Closed fjetter closed 1 week ago
same thing here https://github.com/dask/distributed/actions/runs/11728636755/job/32672527284
I so hope this is it... those failures have been driving me mad
_See test report for an extended history of previous test failures. This is useful for diagnosing flaky tests._
25 files ±0 25 suites ±0 10h 17m 11s ⏱️ -38s 4 129 tests ±0 4 017 ✅ ±0 110 💤 ±0 1 ❌ - 1 1 🔥 +1 47 681 runs +1 45 559 ✅ +2 2 120 💤 - 1 1 ❌ - 1 1 🔥 +1
For more details on these failures and errors, see this check.
Results for commit d9cbd6dc. ± Comparison against base commit 26b10617.
Well, None of the CLI tests crashed.
We're frequently seeing weird deadlocks / stuck tests when running CLI tests. I've tuned the timeouts a couple of times and for some this was indeed helpful.
I've been looking into this again and saw in https://github.com/dask/distributed/actions/runs/11736945020/job/32696964203 an interesting traceback
This suggests that coverage has a python atexit hook that is locking up for some reason. Possibly because it cannot write the coverage data out quickly enough.
I hope that a quick fix is to just not collect coverage for the CLI tests