thisisnic commented on issue #33189: URL: https://github.com/apache/arrow/issues/33189#issuecomment-6764154112
I tried to reproduce this with a reprex that starts 32 threads at once, each calling `FileSystem.from_uri()` on an `s3://` URI and then `pq.read_table()`, in a fresh process each time. It ran cleanly 25 times out of 25 on both pyarrow 9.0.0 and 26.0.1. The backtrace above shows a thread waiting on a lock inside OpenSSL during the bucket region lookup rather than the thread that crashed, so it doesn't tell us where the heap corruption came from. The code on that path (`EnsureS3Initialized()`, `RegionResolver::DefaultInstance()`, `ResolveS3BucketRegion()`) is guarded by locks in both 9.0.0 and current main. Closing as we can't reproduce it and the full `thread apply all bt` output was never provided. If you still see this on a current release, please open a new issue with the backtrace for all threads. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
