Stuck on an issue?

Lightrun Answers was designed to reduce the constant googling that comes with debugging 3rd party libraries. It collects links to all the places you might be looking at while hunting down a tough bug.

And, if you’re still stuck at the end, we’re happy to hop on a call to see how we can help out.

[SUPPORT] Auto-clean doesn't work

See original GitHub issue

I’m trying to use Hudi with Spark EMR. Everything is ok when I run a batch job of S3 data. But when I run it of Kinesis stream it creates tens of versions of the output file and never removes them.

To Reproduce

This is my code

  val hudiOptions = Map[String,String](
    DataSourceWriteOptions.TABLE_TYPE_OPT_KEY -> DataSourceWriteOptions.COW_TABLE_TYPE_OPT_VAL,
    DataSourceWriteOptions.OPERATION_OPT_KEY -> DataSourceWriteOptions.UPSERT_OPERATION_OPT_VAL,
    HoodieStorageConfig.PARQUET_COMPRESSION_CODEC -> "snappy",
    HoodieCompactionConfig.AUTO_CLEAN_PROP -> "true",
    HoodieCompactionConfig.CLEANER_FILE_VERSIONS_RETAINED_PROP -> "1",
    HoodieCompactionConfig.CLEANER_COMMITS_RETAINED_PROP -> "1",
    DataSourceWriteOptions.HIVE_STYLE_PARTITIONING_OPT_KEY -> "true",
    "hoodie.upsert.shuffle.parallelism" -> "5",
    HoodieCompactionConfig.PARQUET_SMALL_FILE_LIMIT_BYTES -> (512 * 1024 * 1024).toString,
    "hoodie.combine.before.insert" -> "true",
    DataSourceWriteOptions.INSERT_DROP_DUPS_OPT_KEY -> "true"
  )

  dataframe
    .write
    .format("org.apache.hudi")
    .options(hudiOptions)
    .option(HoodieWriteConfig.TABLE_NAME, tableName)
    .option(DataSourceWriteOptions.TABLE_NAME_OPT_KEY, tableName)
    .option(DataSourceWriteOptions.PARTITIONPATH_FIELD_OPT_KEY, partitionPathField)
    .option(DataSourceWriteOptions.RECORDKEY_FIELD_OPT_KEY, recordKeyField)
    .option(DataSourceWriteOptions.PRECOMBINE_FIELD_OPT_KEY, precombineFieldKey)
    .mode(SaveMode.Append)
    .save(destinationPath)

Environment Description