| *** masayukig_ is now known as masayukig | 01:10 | |
| *** pojadhav is now known as pojadhav|ruck | 04:57 | |
| *** jpena|off is now known as jpena | 06:53 | |
| *** bhagyashris_ is now known as bhagyashris | 07:46 | |
| yoctozepto | frickler: :-) | 08:43 |
|---|---|---|
| frickler | yoctozepto: same thing for cs8. they have 15:5.2.0-16.el8 from whatever source devstack installs there by default | 08:43 |
| yoctozepto | oh, nice, so they have upgraded too | 08:43 |
| yoctozepto | possibly why the job became flaky | 08:44 |
| yoctozepto | thanks for confirming :-) | 08:44 |
| frickler | yoctozepto: thanks for the idea, now we can tell rh folk to fix it instead of having to deal with qemu ourselves ;) | 09:15 |
| opendevreview | Dr. Jens Harbott proposed openstack/devstack master: DNM: Test debian with OVN https://review.opendev.org/c/openstack/devstack/+/814891 | 10:08 |
| *** dviroel|rover|out is now known as dviroel|rover | 10:51 | |
| *** jpena is now known as jpena|lunch | 11:29 | |
| opendevreview | Dr. Jens Harbott proposed openstack/devstack master: DNM: Test debian with OVN https://review.opendev.org/c/openstack/devstack/+/814891 | 11:36 |
| opendevreview | Lajos Katona proposed openstack/tempest master: Scenario test for routed provider networks https://review.opendev.org/c/openstack/tempest/+/665155 | 12:00 |
| yoctozepto | frickler: :-))) | 12:13 |
| *** jpena|lunch is now known as jpena | 12:22 | |
| *** pojadhav is now known as pojadhav|ruck | 12:33 | |
| *** redrobot is now known as Guest3656 | 12:59 | |
| opendevreview | James Parker proposed openstack/whitebox-tempest-plugin master: Update cpu model for deployment https://review.opendev.org/c/openstack/whitebox-tempest-plugin/+/814914 | 13:15 |
| opendevreview | Lajos Katona proposed openstack/tempest master: Scenario test for routed provider networks https://review.opendev.org/c/openstack/tempest/+/665155 | 13:58 |
| yoctozepto | frickler: wunderbar, ovn works in devstack :-) | 14:03 |
| yoctozepto | testing also in kolla | 14:04 |
| yoctozepto | (I mean Debian ovn) | 14:04 |
| yoctozepto | zigo: ^ | 14:04 |
| zigo | yoctozepto: Ah, great ! :) | 15:17 |
| opendevreview | Clark Boylan proposed openstack/devstack master: Use Nehalem CPU model by default https://review.opendev.org/c/openstack/devstack/+/815020 | 15:19 |
| yoctozepto | zigo: it works in kolla too | 15:31 |
| yoctozepto | zigo: and even better news - Debian will be an official tested platform in Yoga | 15:32 |
| zigo | :) | 15:42 |
| zigo | yoctozepto: Wonderful ! | 15:42 |
| *** jpena is now known as jpena|off | 15:52 | |
| johnsom | gmann We have an interesting situation I would like you advice on around the system_admin credential. refstack is having a job fail due to system_admin being missing: https://storage.gra.cloud.ovh.net/v1/AUTH_dcaab5e32b234d56b626f72581e3644c/zuul_opendev_logs_eaf/809450/10/check/interop-next-refstack-client-wallaby/eaf8996/job-output.txt | 16:16 |
| johnsom | I see they have a accounts override file. My first thought is that just needs to be updated, but I am also hearing that they use an older tempest version pinned. | 16:17 |
| johnsom | Should we be putting conditionals around all of our "credentials = ["admin", "system_admin", "primary"]" lines too? Or is it really that everyone with "use_dynamic_credentials = False" will need to update their files? | 16:19 |
| *** bhagyashris_ is now known as bhagyashris | 16:27 | |
| gmann | johnsom: it is not yet implemented in pre-provisioned account yet ("use_dynamic_credentials = False") . | 17:10 |
| gmann | johnsom: I think we can out a assert/log in tempest/test.py or in pre-provisioned account saying it is not supported yet | 17:11 |
| johnsom | gmann So your advice is to put conditionals around all of the "credentials =" in the test classes? | 17:11 |
| gmann | johnsom: no | 17:11 |
| gmann | johnsom: i was thinking to put a generic condition check and either 1. fail if test need system scoped and env using "use_dynamic_credentials = False" or 2. if "use_dynamic_credentials = False" fallback system admin to project admin | 17:12 |
| gmann | 2nd might not work or confusing when debugging the failure | 17:13 |
| johnsom | It's breaking refstack, so option 1 probably is not good | 17:13 |
| gmann | humm | 17:13 |
| johnsom | The tests themselves all use a conditional, it's just the base credential setup that bombs out. | 17:14 |
| gmann | yeah, we can just skip the creating the system scoped token if enforce_scope if false but for that we need a global enforce_scope_ config along with service wise | 17:15 |
| gmann | which is what we agreed to do in devstack so not bad | 17:15 |
| opendevreview | Lajos Katona proposed openstack/tempest master: Scenario test for routed provider networks https://review.opendev.org/c/openstack/tempest/+/665155 | 17:17 |
| johnsom | I'm going to propose a patch for designate to extend the service level config check to the credential creation. Feel free to comment, etc. Then if that is the right path I will update Octavia was well. | 17:18 |
| johnsom | Thanks! | 17:19 |
| gmann | johnsom: that will be needed in every tests. let me push something in tempest as generically if that works fine ? | 17:19 |
| johnsom | Ok | 17:19 |
| johnsom | Ping me and I will do a designate patch to test it. | 17:20 |
| gmann | k | 17:26 |
| gmann | johnsom: there are two issue here | 18:24 |
| gmann | johnsom: for this failure https://storage.gra.cloud.ovh.net/v1/AUTH_dcaab5e32b234d56b626f72581e3644c/zuul_opendev_logs_eaf/809450/10/check/interop-next-refstack-client-wallaby/eaf8996/job-output.txt | 18:24 |
| gmann | it seems old tempest with master designate is used in the job? | 18:24 |
| gmann | *designate-tempest-plugin | 18:24 |
| opendevreview | Ghanshyam proposed openstack/tempest master: DNM: Test if system creds request pass if enforce scope is disabled https://review.opendev.org/c/openstack/tempest/+/815048 | 18:28 |
| gmann | johnsom: 2nd problem of 'if enforce scope if false then requested system creds fail or skipped?' I am going to test ^^ 815048 | 18:29 |
| gmann | johnsom: for refstack problem you are encountering , compatible tempest and designate-tempest-plugin versions needs to be used. | 18:30 |
| johnsom | gmann That is a refstack job that I am not super familiar with, but I think they said they use a pinned tempest. Not sure. | 18:45 |
| gmann | kopecmartin: do you know if refstack or ansible-role-refstack-client use master Tempest? because it is attribute error which does not exist in old tempest (<26.1.0) and exist in new one | 18:45 |
| gmann | johnsom: yeah, I am trying to debug it in devstack logs of that job but tempest is disabled there and some other way is it is installed not via devstack | 18:45 |
| johnsom | Our main jobs in designate still run with enforce scope disabled just fine. | 18:46 |
| gmann | but only way it is failing there is if tempest <26.1.0 is used with master designate-tempest-plugin | 18:46 |
| johnsom | I wonder why refstack would have it pinned..... | 18:46 |
| gmann | johnsom: yeah, dynamic creds are good and I am going to verify the pre-provisioned creds also in 815048 if we need some condition there if enforce_scope disable | 18:47 |
| johnsom | ack | 18:48 |
| gmann | let's wait for kopecmartin on why tempest is pinned or how it is installed there. | 18:48 |
| kopecmartin | gmann: 2 year old tempest is used in refstack-cient - https://opendev.org/openinfra/refstack-client/src/branch/master/setup_env#L4 | 18:51 |
| gmann | wow | 18:51 |
| kopecmartin | historic reasons, we slowly update that whenever we need newer changes | 18:51 |
| kopecmartin | we can definitely try to bump tempest version higher | 18:51 |
| gmann | kopecmartin: but main issue is *-tempest-plugins are used from master which is never going to be compatible | 18:52 |
| kopecmartin | yeah, true, it was working till now, so I didn't want to pin in it it wasn't necessary | 18:52 |
| gmann | kopecmartin: whatever tempest it use same compatible *-tempest-plugins it needs to use otherwise we will see similar issue | 18:52 |
| gmann | kopecmartin: yeah, just by luck :) | 18:52 |
| kopecmartin | well :) | 18:53 |
| kopecmartin | i tried to pin designate-tempest-plugin - https://review.opendev.org/c/openinfra/ansible-role-refstack-client/+/815030 .. i hit compact error again | 18:54 |
| kopecmartin | i though i fixed that already | 18:54 |
| kopecmartin | *compat | 18:54 |
| gmann | kopecmartin: in destack stable/stein it is fixed https://review.opendev.org/c/openstack/devstack/+/812092 | 18:56 |
| gmann | johnsom: kopecmartin right way is to pin all the tempest-plugins in refstack with the compatible tempest it use. | 18:56 |
| johnsom | With the services starting to test the secure RBAC stuff, it might be a good time to bump the pin | 18:56 |
| gmann | johnsom: yeah, it was attribute error so easily located otherwise it might become hard to debug | 18:57 |
| gmann | and I will continue the pre-provisioned creds condition tomorrow or next week based on 815048 result | 18:58 |
| johnsom | Ok | 19:00 |
| redrobot | Hi QA friends! | 19:18 |
| redrobot | I have a question about how y'all handle Tempest stuff on stable branches. | 19:18 |
| redrobot | We landed some fixes on master that I need to backport. It's a chain of 3 patches that fix a bug, but I can't land the first one because the gate is failing on stable/xena | 19:19 |
| redrobot | looks like the tempest test being run are from master, which is kind of expected since we don't branch our tempest plugin | 19:20 |
| redrobot | Not sure if I need to squash all patches into one, or disable the gate job temporarily while I land all three patches and then re-enable it | 19:20 |
| clarkb | redrobot: how did you land them on master? | 19:22 |
| gmann | redrobot: that is main reason of using tempest/plugins master on stable | 19:22 |
| gmann | it should run/pass as it was in master | 19:22 |
| gmann | redrobot: if test are modified in master ( hoping in backward compatible change modification) and gate is passing then it should pass in stable branch(using master tempest) also | 19:23 |
| redrobot | so, we landed the patches in Barbican proper, then added the tests to barbican-tempest-plugin | 19:26 |
| gmann | redrobot: or if it is depends on backport to pass test then we go like this process - https://docs.openstack.org/tempest/latest/HACKING.html#bug-fix-on-core-project-needing-tempest-changes | 19:26 |
| gmann | ok | 19:26 |
| redrobot | it was a chain of 3 patches, then a Depends-On: patch for the plugin | 19:26 |
| gmann | redrobot: i see, in that case you can skip the tests and enable it again during backports. | 19:27 |
| redrobot | gmann, ack, that makes sense. I wish I'd seen that HACKING note before I got into this mess, lol | 19:27 |
| gmann | it could be added with feature flag though like what we do in tempest https://docs.openstack.org/tempest/latest/HACKING.html#new-tests-for-new-features | 19:28 |
| gmann | redrobot: we tried to document such info of getting it done in branchless nature in HACKING as that is a common things to take care for branchless tests | 19:29 |
| redrobot | I'll have to review that stuff. Thanks for the help, y'all! | 19:29 |
| *** dviroel|rover is now known as dviroel|rover|afk | 21:58 | |
| clarkb | QA folks review of https://review.opendev.org/c/openstack/devstack/+/815020 would probably start to be good at this point. I haven't been able to find any test failures related to teh chagne yet and this allows centos9-stream/rhel9 to boot on devstack | 22:17 |
Generated by irclog2html.py 2.17.2 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!