ehurheap opened a new pull request, #19828:
URL: https://github.com/apache/hudi/pull/19828

   feat(cli): pass system properties through to archiver HoodieWriteConfig
   
   The ARCHIVE command in SparkMain accepts only positional args
   (minCommits, maxCommits, commitsRetained, enableMetadata, basePath) and
   builds HoodieWriteConfig from just those. Any hoodie.* property set via
   spark-submit — whether on SparkConf or on driver JVM opts — was silently
   ignored, so tuning e.g. hoodie.commits.archival.batch.size at run time
   was not possible.
   
   Merge System.getProperties() into the builder so that
   `-Dhoodie.*=...` in spark.driver.extraJavaOptions reaches the archiver
   config. No effect when no such properties are set.
   
   Immediate motivation: US events archiver hits the JVM 2GB byte-array
   ceiling inside HoodieAvroDataBlock.serializeRecords when batching
   completed clean/rollback/commit metadata for a 6.7M-partition table.
   Setting hoodie.commits.archival.batch.size=1 is the cheapest experiment
   to determine whether batching alone is the problem or a single instant's
   serialized metadata exceeds 2GB on its own — but requires this
   passthrough to take effect.
   


-- 
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]

Reply via email to