[SPARK-7837] NPE when save as parquet in speculative tasks

• Type: Bug
• Status: Resolved
• Priority: Critical
• Resolution: Fixed
• Affects Version/s: 1.4.0
• Fix Version/s: 1.5.0
• Component/s: SQL
• Target Version/s: 1.5.0
• Assignee: Cheng Lian
• Reporter: Yin Huai
• Created: 23/May/15 00:15
• Updated: 21/Jul/22 12:46
• Resolved: 17/Aug/15 16:59

Description

The query is like df.orderBy(...).saveAsTable(...).

When there is no partitioning columns and there is a skewed key, I found the following exception in speculative tasks. After these failures, seems we could not call SparkHadoopMapRedUtil.commitTask correctly.

java.lang.NullPointerException
	at parquet.hadoop.InternalParquetRecordWriter.flushRowGroupToStore(InternalParquetRecordWriter.java:146)
	at parquet.hadoop.InternalParquetRecordWriter.close(InternalParquetRecordWriter.java:112)
	at parquet.hadoop.ParquetRecordWriter.close(ParquetRecordWriter.java:73)
	at org.apache.spark.sql.parquet.ParquetOutputWriter.close(newParquet.scala:115)
	at org.apache.spark.sql.sources.DefaultWriterContainer.abortTask(commands.scala:385)
	at org.apache.spark.sql.sources.InsertIntoHadoopFsRelation.org$apache$spark$sql$sources$InsertIntoHadoopFsRelation$$writeRows$1(commands.scala:150)
	at org.apache.spark.sql.sources.InsertIntoHadoopFsRelation$$anonfun$insert$1.apply(commands.scala:122)
	at org.apache.spark.sql.sources.InsertIntoHadoopFsRelation$$anonfun$insert$1.apply(commands.scala:122)
	at org.apache.spark.scheduler.ResultTask.runTask(ResultTask.scala:63)
	at org.apache.spark.scheduler.Task.run(Task.scala:70)
	at org.apache.spark.executor.Executor$TaskRunner.run(Executor.scala:213)
	at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1145)
	at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:615)
	at java.lang.Thread.run(Thread.java:745)

Issue Links

• Is contained by: SPARK-9072 Parquet : Writing data to S3 very slowly
• links to: [Github] Pull Request #8236 (liancheng)
• links to: [Github] Pull Request #37245 (LuciferYang)
• links to: [Github] Pull Request #37245 (LuciferYang)