| Posted | Nick | Remark | |
|---|---|---|---|
| #openstack-nova - 2023-09-18 | |||
| 12:16:42 | sean-k-mooney | it passes with ml2/ovs | |
| 12:16:46 | sean-k-mooney | so this is not a nova issue | |
| 12:17:03 | sean-k-mooney | its a neutorn one its also not an os-vif issues as far as i am aware | |
| 12:17:11 | sean-k-mooney | i have not looked that this in a while however | |
| 12:17:52 | sean-k-mooney | so i still stand by the invaldifor nova and incomplte/invalid for os-vif tatus | |
| 12:18:09 | sean-k-mooney | this is a neutron bug only when using ml2/ovn | |
| 12:18:50 | sean-k-mooney | we can skip that test if we want by default and leave it on in the hybrid plug job which uses ml2/ovs | |
| 12:19:02 | sean-k-mooney | until they fix it | |
| 12:20:55 | sean-k-mooney | bauzas: fyi i explained why this is a neutron bug in https://bugs.launchpad.net/neutron/+bug/1940425/comments/23 | |
| 12:21:29 | bauzas | yup, we know this is a neutron test and ralonsoh works on it AFAIK | |
| 12:21:55 | bauzas | context : https://review.opendev.org/c/openstack/nova/+/894945 is getting trampled a lot due to this | |
| 12:22:42 | sean-k-mooney | ya so in https://review.opendev.org/c/openstack/nova/+/893540 | |
| 12:23:02 | sean-k-mooney | i mentioned to dan that we could skip it in the other jobs where its failing and nova-ovs-hybrid-plug | |
| 12:23:06 | sean-k-mooney | would still give us coverage | |
| 12:23:11 | sean-k-mooney | want me to do that now? | |
| 12:23:38 | sean-k-mooney | we should unskip it in nova when they fix the bug but there is no reason to keep testing somethign that is broken in the interim | |
| 12:33:49 | bauzas | let's wait for dan to be here, but meh | |
| 12:34:00 | bauzas | both the changes I want to merge are in the gate by nox | |
| 12:39:16 | sean-k-mooney | ill put up a patch and ye can review | |
| 12:46:06 | bauzas | ++ | |
| 12:50:03 | opendevreview | sean mooney proposed openstack/nova master: disable ovn based testing of test_live_migration_with_trunk https://review.opendev.org/c/openstack/nova/+/895655 | |
| 12:50:18 | sean-k-mooney | i think ^ is correct but lets see what the ci says | |
| 12:53:23 | bauzas | ack | |
| 13:41:24 | opendevreview | sean mooney proposed openstack/nova-specs master: nova-specs shoudl use upper-constraints https://review.opendev.org/c/openstack/nova-specs/+/895683 | |
| 13:41:32 | opendevreview | Merged openstack/nova master: Revert "Add nova-manage ironic-compute-node-move" https://review.opendev.org/c/openstack/nova/+/894945 | |
| 13:46:25 | sean-k-mooney | bauzas: ok so nova-specs was not using UC so it got broken by the new release. since then there has also been a new release of stestr that corects the probelm | |
| 13:46:52 | sean-k-mooney | so i abandoned the uc clamp patch and made nova-spec actully use upper-constraits | |
| 13:47:31 | sean-k-mooney | it will also resolve itself once stestr 4.1.0 makes it way into the pip caches of the relvent ci providers so we should be unblokced there | |
| 13:47:33 | bauzas | huzzah, one last merge todo :) | |
| 13:51:25 | dansmith | bauzas: yeah we need to get that last one landed | |
| 13:56:31 | bauzas | yup | |
| 13:56:40 | bauzas | I'll also click on the nova-specs | |
| 14:37:49 | sahid | o/ | |
| 14:38:22 | sahid | qucik question regarding event vif-plug | |
| 14:38:29 | sahid | how nova neutron are communicating? | |
| 14:38:47 | sahid | we have lot of timout vif plug creation and are not able to find why | |
| 14:38:55 | dansmith | neutron calls to nova's api to post an event | |
| 14:39:13 | dansmith | it's been that way since icehouse, so nothing new | |
| 14:42:06 | sahid | ah yes perfect | |
| 15:41:50 | opendevreview | Sylvain Bauza proposed openstack/nova master: Add a Bobcat prelude section https://review.opendev.org/c/openstack/nova/+/894940 | |
| 15:42:11 | bauzas | dansmith: sean-k-mooney: gmann: we need some approval of the prelude above before RC1 ^ | |
| 19:07:57 | opendevreview | Merged openstack/nova master: Revert "Limit nodes by ironic shard key" https://review.opendev.org/c/openstack/nova/+/894946 | |
| #openstack-nova - 2023-09-19 | |||
| 03:50:31 | opendevreview | Nobuhiro MIKI proposed openstack/nova master: libvirt: Add config option to support maxphysaddr https://review.opendev.org/c/openstack/nova/+/895759 | |
| 03:52:06 | opendevreview | Nobuhiro MIKI proposed openstack/nova master: scheduler: Request COMPUTE_ADDRESS_SPACE_* traits based on flavor https://review.opendev.org/c/openstack/nova/+/895761 | |
| 04:43:02 | opendevreview | Takashi Kajinami proposed openstack/nova master: Fix missing oslo.versionedobjects library option https://review.opendev.org/c/openstack/nova/+/890718 | |
| 08:34:08 | opendevreview | Dr. Jens Harbott proposed openstack/nova master: Fix URLs in status check results https://review.opendev.org/c/openstack/nova/+/895770 | |
| 08:34:41 | frickler | ^^ does this prove that nobody reads upgrade check results? ;) | |
| 09:02:08 | bauzas | frickler: operators tend to run it automatically by tools :) | |
| 09:02:14 | bauzas | and just look at the status code | |
| 09:11:46 | frickler | bauzas: yes, I also noticed that kolla-ansible only fails and doesn't even show the result. need to fix that, too | |
| 09:16:11 | opendevreview | Jorhson Deng proposed openstack/nova master: Use the VersionedObjectSerializer as parent class in NovaObjectRegistry https://review.opendev.org/c/openstack/nova/+/894892 | |
| 09:32:54 | opendevreview | Sylvain Bauza proposed openstack/nova master: Add a Bobcat prelude section https://review.opendev.org/c/openstack/nova/+/894940 | |
| 09:36:08 | bauzas | dansmith: melwitt: gmann: this is way early for you, but could you guys swing quickly https://review.opendev.org/c/openstack/nova/+/894940 so we could have a late-but-hopefully-soon RC1 ? | |
| 11:47:14 | opendevreview | Jorhson Deng proposed openstack/nova master: Use the VersionedObjectSerializer as parent class in NovaObjectRegistry https://review.opendev.org/c/openstack/nova/+/894892 | |
| 11:54:57 | opendevreview | Jorhson Deng proposed openstack/nova master: Use the VersionedObjectSerializer as parent class in NovaObjectRegistry https://review.opendev.org/c/openstack/nova/+/894892 | |
| 12:39:19 | sean-k-mooney | bauzas: dansmith this is the patch i submitted to disabel the neutron trunk migration test | |
| 12:39:23 | sean-k-mooney | https://review.opendev.org/c/openstack/nova/+/895655 | |
| 12:39:30 | sean-k-mooney | looks like it pased ci | |
| 12:45:58 | Uggla | sean-k-mooney, bauzas, I wrote what I have found regarding the issue with service token. If you have a look --> https://paste.opendev.org/show/blcYEIcNRW24u8lVOqIH/ | |
| 12:46:32 | sean-k-mooney | can you point me at the ptach wehre you are trying to use them | |
| 12:46:38 | bauzas | sean-k-mooney: okidoki | |
| 12:46:50 | bauzas | sean-k-mooney: I was wondering when your PTO days were starting | |
| 12:47:01 | sean-k-mooney | monday to firday next week | |
| 12:47:09 | bauzas | if you are still on, can you review https://review.opendev.org/c/openstack/nova/+/894940 ? | |
| 12:47:17 | bauzas | prelude ^ | |
| 12:47:23 | sean-k-mooney | so i +2'd that this morning | |
| 12:47:23 | bauzas | we need it for the RC1 | |
| 12:47:42 | bauzas | oh sorry, I haven't seen on the emails | |
| 12:47:48 | bauzas | for Gerrit | |
| 12:47:52 | sean-k-mooney | no worries | |
| 12:59:37 | Uggla | sean-k-mooney, yep give me a sec | |
| 13:04:03 | bauzas | Uggla: lemme look too | |
| 13:07:11 | Uggla | sean-k-mooney, bauzas changes as simple as possible here: https://github.com/uggla/nova/blob/5341a726baaa54289838655d6b231a4716872556/nova/utils.py#L913 | |
| 13:07:37 | Uggla | L913 --> 916 which is basically passing an auth with service token. | |
| 13:08:33 | bauzas | I do remember this | |
| 13:08:55 | bauzas | Uggla: interesting thoughts you made | |
| 13:09:04 | Uggla | I think I can adapt the method in plugin.py, based on the one in identity to get the proper endpoint. But I don't know if that's a correct approach | |
| 13:09:19 | bauzas | can we somehow force the endpoint discovery with the wrapper ? | |
| 13:10:39 | Uggla | bauzas, I think instead of returning None in the plugin.py method, maybe I can get add the code to get the endpoint. | |
| 13:11:26 | bauzas | in keystoneauth ? | |
| 13:11:39 | Uggla | I mean here: https://github.com/openstack/keystoneauth/blob/e6f3999c6f2f846e3dda505343166ab8c8346c2a/keystoneauth1/plugin.py#L136 | |
| 13:11:43 | Uggla | yep | |
| 13:11:51 | bauzas | I was just wondering whether endpoint discovery was a public API argument | |
| 13:13:05 | sean-k-mooney | Uggla i was asking for your gerrit review | |
| 13:13:17 | sean-k-mooney | so i coudl see what you were changing | |
| 13:13:38 | sean-k-mooney | but also im not conviced you shoudl be changing it there | |
| 13:14:58 | Uggla | sean-k-mooney, so far I have not pushed to gerrit because it was not working. But I gave you the changes in my github. | |
| 13:16:14 | sean-k-mooney | so im not really expecting you to modify nova/utils.py | |
| 13:17:36 | bauzas | sean-k-mooney: afaik, Uggla did this on a testing purpose | |
| 13:17:59 | bauzas | because he was stuck with this error | |
| 13:18:42 | bauzas | so the code is damn simple and shouldn't probably be automatically creating a service token for *all* calls but for the moment, Uggla just wonders *why* keystoneauth gets an exception | |
| 13:18:54 | bauzas | and if you see the problem, it's due to the endpoint | |
| 13:22:48 | sean-k-mooney | so we currently have service token support in https://github.com/openstack/nova/blob/master/nova/service_auth.py | |
| 13:24:50 | sean-k-mooney | so we shoudl be reusing that as a start | |
| 13:26:31 | sean-k-mooney | and im expectign this to be used form the manilaclient function | |
| 13:29:45 | dansmith | bauzas: done | |
| 13:29:52 | bauzas | dansmith: <3 | |
| 13:31:58 | sean-k-mooney | i am going to redpeloy my devstack vm and Uggla do you have a local.conf for manilla | |
| 13:33:59 | bauzas | dansmith: you may be interested in unblocking our placement gate https://review.opendev.org/c/openstack/nova-specs/+/895683 | |
| 13:35:25 | Uggla | sean-k-mooney, local.conf --> https://paste.opendev.org/show/bfVUU744TgKbsFarvc97/ with ceph | |