shubhambg95 opened a new issue, #7125: URL: https://github.com/apache/hudi/issues/7125
**Describe the problem you faced** We currently have a spark streaming job that writes data from a Kafka topic to a Hudi MOR table. We want to enable column stats indexing to leverage data skipping while reading from this table. To do this we basically set the ```hoodie.metadata.index.column.stats.enable = true```. When we enable this change we are seeing some issue with the Hudi cleaner. We are running into clean operations that never finish and remain at ```clean.inflight``` stage. This is also causing an issue with the MDT tasks which fail with ```java.lang.StringIndexOutOfBoundsException at getColumnStatsRecords(HoodieTableMetadataUtil.java:1158)``` When we run without setting ``` hoodie.metadata.index.column.stats.enable = true```, we are not seeing any issues. Hudi Configurations ``` hoodie.datasource.write.table.type: "MERGE_ON_READ" hoodie.datasource.write.precombine.field: preCombineField hoodie.datasource.write.recordkey.field: recordKeyField hoodie.datasource.write.partitionpath.field: partitionPathField hoodie.table.name: tableName hoodie.index.type: "SIMPLE" hoodie.datasource.write.operation: "INSERT" hoodie.datasource.write.hive_style_partitioning: true hoodie.insert.shuffle.parallelism: 54 hoodie.finalize.write.parallelism: 54 hoodie.cleaner.commits.retained: 300 hoodie.keep.min.commits: 325 hoodie.keep.max.commits: 350 hoodie.sql.insert.mode: "upsert" hoodie.parquet.compression.codec: "snappy" hoodie.metadata.index.column.stats.enable: true ``` **Environment Description** * Hudi version : 12.1 * Spark version : 3.1.1 * Hive version : 2.3.7 * Hadoop version : 3.2.0 * Storage (HDFS/S3/GCS..) : GCS * Running on Docker? (yes/no) : yes **Additional context** Add any other context about the problem here. **Stacktrace** ```Add the stacktrace of the error.``` -- 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]
