szehon-ho opened a new pull request, #58053: URL: https://github.com/apache/spark/pull/58053
### What changes were proposed in this pull request? Document how to fix local unit-test failures caused by Spark binding to a non-loopback network interface (common with VPN / multi-homed machines): - Add a short **Local network binding (`RemoteClassLoaderError`)** section under Running Tests in `docs/building-spark.md`, recommending `export SPARK_LOCAL_IP=127.0.0.1`. - Add a one-line pointer in `AGENTS.md` so coding agents that run Spark tests hit the same guidance. ### Why are the changes needed? `SPARK_LOCAL_IP` is already listed in the Configuration Guide as a deploy-time env var, but contributor / test docs do not mention this failure mode. Local tests can fail widely with `RemoteClassLoaderError` (e.g. fetching `.../Object.class`) and Netty `Connection reset by peer` / `Broken pipe` even though the code under test is fine. ### Does this PR introduce _any_ user-facing change? No. Documentation / agent-instructions only. ### How was this patch tested? N/A (docs). The same `SPARK_LOCAL_IP=127.0.0.1` workaround was verified locally when running AutoCDC suites (`pipelines/testOnly *AutoCdc*`: 193/193 passed after setting it; previously failed with `RemoteClassLoaderError`). ### Was this patch authored or co-authored using generative AI tooling? Generated-by: Cursor Grok 4.5 -- 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] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
