will try my luck there, meanwhile I thought to try my luck here as I'm out of it now
19:51:19 mriedem melwitt: i'm just going through this ironic driver traits series
19:51:30 mriedem https://review.openstack.org/#/c/508116/
19:58:28 izombie guess, I won't fix it today.
19:58:50 artom I mean, I did ask them --debug output :/
19:59:44 openstackgerrit Artom Lifshitz proposed openstack/nova master: Refactor _build_device_metadata https://review.openstack.org/533804
19:59:45 openstackgerrit Artom Lifshitz proposed openstack/nova master: Consider hostdev devices when building metadata https://review.openstack.org/533805
20:00:31 efried mriedem The ironic traits thing, we're trying to get it in?
20:00:59 efried mriedem I can probably do any last-minute changes if you didn't get another volunteer.
20:01:17 mriedem we'll see
20:01:25 mriedem i'm reviewing the changes now yes,
20:01:25 jroll I'm happy to help too
20:01:36 mriedem which requires looking at some of the provider tree stuff that already merged, like set_traits_for_provider
20:01:53 efried mriedem Yes, that was a prereq. Let me know if you want the exec summary on that.
20:03:03 mriedem i don't suppose there are any integration tests anywhere in ironic that run something that plumbs this all through
20:03:31 mriedem like, create a flavor with a required trait, decorate a node with it, pass the trait from the node to the resource provider in placement, etc
20:03:54 jroll nope
20:04:18 jroll not terribly hard to do other than the time crunch
20:05:06 mriedem you gonna tattoo a todo for yourself?
20:05:37 jroll I'll think about it :P
20:05:41 mriedem we could also have a simple functional test in-tree that has a fake virt driver that returns a fake trait, and we make sure that the RT and such creates it in placement for us
20:05:50 jroll I can add it to the existing job pretty easy
20:05:51 mriedem rather than just these lowly unit test
20:12:41 mriedem jroll: that test_get_traits_with_no_traits_attr test can also go away once we require 1.37 and depend on ironicclient to give us a node object that has a traits attribute
20:13:11 mriedem jroll: you want to update https://review.openstack.org/#/c/532288/ then?
20:13:19 mriedem i'll be +2 after removing the hasattr thing
20:13:50 jroll mriedem: heh, we wrote a test for that? fun
20:14:00 jroll I can do that but it'll fail hard until we release a new client
20:14:37 mriedem welp,
20:14:40 mriedem alternative is,
20:15:06 mriedem we push a change on top that removes that, and has a dependency on whatever g-r bump we're going to have for the new ironicclient release version
20:15:39 jroll oh, I'm fine with the former, just letting you know it won't be landable
20:15:49 jroll (see also my -1 for the hasattr thing)
20:15:52 mriedem i'm ok with it not being landable until the deps are done
20:15:55 mriedem i did
20:16:02 mriedem i'll FFE this
20:16:06 jroll I'll hack out some sort of functional or integration test too
20:16:41 mriedem alright, you might want to save any in-tree functional test patch for something that goes on top of the series
20:16:52 jroll sure
20:24:26 openstackgerrit Claudiu Belu proposed openstack/nova master: tests: refactors and cleans up test_rbd.py https://review.openstack.org/538022
20:24:59 openstackgerrit Jim Rollenhagen proposed openstack/nova master: Implement get_traits() for the ironic virt driver https://review.openstack.org/532288
20:25:01 jroll mriedem: efried: there you are
20:26:26 mriedem jroll: likely somewhere in nova.tests.functional.test_servers
20:26:38 mriedem there are lots of examples of tests using a specific fake virt driver,
20:26:42 jroll mriedem: I'm liking nova/tests/functional/compute/test_resource_tracker.py
20:26:43 mriedem and using a placement api client to assert things
20:26:49 jroll has a bunch of RT + ironic things already
20:27:04 mriedem this doesn't necessarily need to be an ironic virt driver functional test
20:27:19 mriedem just a fake driver (see nova/virt/fake.py) that returns a standard and custom trait
20:27:30 jroll true
20:27:38 mriedem test_resource_tracker.py is probably fine as a new test class
20:27:57 mriedem *with a new test class
20:28:03 jroll yeah, I'll poke around, thanks
20:43:29 openstackgerrit Claudiu Belu proposed openstack/nova master: tests: fixes mock autospec usage https://review.openstack.org/447505
20:43:36 mriedem fyi, the cinder backup bug is still in the gate
20:43:39 mriedem there is another thing going on
20:43:52 mriedem swiftclient is logging within the thread
20:43:52 dansmith still in the gate as in, unmerged?
20:43:55 mriedem no,
20:43:59 mriedem another issue, same bug
20:44:06 mriedem http://logs.openstack.org/05/266605/33/check/neutron-tempest-dvr/8a5dd5f/logs/screen-c-bak.txt.gz#_Jan_25_19_26_01_896832
20:44:16 dansmith ah
20:44:17 mriedem swiftclient is logging while their are in a thread and it switches
20:44:23 mriedem *they are
20:44:54 efried shitballs, does this mean stuff that's been sitting in the gate for 18h is now going to fail??
20:44:57 mriedem seeing if i can patch cinder-backup to not do debug logging for swiftclient if using that backup driver
20:45:06 mriedem efried: it means it has a good chance of failing

Earlier   Later