| Posted | Nick | Remark | |
|---|---|---|---|
| #openstack-nova - 2021-12-10 | |||
| 18:37:52 | opendevreview | Dmitriy Rabotyagov proposed openstack/nova stable/wallaby: Ensure MAC addresses characters are in the same case https://review.opendev.org/c/openstack/nova/+/816883 | |
| 20:06:49 | jamesbenson | Can anyone help with refstack tests failing due to additional-properties? | |
| 20:10:53 | jamesbenson | These are, for example this test: tempest.api.compute.servers.test_instance_actions.InstanceActionsTestJSON.test_get_instance_action | |
| 20:11:01 | jamesbenson | Gives me a: jsonschema.exceptions.ValidationError: Additional properties are not allowed ('updated_at' was unexpected) | |
| 20:12:38 | gmann | jamesbenson: hi | |
| 20:13:05 | gmann | jamesbenson: is it with upstream code or modified API? | |
| 20:13:48 | gmann | In tempest, we have strict validation on API response on additional-properties and t will fail if any API return any additional field in response without microversion or in downstream | |
| 20:15:09 | jamesbenson | okay, so I need to add the microversions to my tempest.conf then? Currently I don't have those. | |
| 20:15:39 | jamesbenson | It with a local deployment | |
| 20:17:42 | jamesbenson | @gmann https://gitlab.com/utsa-ics/osias We do regular testing and only ran into this issue with Victoria on. Ussuri wasn't giving any issues. But we haven't merged code for Victoria, Wallaby, or Xena yet because of these issues. | |
| 20:19:59 | gmann | jamesbenson: i see, please cap the max version for victoria in tempest conf as 2.87 max_microversion https://github.com/openstack/tempest/blob/93a902072fd9986f2bb660166552f37d9eb5bdbb/tempest/config.py#L381 | |
| 20:20:50 | jamesbenson | @gmann, Yeah, these are version locks we have: https://gitlab.com/utsa-ics/osias/-/blob/master/osias_variables.py#L67 | |
| 20:21:32 | jamesbenson | We need to specify both min and max version, correct? | |
| 20:22:13 | gmann | jamesbenson: min version can be None as Nova has not bumped the min version and its 2.1 only. min_microversion as None in tempest conf means nova 2.1 | |
| 20:22:27 | gmann | jamesbenson: so you are getting error with these cap version right? | |
| 20:23:43 | jamesbenson | @gmann, currently we don't have these caps used in our testing and are getting the errors. | |
| 20:24:20 | gmann | jamesbenson: so basically with your current configuration, tempest will request nova microversion 2.80 for victoria | |
| 20:24:42 | jamesbenson | We did use them at some point, but I think we were getting additional errors, so we stopped. But I'm running tests now, so I should know in a couple of hours. | |
| 20:24:52 | gmann | jamesbenson: make min_microversion as None and max_microversion as you have currently in https://gitlab.com/utsa-ics/osias/-/blob/master/osias_variables.py#L72 | |
| 20:24:56 | jamesbenson | Yes, these are what would get put into the tempest.conf file | |
| 20:25:17 | jamesbenson | okay, will do! | |
| 20:25:29 | gmann | jamesbenson: because tempest might fail with min_microversion as 2.80 because few tests might need modification on asserts/schema | |
| 20:26:00 | jamesbenson | gotcha, should I set them all to "None"? | |
| 20:26:07 | gmann | yeah | |
| 20:26:34 | jamesbenson | Xena I have a max of: 2.90 | |
| 20:26:59 | gmann | yeah, that's correct https://docs.openstack.org/nova/latest/reference/api-microversion-history.html#maximum-in-xena | |
| 20:27:02 | jamesbenson | Does the tempest version or other things matter? | |
| 20:27:26 | jamesbenson | (And any thoughts on the project, feel free to share, we love input) | |
| 20:27:28 | gmann | jamesbenson: for victoria no. Tempest master works for ussuri - current openstack master | |
| 20:28:07 | gmann | that is what we test at upstream also, tempest master to test ussuri, victoria, wallaby, xena and current yoga | |
| 20:28:48 | jamesbenson | okay, cool. I'll update tempest u->x to 29.1.0 | |
| 20:29:52 | gmann | +1, please let me know if tempest fail in victoria as min_microversion=None and max_microversion='2.87' | |
| 20:30:58 | jamesbenson | sure will do | |
| 20:31:02 | jamesbenson | one other thing fails too | |
| 20:31:34 | jamesbenson | we use cirros as our test image, and gocubsgo works, but then some tests use random passwords, and those fail | |
| 20:31:39 | jamesbenson | Any ideas on that? | |
| 20:31:42 | jamesbenson | microversion issue too? | |
| 20:33:13 | gmann | jamesbenson: cirros is one we use in upstream too and password is configurable in tempest config. But tempest is not ready to run for all combination of micrversion. | |
| 20:33:38 | gmann | I will say try with min_microversion=None and max_microversion='2.87' and then we can see if they still fail then its tempest issue | |
| 20:34:29 | jamesbenson | ok, sounds good. These are our setups for tempest.conf | |
| 20:34:31 | jamesbenson | https://gitlab.com/utsa-ics/osias/-/blob/master/test_setup.sh | |
| 20:36:11 | gmann | yeah, that seems correct image_ssh_password but there might be some test issue, if you face issue ping me on #openstack-qa or file bug in tempest i can fix that | |
| 20:36:37 | jamesbenson | you are a life saver! Thank you! | |
| 20:37:13 | gmann | np! | |
| 20:39:53 | jamesbenson | I do have one re-occuring bug, it always happens in ussuri up in the tearDownClass (tempest.api.compute.servers.test_multiple_create.MultipleCreateTestJSON). Details: {'type': 'SubnetInUse', 'message': 'Unable to complete operation on subnet 5be11612-2a84-4723-b683-3cf06e3567d4: One or more ports have an IP allocation from this subnet.', 'detail': ''} | |
| 20:40:22 | jamesbenson | It can't destroy the subnet it creates.... | |
| 20:41:54 | gmann | jamesbenson: can you please log bug in tempest for that, I will check on Monday | |
| 20:42:08 | gmann | here https://bugs.launchpad.net/tempest | |
| 20:42:15 | jamesbenson | sure thing | |
| 20:42:33 | gmann | thanks | |
| 20:44:14 | jamesbenson | Any special logs you need or just the output from Refstack on that? | |
| 20:44:49 | gmann | you can add tempest log also | |
| 20:50:23 | jamesbenson | https://bugs.launchpad.net/tempest/+bug/1954535 | |
| 23:05:31 | jamesbenson | @gmann, 'None' had issues so I changed it to 2.1 for the min. I'm rerunning it now, I'll report back on Monday. Thanks again and have a great weekend! | |
| 23:07:11 | gmann | jamesbenson: sure, you too. | |
| #openstack-nova - 2021-12-11 | |||
| 21:11:22 | opendevreview | Stanislav Dmitriev proposed openstack/nova master: Fix PCI passthrough race on reschedule (claims) https://review.opendev.org/c/openstack/nova/+/710847 | |
| 21:36:09 | opendevreview | Stanislav Dmitriev proposed openstack/nova master: Fix PCI passthrough race on reschedule (claims) https://review.opendev.org/c/openstack/nova/+/710847 | |
| 22:00:16 | opendevreview | Stanislav Dmitriev proposed openstack/nova master: Fix PCI passthrough race on reschedule (refresh) https://review.opendev.org/c/openstack/nova/+/710848 | |
| #openstack-nova - 2021-12-12 | |||
| 03:13:39 | opendevreview | Stanislav Dmitriev proposed openstack/nova master: Fix PCI passthrough race on reschedule (claims) https://review.opendev.org/c/openstack/nova/+/710847 | |
| 03:40:56 | opendevreview | Stanislav Dmitriev proposed openstack/nova master: Fix PCI passthrough race on reschedule (refresh) https://review.opendev.org/c/openstack/nova/+/710848 | |
| 14:28:33 | opendevreview | Stanislav Dmitriev proposed openstack/nova master: Fix PCI passthrough race on reschedule (claims) https://review.opendev.org/c/openstack/nova/+/710847 | |
| 16:08:14 | opendevreview | Stanislav Dmitriev proposed openstack/nova master: Fix PCI passthrough race on reschedule (refresh) https://review.opendev.org/c/openstack/nova/+/710848 | |
| 16:47:06 | opendevreview | Stanislav Dmitriev proposed openstack/nova master: Functional test test_boot_reschedule_with_proper_pci_device_count https://review.opendev.org/c/openstack/nova/+/760354 | |
| 16:48:58 | opendevreview | Stanislav Dmitriev proposed openstack/nova master: Functional test test_boot_reschedule_with_proper_pci_device_count https://review.opendev.org/c/openstack/nova/+/760354 | |
| #openstack-nova - 2021-12-13 | |||
| 01:56:52 | sdmitriev | https://review.opendev.org/c/openstack/nova/+/760354/5 | |
| 01:56:54 | sdmitriev | https://review.opendev.org/c/openstack/nova/+/710848/6 | |
| 01:57:06 | sdmitriev | https://review.opendev.org/c/openstack/nova/+/710847/8 | |
| 01:57:23 | sdmitriev | Hello there, just rebased these commits above, would really appreciate if someone could review them | |
| 01:57:33 | sdmitriev | I'm really interested in fixing that bug since deployments are pretty much affected | |
| 08:47:49 | opendevreview | Pierre-Samuel Le Stang proposed openstack/nova-specs master: Implements: blueprint soft-delete-instance-actions https://review.opendev.org/c/openstack/nova-specs/+/821387 | |
| 09:42:23 | gibi | bauzas, sean-k-mooney: this is my last week before vacation and I will not be back until M2 so I'd like to get over with the placement spec https://review.opendev.org/q/topic:any-traits-support this week. | |
| 09:42:42 | bauzas | gibi: ack good point | |
| 09:42:46 | gibi | it was already approved a long time ago so I don't think it is controversioal | |
| 09:42:52 | bauzas | and reminder that we have our spec review day tomorrow | |
| 09:43:03 | gibi | yepp, tomorrow is good for me | |
| 09:43:04 | gibi | :) | |
| 09:50:36 | gibi | as a trade I'm willing to look at many reviews this week as I don't want to start any new on my side just closing down open things | |
| 09:50:41 | gibi | so hit me with reviews | |
| 11:03:45 | gibi | sean-k-mooney: I'm +2 on the recent version of the instance action softdelete spec https://review.opendev.org/c/openstack/nova-specs/+/821387 | |
| 11:11:25 | gibi | bauzas: and easy stable/xena backport needing a second core (lyarwood is already on PTO) https://review.opendev.org/q/topic:%2522bug/1945310%2522+branch:stable/xena | |
| 11:17:48 | bauzas | gibi: done | |
| 11:18:14 | bauzas | gibi: I'm just reviewing pslestang's spec already :) | |
| 11:50:07 | sean-k-mooney | gibi: just reading back | |
| 11:50:33 | sean-k-mooney | ill look at the placement one shortly | |
| 11:51:27 | sean-k-mooney | gibi: i was pretty happy with the instance action spec so ill look at that again this morning since its been revised | |
| 11:52:57 | sean-k-mooney | gibi: by the way i dong thave +2 rights on placemt/placemsnt spec directory since its in tree but ill review it anyway | |
| 11:56:12 | sean-k-mooney | gibi: oh this is the spect that like yuou do required=in:TRAIT1,TRAIT2 | |
| 11:57:55 | opendevreview | Wenping Song proposed openstack/nova master: Fill the AcceleratorRequestBindingFailed exception msg info https://review.opendev.org/c/openstack/nova/+/817326 | |
| 12:07:30 | gibi | sean-k-mooney: yes, that is in :TRAIT1,TRAIT2 | |
| 12:07:58 | sean-k-mooney | gibi: i have one open qustion/suggestion for that spec | |
| 12:09:52 | sean-k-mooney | gibi: https://review.opendev.org/c/openstack/placement/+/649992/5/doc/source/specs/yoga/approved/2005346-any-traits-in-allocation_candidates-query.rst#68 | |
| 12:10:59 | sean-k-mooney | oh i see the last spec is for mixing which is what i was asking us to support | |
| 12:11:50 | sean-k-mooney | the proposal for mixing will work for placment but now for nova/glance | |
| 12:12:47 | sean-k-mooney | i tink just using ; is simpler | |
| 12:24:56 | gibi | sean-k-mooney: it is almost directly following how member_of is modelled in the placement API | |
| 12:25:18 | gibi | so repeat means AND. OR expressed with in: prefix | |
| 12:26:02 | gibi | sure ';' could be used instead of repeat of the required query param but then we would create inconsistency with member_of | |
| 12:32:00 | opendevreview | Pierre-Samuel Le Stang proposed openstack/nova-specs master: Implements: blueprint soft-delete-instance-actions https://review.opendev.org/c/openstack/nova-specs/+/821387 | |
| 12:33:07 | sean-k-mooney | gibi: it would yes but the ofllow on proposal of using multiple &required=...&required=... | |
| 12:33:17 | sean-k-mooney | cant be modled in flavors or images | |