leixm opened a new pull request, #58697: URL: https://github.com/apache/spark/pull/58697
### What changes were proposed in this pull request? This PR associates the representative `BlockStatus` returned by `BlockLocationsAndStatus` with the `BlockManagerId` that owns it. `BlockManager` now accepts an empty remote response only when it is fetched from that exact location and the location reports a deserialized in-memory block. Empty responses from external shuffle services or other locations continue to be treated as fetch failures. The PR also adds regression tests for: - accepting an empty response from the matching deserialized-memory executor; - rejecting an empty external shuffle service response; and - rejecting an empty response from an executor other than the status owner. ### Why are the changes needed? For deserialized in-memory blocks, `memSize` is an estimated size of JVM objects, not the size of the serialized data sent over the network. Therefore, an empty iterator can have a positive `memSize` while its serialized response is empty. The previous check interpreted every empty response for a positive block size as a failed fetch. A simple exception based only on the representative status would be unsafe because the status can belong to a different replica. This PR keeps the corruption and stale-location protection by applying the exception only to the status-owning executor. ### Does this PR introduce _any_ user-facing change? Yes. Remote reads of an affected empty deserialized RDD cache block can now succeed instead of being incorrectly reported as unavailable. There are no public API or configuration changes. ### How was this patch tested? ```bash build/sbt 'core/testOnly org.apache.spark.storage.BlockManagerSuite' The suite completed successfully with 133 tests passed. ###Was this patch authored or co-authored using generative AI tooling? Generated-by: OpenAI Codex 5.6 Terra. -- 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]
