Mark test_stress_creation_and_deletion as xfail

[jrbourbeau]

xref #5388

cc @crusaderky

[crusaderky]

I'm running stress testing to see if the flaky flag actually makes the test pass. We've had several cases in the past where a flaky test would either pass on the first go or fail 10 times in a row (which highlights that the whole VM, container, or process is borked).

My stress test will take 1-2 days to run. Happy to have this merged into main as a tentative fix while we wait.

[jrbourbeau]

That makes sense. FWIW anecdotally this seems to help based on the CI runs for this commit that already finished over in my fork of distributed. Though I agree we should revisit based on the results on your stress tests (thank you for taking the time/effort to run those btw)

[jakirkham]

Yeah it seems it still fails in this CI job

Also seeing this failure ( #3574 ) in this CI job.

[crusaderky]

As I feared.

distributed/tests/test_stress.py::test_stress_creation_and_deletion RERUN [ 86%]
distributed/tests/test_stress.py::test_stress_creation_and_deletion RERUN [ 86%]
distributed/tests/test_stress.py::test_stress_creation_and_deletion RERUN [ 86%]
distributed/tests/test_stress.py::test_stress_creation_and_deletion RERUN [ 86%]
distributed/tests/test_stress.py::test_stress_creation_and_deletion RERUN [ 86%]
distributed/tests/test_stress.py::test_stress_creation_and_deletion RERUN [ 86%]
distributed/tests/test_stress.py::test_stress_creation_and_deletion RERUN [ 86%]
distributed/tests/test_stress.py::test_stress_creation_and_deletion RERUN [ 86%]
distributed/tests/test_stress.py::test_stress_creation_and_deletion RERUN [ 86%]
distributed/tests/test_stress.py::test_stress_creation_and_deletion RERUN [ 86%]
distributed/tests/test_stress.py::test_stress_creation_and_deletion FAILED [ 86%]

[crusaderky]

Marker should be changed to xfail

[jrbourbeau]

Thanks for reviewing @jakirkham @crusaderky! CI failures are unrelated