-
Notifications
You must be signed in to change notification settings - Fork 1
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
The centralized nightlies job failed on Friday (2024-08-16) #18
Comments
The centralized nightlies job failed on Saturday (2024-08-17) in run 10437049686 |
The centralized nightlies job failed on Sunday (2024-08-18) in run 10445960340 |
Both TileDB-Go and TileDB-SOMA-Py started failing their tests on Friday night. I had expected these failures because they both also failed earlier on Friday at noon. |
The TileDB-Go test failure has been consistent:
I don't see any open PR/Issue in TileDB-Go related to this. Also, the test file itself, Click here for full traceback
|
Friday night a TileDB-SOMA-Py test caused a segfault:
I assume Saturday night also failed with the same segfault. Unfortunately GitHub didn't save the log for this build, but the tests ran for the same amount of time as Friday prior to failing. I couldn't find a potential cause for this failure. The test file, Click here for full traceback
|
And Sunday night a TileDB-Py test caused a segfault (thus TileDB-SOMA-Py was never built).
This same error also broke the upstream TileDB-Py nightlies over the weekend (TileDB-Inc/TileDB-Py#2040 (comment)) Given that this is also related to Click here for full traceback
|
Actually, skimming |
The centralized nightlies job failed on Monday (2024-08-19) in run 10464107494 |
The TileDB-Py test failure was fixed today in TileDB-Inc/TileDB-Py#2042, which I confirmed via a manual nightly run on main. Thanks @kounelisagis! Note that TileDB-Go and TileDB-SOMA-Py continue to fail their tests |
The centralized nightlies job failed on Tuesday (2024-08-20) in run 10482317549 |
As anticipated, TileDB-Go and TileDB-SOMA-Py continue to fail their tests |
The centralized nightlies job failed on Wednesday (2024-08-21) in run 10500657691 |
The centralized nightlies job failed on Thursday (2024-08-22) in run 10518723380 |
The centralized nightlies job failed on Friday (2024-08-23) in run 10518723380 |
1 similar comment
The centralized nightlies job failed on Friday (2024-08-23) in run 10518723380 |
Lots going on here. Last night there were two failures: TileDB-Go (expected) and TileDB-Py (unexpected) TileDB-Py failed a single test, and it passed on the second attempt =========================== short test summary info ============================
FAILED tiledb/tests/test_libtiledb.py::DenseArrayTest::test_open_with_timestamp[False] - AssertionError
== 1 failed, 961 passed, 6 skipped, 8 xfailed, 1 xpassed in 88.51s (0:01:28) === After TileDB-Py passed, there were two failures: TileDB-SOMA-Py (expected) and TileDB-VCF-Py (unexpected) I restarted TileDB-VCF-Py, but it failed a second time with the same error while building the wheel:
I don't understand why this error is just surfacing now. There have been no updates to TileDB-VCF and none related to the CMake configuration of TileDB. |
The centralized nightlies job failed on Friday (2024-08-23) in run 10534890732 |
The centralized nightlies job failed on Saturday (2024-08-24) in run 10543362914 |
The centralized nightlies job failed on Sunday (2024-08-25) in run 10552357727 |
TileDB-Go and TileDB-SOMA-Py continue to fail their tests. TileDB-VCF-Py continues to fail to build a wheel due to some CMake configuration issue |
Investigating more what could have changed in the TileDB-VCF-Py wheel build. The last passing build was the evening of Wed, Aug 21st. The first failing build was the morning of Fri, Aug 23nd. It didn't run Thursday evening because the TileDB-Py build spuriously failed, so the problem didn't appear until I restarted it Friday morning. Looking for changes in dependencies. The version of CMake is the same, 3.30.2. The version of scikit-build-core changed from 0.10.4 to 0.10.5, but I don't see anything obvious in the 0.10.5 changelog that would affect this behavior. packaging, setuptools, and setuptools-scm are all the same version. I noticed that the failing build checked for the CXX compiler, but presumably the passing build retrieved this from the cache. Maybe the cache has been masking this error until now.
|
The centralized nightlies job failed on Monday (2024-08-26) in run 10570516527 |
The TileDB-Go and TileDB-SOMA-Py test failures were fixed by TileDB-Inc/TileDB#5256 |
So now the only remaining nightly error is the TileDB-VCF-Py wheel build |
The centralized nightlies job failed on Tuesday (2024-08-27) in run 10589002726 |
Still just the failing TileDB-VCF-Py wheel build |
The centralized nightlies job failed on Wednesday (2024-08-28) in run 10607538194 |
Turns out I had seen this same error previously for the TileDB-VCF-Py build (https://github.com/jdblischak/centralized-tiledb-nightlies/issues/6#issuecomment-2113166645), which I had fixed in jdblischak@61765c5 by setting Here's what I don't understand though: why was the behavior so different for the standard build/installation via |
The centralized nightlies job failed on Friday (2024-08-16) in run 10428720345
The text was updated successfully, but these errors were encountered: