| Posted | Nick | Remark | |
|---|---|---|---|
| #openstack-nova - 2020-04-10 | |||
| 20:30:08 | melwitt | I dunno yet :/ https://zuul.opendev.org/t/openstack/build/b460a8c59ad64e57b871aa3c00638b01 | |
| 20:30:40 | gmann | one is tempest issue happening for tempest py2 jobs on stable/trian and stable stein - http://lists.openstack.org/pipermail/openstack-discuss/2020-April/014010.html | |
| 20:30:48 | gmann | checking.. | |
| 20:31:14 | melwitt | yeah this isn't tempest. first time I've seen it | |
| 20:32:44 | gmann | most probably similar where stable py2 jobs end up using the master version which need >py3.6 | |
| 20:33:19 | melwitt | ah I see | |
| 20:34:41 | gmann | yeah, it is trying to clone master - https://zuul.opendev.org/t/openstack/build/b460a8c59ad64e57b871aa3c00638b01/log/job-output.txt#781 | |
| 20:35:23 | melwitt | ok. how to fix? | |
| 20:36:43 | gmann | checking why it does not use already cloned stable/train version of placement | |
| 20:46:00 | melwitt | yeah ... you'd think it's all automatic to use stable/train on ... stable/train | |
| 20:46:43 | gmann | yeah but seems this causing it https://github.com/openstack/nova/blob/28324e6aa7f1286e92c4e3292c57b9df24309cc8/.zuul.yaml#L57 | |
| 20:47:16 | melwitt | fyi the gate bug on "timed out waiting for response from cell" will still hit some failures until we merge the stable/train fix. because it can currently fail on the old side of grenade | |
| 20:48:01 | melwitt | oh hm ... | |
| 20:48:12 | gmann | ohk, is it happening all time i mean timeout ? | |
| 20:48:43 | gmann | i can try override-checkout to fix it fast and then debug why stable/train not used for required-projects | |
| 20:49:02 | melwitt | no it is only sometimes. it's just I was confused at first when I saw the logstash still was having new hits. and I realized we would have to merge fix on stable/train too in order to see all hits stop | |
| 20:49:33 | melwitt | since it is grenade running both stable/train and then upgrade, and then run master | |
| 20:50:33 | melwitt | ok, cool. well, besides that I don't know why the functional test job should need openstack/placement. oh, maybe for the openstack/placement test fixtures? | |
| 20:51:23 | gmann | ok. | |
| 20:56:16 | openstackgerrit | Ghanshyam Mann proposed openstack/nova stable/train: Use placement stable branch version for functional py2 job https://review.opendev.org/719105 | |
| 20:56:18 | gmann | melwitt: ^^ let's see if that fix | |
| 20:57:07 | melwitt | cool thank you! I will keep 👀 on it | |
| 20:57:07 | melwitt | cool thank you! I will keep 👀 on it | |
| 20:58:00 | melwitt | also noting that if this is the recommended way to do it (by -infra) then we will need to fix all of our stable branches I would think | |
| 20:58:21 | melwitt | actually I will ask them to look and confirm | |
| 20:59:56 | gmann | yeah, required-project on stable branch should use stable version | |
| 21:04:29 | artom | Have we considered just removing placement from required projects? The error is from cloning it, not using it. Why are we using placement in func jobs? | |
| 21:05:36 | artom | I guess 787bb336065 is why... | |
| 21:17:02 | melwitt | artom: there's more info in -infra, but this is some explanation https://github.com/openstack/nova/blob/stable/train/tox.ini#L76 | |
| 21:19:28 | dansmith | artom: we use it for the fixture in func tests, IIRC | |
| 21:29:26 | melwitt | yeah, tl;dr is we use it for the fixtures, and we don't want to install it for literally every test env, only func test envs, so we have to handle it in tox.ini | |
| 21:30:32 | gmann | melwitt: dansmith is it too bad to hav placement installed for unit test case - https://github.com/openstack/nova/blob/stable/train/tox.ini#L83 | |
| 21:30:54 | gmann | if no then having it in test-requirement.txt is much better solution | |
| 21:32:30 | melwitt | yeah, I have no idea really | |
| 21:34:14 | gmann | because we cannot solve stable branch and local run things without having constrain in tox.ini which need update on every placement new version | |
| 21:34:40 | melwitt | unless we do the thing you suggested earlier right, just prescribe the zuul version | |
| 21:35:03 | gmann | we can solve it for stable branch in generic way without need of update evrytime but local env run require manual handling of right version | |
| 21:35:08 | gmann | yeah | |
| 21:35:18 | melwitt | oh, so that's the snag. I see | |
| 21:36:55 | gmann | that work if we leave local run case which will be only case where we want to run functional tests for stable fix locally. | |
| 21:37:24 | gmann | which should be min cases ? | |
| 21:37:44 | melwitt | well, I think it's pretty normal to run them locally for a stable backport. I do | |
| 21:38:35 | gmann | ok | |
| 21:39:11 | melwitt | I guess so far I'm thinking,if test-requirement.txt would be the idea solution, we would have to maintain version ranges for that per stable branch anyway. and if we're doing that anyway, doing it in tox.ini isn't any worse, like clarkb said | |
| 21:40:20 | gmann | true. and having unused placement installed for unit test is not so bad | |
| 21:40:51 | gmann | i prefer that for simplicity of maintenance | |
| 21:40:53 | melwitt | well, I was thinking we do it in tox.ini at first and have a wider talk about putting into test-requirements.txt | |
| 21:41:02 | gmann | hummm | |
| 21:41:39 | gmann | in tox we need to update our-self on every branch but test-requirements.txt is automatic | |
| 21:41:40 | melwitt | I dunno, I just don't want to make that decision with just a couple of us I guess | |
| 21:42:39 | melwitt | oh, so you're saying we wouldn't have to do a version range in test-requirements? | |
| 21:44:09 | gmann | yeah, that will be handled by upper-constraint file. - https://github.com/openstack/nova/blob/stable/train/tox.ini#L18 | |
| 21:44:29 | melwitt | and clark said in tox.ini we get tox-siblings, but that doesn't do the same thing? | |
| 21:47:21 | gmann | ah you mean just have lower cap and then tox-siblings take zuul version to use | |
| 21:47:38 | gmann | let me try that. | |
| 21:47:42 | melwitt | "I" don't mean it, just trying to repeat what clarkb said | |
| 21:49:10 | melwitt | I don't understand this stuff hardly at all | |
| 21:49:47 | gmann | that should work i think. let me push that also in train gate and see | |
| 21:52:55 | gmann | melwitt: btw, keypair policy series is ready - https://review.opendev.org/#/c/718621/4 | |
| 21:53:28 | gmann | and quota class -https://review.opendev.org/#/c/719100/ | |
| 21:55:02 | melwitt | ok. if it gets too ugly, then I guess we just put into test-requirements | |
| 21:55:08 | melwitt | cool, thanks | |
| 22:04:31 | openstackgerrit | Ghanshyam Mann proposed openstack/nova stable/train: Use placement stable branch version for functional job https://review.opendev.org/719105 | |
| 22:05:03 | gmann | melwitt: let's see if that work fine. I will propose it from master then backport | |
| 22:05:19 | gmann | ^^ | |
| 22:05:42 | melwitt | ack | |
| 22:08:57 | eandersson | Before I forget. I know someone mentioned that they had a custom patch to expose numa pinning over api. | |
| 22:09:09 | eandersson | Is that something someone can share? :D | |
| 22:09:17 | eandersson | Also, what is the status of moving some of that to the placement api? | |
| 22:10:00 | melwitt | nothing in the placement api yet, nothing in progress | |
| 22:10:18 | melwitt | let me see if I can find the numa topo over api proposal | |
| 22:10:33 | eandersson | I remember someone... probably years ago mentioned that they changed the api to expose it in a custom patch :D | |
| 22:10:36 | openstackgerrit | Ghanshyam Mann proposed openstack/nova master: Add new default roles in servers policies https://review.opendev.org/662971 | |
| 22:10:38 | eandersson | thanks melwitt | |
| 22:11:02 | eandersson | We have a use case where we might need it, and thinking either a custom patch and then maybe (no promise) try to help upstream for placement | |
| 22:11:15 | eandersson | or a custom api outside of nova | |
| 22:12:04 | melwitt | ok, looks like that landed https://review.opendev.org/#/q/topic:bp/show-server-numa-topology+(status:open+OR+status:merged) in train with microversion 2.78 | |
| 22:12:41 | melwitt | https://docs.openstack.org/nova/latest/reference/api-microversion-history.html#id71 | |
| 22:12:49 | eandersson | Awesome! | |
| 22:13:03 | eandersson | We can just back-port that until we can upgrade to Train. | |
| 22:26:53 | openstackgerrit | Ghanshyam Mann proposed openstack/nova stable/train: Use placement stable branch version for functional job https://review.opendev.org/719105 | |
| 22:34:21 | openstackgerrit | melanie witt proposed openstack/nova stable/train: Reset the cell cache for database access in Service https://review.opendev.org/718934 | |
| 22:40:06 | openstackgerrit | Ghanshyam Mann proposed openstack/nova master: Add new default roles in os-evacuate policies https://review.opendev.org/708237 | |
| 22:46:19 | openstackgerrit | Ghanshyam Mann proposed openstack/nova master: Pass the target in os-services APIs policy https://review.opendev.org/676688 | |
| 22:53:29 | openstackgerrit | Merged openstack/python-novaclient master: Microversion 2.87 - Stable device boot from volume rescue https://review.opendev.org/714956 | |
| 22:55:24 | openstackgerrit | Ghanshyam Mann proposed openstack/nova master: Use placement stable version for functional job https://review.opendev.org/719115 | |
| 23:09:11 | gmann | melwitt: it worked - https://review.opendev.org/#/c/719105/ | |
| 23:12:41 | openstackgerrit | Ghanshyam Mann proposed openstack/nova master: Pass the target in os-services APIs policy https://review.opendev.org/676688 | |
| 23:13:22 | openstackgerrit | Ghanshyam Mann proposed openstack/nova master: Pass the actual target in server group policy https://review.opendev.org/717176 | |
| 23:13:42 | openstackgerrit | Ghanshyam Mann proposed openstack/nova master: Add new default roles in server group policies https://review.opendev.org/717175 | |
| 23:14:13 | openstackgerrit | Ghanshyam Mann proposed openstack/nova master: Add new default roles in server group policies https://review.opendev.org/717175 | |
| 23:14:27 | openstackgerrit | Ghanshyam Mann proposed openstack/nova master: Fix servers policy for admin_or_owner https://review.opendev.org/718501 | |
| 23:14:38 | openstackgerrit | Ghanshyam Mann proposed openstack/nova master: Add test coverage of existing server policies https://review.opendev.org/718348 | |
| 23:14:52 | openstackgerrit | Ghanshyam Mann proposed openstack/nova master: Introduce scope_types in servers Policies https://review.opendev.org/662968 | |
| 23:15:08 | openstackgerrit | Ghanshyam Mann proposed openstack/nova master: Add new default roles in servers policies https://review.opendev.org/662971 | |
| 23:25:21 | openstackgerrit | Ghanshyam Mann proposed openstack/nova master: Use placement stable version for functional job https://review.opendev.org/719115 | |
| 23:29:27 | openstackgerrit | Ghanshyam Mann proposed openstack/nova master: Use placement stable version for functional job https://review.opendev.org/719119 | |
| 23:38:16 | openstackgerrit | Ghanshyam Mann proposed openstack/nova stable/train: Use placement stable version for functional job https://review.opendev.org/719120 | |
| 23:40:26 | openstackgerrit | Ghanshyam Mann proposed openstack/nova stable/stein: Use placement stable version for functional job https://review.opendev.org/719121 | |
| 23:42:17 | gmann | melwitt: dansmith artom : pushed the functional tox fix changes from master till stein - https://review.opendev.org/#/q/I96bb9fcb55f2a4fc3aefa01db70b45740db166c8 | |