ajithme commented on issue #24142: [SPARK-27194][core] Job failures when task attempts do not clean up spark-staging parquet files URL: https://github.com/apache/spark/pull/24142#issuecomment-474991798 Okay i see the point. So Currently i see in org.apache.spark.internal.io.HadoopMapReduceCommitProtocol#commitJob when ``dynamicPartitionOverwrite`` is true, we do this ``` fs.rename(new Path(stagingDir, part), finalPartPath) ``` This is where we will get the old and the new task file causing duplicates. Would it be better if we could eliminate speculated old tasks files from this operation.?
---------------------------------------------------------------- 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. For queries about this service, please contact Infrastructure at: [email protected] With regards, Apache Git Services --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
