[
https://issues.apache.org/jira/browse/SPARK-57593?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
ASF GitHub Bot updated SPARK-57593:
-----------------------------------
Labels: pull-request-available (was: )
> Python UDF: instrument and byte-bound the pickle input batch
> ------------------------------------------------------------
>
> Key: SPARK-57593
> URL: https://issues.apache.org/jira/browse/SPARK-57593
> Project: Spark
> Issue Type: Task
> Components: PySpark
> Affects Versions: 4.2.0
> Reporter: Tengfei Huang
> Priority: Major
> Labels: pull-request-available
>
> For a regular (non-Arrow, pickle-serialized) Python UDF,
> `BatchEvalPythonExec` groups input rows into batches purely by row count
> (`spark.sql.execution.python.udf.maxRecordsPerBatch`, default 100) and
> pickles the whole batch into one contiguous byte array before sending it to
> the Python worker.
> When the rows are wide (large strings, binary, or decimal columns), a single
> row-count batch can be very large, so the JVM must buffer one oversized
> pickled array on the heap -- amplified by the `ByteArrayOutputStream`
> power-of-two grow-copy and the final toByteArray copy (~2x transient). On
> memory-constrained executors this can drive the executor into an
> `OutOfMemoryError`, with no way to bound it other than lowering the row count
> for every UDF.
> The Arrow Python UDF path already byte-bounds its batches (it checks
> arrowWriter.sizeInBytes() per row and slices before crossing the limit). The
> pickle path has no equivalent: it is bounded only by row count. To avoid
> memory issues crashing executor, we should also bound the batch by size for
> pickle-serialized UDF.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
