brijrajk opened a new pull request, #12333:
URL: https://github.com/apache/gluten/pull/12333

   ## What changes are proposed in this pull request?
   
   When `spark.io.compression.codec` is set to a codec not natively
   supported by the Gluten shuffle writer (e.g. `snappy`, `none`),
   `GlutenShuffleUtils.getCompressionCodec()` was throwing
   `IllegalArgumentException` inside executor tasks, crashing the job.
   
   Root cause: the `case None =>` path (no explicit Gluten codec configured)
   picked up `spark.io.compression.codec` and validated it against the
   backend's supported set (`lz4`, `zstd` for Velox). Because Gluten's
   native Arrow IPC-based shuffle is independent of Spark's RDD/shuffle
   compression, an unsupported value in `spark.io.compression.codec` should
   not be fatal.
   
   Fix: when `spark.io.compression.codec` contains an unsupported codec and
   no explicit Gluten codec is set, log a warning and fall back to `zstd`.
   The strict error is preserved for `spark.gluten.sql.columnar.shuffle.codec`
   (the user explicitly asked Gluten to use that codec).
   
   ## How was this patch tested?
   
   Unit-style test in `MiscOperatorSuite` that sets 
`spark.io.compression.codec=snappy`
   in a cloned `SparkConf` and asserts `getCompressionCodec` returns `"zstd"`.
   
   ## Was this patch authored or co-authored using generative AI tooling?
   
   Generated-by: Claude Code
   
   Related issue: #11539


-- 
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]

Reply via email to