| -@gerrit:opendev.org- Zuul merged on behalf of Monty Taylor https://matrix.to/#/@mordred:inaugust.com: [zuul/zuul-jobs] 922515: Support .python-version files in ensure-python https://review.opendev.org/c/zuul/zuul-jobs/+/922515 | 02:53 | |
| @anoop-jose:matrix.org | While placing a PR in github, it is failing with following message: | 11:16 |
|---|---|---|
| ``` | ||
| Warning: Failed to create check run repo/check: 422 Client Error: Unprocessable Entity for url: https://api.github.com/repos/organisation/repo/check-runs | ||
| ``` | ||
| I have provided read-write permissions in check-runs, status, etc. Has anyone of you have suggestions on it? | ||
| There are no other errors. | ||
| -@gerrit:opendev.org- Benedikt Löffler proposed: [zuul/nodepool] 922014: Relax image status check https://review.opendev.org/c/zuul/nodepool/+/922014 | 11:22 | |
| -@gerrit:opendev.org- Simon Westphahl proposed: | 12:04 | |
| - [zuul/zuul] 921949: Add basic zuul-launcher client/server skeleton https://review.opendev.org/c/zuul/zuul/+/921949 | ||
| - [zuul/zuul] 922019: Properly sort imports in model.py https://review.opendev.org/c/zuul/zuul/+/922019 | ||
| - [zuul/zuul] 922310: Add ZKObject-based launcher API https://review.opendev.org/c/zuul/zuul/+/922310 | ||
| - [zuul/zuul] 922469: Support provider node subclasses in launcher API https://review.opendev.org/c/zuul/zuul/+/922469 | ||
| - [zuul/zuul] 922802: wip: Add launcher client/server implementation https://review.opendev.org/c/zuul/zuul/+/922802 | ||
| -@gerrit:opendev.org- Simon Westphahl proposed: [zuul/zuul] 922802: wip: Add launcher client/server implementation https://review.opendev.org/c/zuul/zuul/+/922802 | 13:17 | |
| @clarkb:matrix.org | > <@anoop-jose:matrix.org> While placing a PR in github, it is failing with following message: | 14:12 |
| > | ||
| > ``` | ||
| > Warning: Failed to create check run repo/check: 422 Client Error: Unprocessable Entity for url: https://api.github.com/repos/organisation/repo/check-runs | ||
| > ``` | ||
| > | ||
| > I have provided read-write permissions in check-runs, status, etc. Has anyone of you have suggestions on it? | ||
| > | ||
| > There are no other errors. | ||
| The GitHub API docs indicate only GitHub apps can request check runs. Is your zuul connection set up with an app? | ||
| -@gerrit:opendev.org- Clark Boylan proposed: [zuul/zuul-jobs] 922643: Switch fips fole testing to CentOS 9 Stream https://review.opendev.org/c/zuul/zuul-jobs/+/922643 | 19:53 | |
| @jim:acmegating.com | release announcement sent: https://lists.zuul-ci.org/archives/list/zuul-announce@lists.zuul-ci.org/thread/X4KBKCKSVR2WCTEVFP4MCA6QM3X6PHN3/ | 20:38 |
| @clarkb:matrix.org | oh shoot I was going to check the release proposal then spaced on ti. But I think getting a release out with those bugfixes was a good idea | 20:39 |
| @clarkb:matrix.org | fungi: so 922643 works for fips testing but now python3.12 testing fails against the mimetypes override for yaml files in the testing for swift log uploads | 20:40 |
| @clarkb:matrix.org | it is almost like the from .foo import bar isn't importing all of foo to then namespace bar in but instead only loading bar which means the top level mimetypes override in foo isn't executing under python3.12 | 20:40 |
| @clarkb:matrix.org | https://zuul.opendev.org/t/zuul/build/291d252fc2ed4a1f992503dcd3f6989f/log/job-output.txt#1583-1599 is the failure and I think it is due to https://opendev.org/zuul/zuul-jobs/src/branch/master/roles/test-upload-logs-swift/library/zuul_swift_upload.py#L64 not being in place when https://opendev.org/zuul/zuul-jobs/src/branch/master/roles/test-upload-logs-swift/library/test_zuul_swift_upload.py#L256 runs | 20:43 |
| @clarkb:matrix.org | I cannot reproduce this running `tox -epy312 -- test_index_files_trailing_slash` against python3.12 on my tumbleweed install. That job appears to have run with the noble python3.12 | 20:53 |
| @fungicide:matrix.org | Clark: yep, just spotted that and started looking into it myself. it's strange, definitely impacting a swath of tests under 3.12 | 20:58 |
| @clarkb:matrix.org | its all the tests that check mimetypes and check a yaml file value | 20:58 |
| @clarkb:matrix.org | but my local python3.12 doesn't reproduce so potentially something weird about ubuntu's python3.12? | 20:58 |
| @clarkb:matrix.org | looks like explicit init() calls are not necessary as the other functions call it implicity if necessary. I half wonder if we're reiniting and removing the old types that were added | 21:00 |
| @clarkb:matrix.org | I have confirmed that if you call init() again after add_type() the added type is forgotten | 21:01 |
| @fungicide:matrix.org | and only under (a recent point release of) 3.12? | 21:01 |
| @fungicide:matrix.org | or possibly with some ubuntu-specific patch over 3.12 | 21:02 |
| @clarkb:matrix.org | I've only confirmed that under python3.12. Let me see what 3.11 does | 21:02 |
| @clarkb:matrix.org | 3.11 has the same behavior | 21:02 |
| @clarkb:matrix.org | but maybe only 3.12 is calling init() more for us? I'm mostly operating on a hunch here and don't have any concrete evidence for that as I haven't been able to reproduce locally yet. Maybe we should just hold a node | 21:03 |
| @fungicide:matrix.org | but passes tests | 21:03 |
| @fungicide:matrix.org | i have a from-source build of 3.12.3 on debian/unstable i can test with | 21:05 |
| @clarkb:matrix.org | fungi: yes my local python3.12 passes the tests that fil | 21:07 |
| @clarkb:matrix.org | though I did only run the one test case and maybe it is a side effect of other test cases? | 21:08 |
| @fungicide:matrix.org | "requests.exceptions.RequestException: Failed for a reason" | 21:08 |
| @fungicide:matrix.org | i'm clearly making poor life choices | 21:08 |
| @fungicide:matrix.org | oh, apparently that's a successful failure | 21:09 |
| @fungicide:matrix.org | py312: OK (208.91=setup[143.86]+cmd[64.72,0.33] seconds) congratulations :) (208.98 seconds) | 21:09 |
| @clarkb:matrix.org | yes those exceptions are in the job output too. I suspect we're testing negative cases as well as positive ones | 21:10 |
| @clarkb:matrix.org | the actual reason the test case seems to fail is the mismatched mimetype | 21:10 |
| @fungicide:matrix.org | could be something more insidious like mime data changes on jammy | 21:10 |
| @clarkb:matrix.org | noble in this case, but we're explicitly setting the mimetype | 21:11 |
| @clarkb:matrix.org | which is why this is confusing we call mimetypes.add_type('text/plain', '.yaml') so even if the platform has different ypes we should override | 21:12 |
| @clarkb:matrix.org | but we seem to get back the default of application/yaml and then fail the check that we get 'text/plain' | 21:12 |
| @clarkb:matrix.org | I think I'll recheck and set a hold | 21:19 |
| @clarkb:matrix.org | it does reproduce reliably on the held node | 21:33 |
| @clarkb:matrix.org | if I manually run mimetypes stuff it works the way I woudl expect. So now to dive into the test cases themselves I guess | 21:35 |
| @clarkb:matrix.org | modifying the code experimentall I can show that the type starts as application/yaml then is changed to text/plain after we set it and then something switches it back again | 21:51 |
| @fungicide:matrix.org | but only on noble | 21:52 |
| @fungicide:matrix.org | chaos in action | 21:53 |
| @clarkb:matrix.org | ok I got the test to pass. It appears that the init in roles/generate-zuul-manifest/library/generate_manifest.py is reseting things for some reason | 21:58 |
| @clarkb:matrix.org | I don't see where we're importing that yet but we must be. I suspect we can simply remove the mimetypes.init() there since init() is supposed to be called for us when needed | 21:58 |
| @clarkb:matrix.org | the test cases for generate_manifest import generate_manifest so this may be an import order issue | 22:01 |
| @clarkb:matrix.org | I'll write a patch | 22:01 |
| -@gerrit:opendev.org- Clark Boylan proposed: | 22:07 | |
| - [zuul/zuul-jobs] 922643: Switch fips fole testing to CentOS 9 Stream https://review.opendev.org/c/zuul/zuul-jobs/+/922643 | ||
| - [zuul/zuul-jobs] 922893: Be more cautious initing mimetypes https://review.opendev.org/c/zuul/zuul-jobs/+/922893 | ||
| @fungicide:matrix.org | still odd that it seems to work on 3.12 everywhere else just not on ubuntu | 22:11 |
| @clarkb:matrix.org | fungi: ya digging into that I think I may know the reason. mimetypes.types_map has .yaml in it on ubuntu by default but not on my tumbleweed install | 22:15 |
| @clarkb:matrix.org | so maybe when it resets on tumbleweed it isn't resetting that value and/or we're falling back to text/plain which is the zuul job role default | 22:16 |
| @clarkb:matrix.org | I'm beginning to think that noble just has more info in its mimetypes files lookup info | 22:16 |
| @clarkb:matrix.org | ya on tumbleweed /etc/mime.types has no yaml entry that same fiel does have yaml on noble | 22:17 |
| @clarkb:matrix.org | I'm guessing this is the platform difference that is hitting us. fungi maybe you can check on your debian instal | 22:17 |
| @fungicide:matrix.org | Clark: neither my custom from-source 3.12 build nor the one shipped in debian/sid's packaging has a `mimetypes.types_map['.yaml']` | 22:40 |
| @fungicide:matrix.org | `KeyError: '.yaml'` in both cases | 22:41 |
| @fungicide:matrix.org | must be an ubuntu-specific patch | 22:41 |
| @clarkb:matrix.org | fungi: ya I think it comes from /etc/mime.types | 22:47 |
| @clarkb:matrix.org | and I'm guessing thati s the underlying source of the behavior change | 22:47 |
| @fungicide:matrix.org | i do have `application/yaml` in my /etc/mime.types | 22:54 |
| -@gerrit:opendev.org- Zuul merged on behalf of Clark Boylan: | 22:56 | |
| - [zuul/zuul-jobs] 922893: Be more cautious initing mimetypes https://review.opendev.org/c/zuul/zuul-jobs/+/922893 | ||
| - [zuul/zuul-jobs] 922643: Switch fips fole testing to CentOS 9 Stream https://review.opendev.org/c/zuul/zuul-jobs/+/922643 | ||
| @fungicide:matrix.org | but `'application/yaml' in mimetypes.types_map.values()` returns `False` for me too | 22:56 |
| @clarkb:matrix.org | huh | 23:00 |
| -@gerrit:opendev.org- Aurelio Jargas proposed: [zuul/zuul-jobs] 922286: Add ensure-poetry role https://review.opendev.org/c/zuul/zuul-jobs/+/922286 | 23:33 | |
| @jim:acmegating.com | the quick-start tests are frequently (but not 100%) failing. from logs i've come up with a hypothesis that between the keycloak and tracing parts of the test, the "podman stop" command is issued and the scheduler does not shut down gracefully; that means there's a delay on startup while it waits for the election to time out and it misses the gerrit event.. it stops gracefully as expected locally, so i'm waiting for an autohold to hit so i can inspect it in the test environment. unsure what's different. we might need to adjust podman compose signal timeouts or something. | 23:36 |
| -@gerrit:opendev.org- Zuul merged on behalf of James E. Blair https://matrix.to/#/@jim:acmegating.com: [zuul/zuul] 921479: Add image configuration object https://review.opendev.org/c/zuul/zuul/+/921479 | 23:52 | |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!