[HIVE-8508] UT: fix bucketsort_insert tests - related to SMBMapJoinOperator

• Type: Sub-task
• Status: Resolved
• Priority: Major
• Resolution: Fixed
• Fix Version/s: 1.1.0
• Component/s: Spark
• Assignee: Chinna Rao Lalam
• Reporter: Thomas Friedrich
• Created: 17/Oct/14 22:19
• Updated: 29/May/15 02:30
• Resolved: 11/Dec/14 15:02

Description

The 4 tests
bucketsortoptimize_insert_2
bucketsortoptimize_insert_4
bucketsortoptimize_insert_6
bucketsortoptimize_insert_7
bucketsortoptimize_insert_8

all fail with the same NPE related in SMBMapJoinOperator:

order object is null in SMBMapJoinOperator:
// fetch the first group for all small table aliases
for (byte pos = 0; pos < order.length; pos++)

{ if (pos != posBigTable) { fetchNextGroup(pos); }

Daemon Thread [Executor task launch worker-3] (Suspended (exception NullPointerException))
SMBMapJoinOperator.processOp(Object, int) line: 258
FilterOperator(Operator<T>).forward(Object, ObjectInspector) line: 799
FilterOperator.processOp(Object, int) line: 137
TableScanOperator(Operator<T>).forward(Object, ObjectInspector) line: 799
TableScanOperator.processOp(Object, int) line: 95
MapOperator(Operator<T>).forward(Object, ObjectInspector) line: 799
MapOperator.process(Writable) line: 536
SparkMapRecordHandler.processRow(Object, Object) line: 139
HiveMapFunctionResultList.processNextRecord(Tuple2<BytesWritable,BytesWritable>) line: 47
HiveMapFunctionResultList.processNextRecord(Object) line: 28
HiveBaseFunctionResultList$ResultIterator.hasNext() line: 108
Wrappers$JIteratorWrapper<A>.hasNext() line: 41
Iterator$class.foreach(Iterator, Function1) line: 727
Wrappers$JIteratorWrapper<A>(AbstractIterator<A>).foreach(Function1<A,U>) line: 1157
RDD$$anonfun$foreach$1.apply(Iterator<T>) line: 760
RDD$$anonfun$foreach$1.apply(Object) line: 760
SparkContext$$anonfun$runJob$3.apply(TaskContext, Iterator<T>) line: 1118
SparkContext$$anonfun$runJob$3.apply(Object, Object) line: 1118
ResultTask<T,U>.runTask(TaskContext) line: 61
ResultTask<T,U>(Task<T>).run(long) line: 56
Executor$TaskRunner.run() line: 182
ThreadPoolExecutor.runWorker(ThreadPoolExecutor$Worker) line: 1145
ThreadPoolExecutor$Worker.run() line: 615
Thread.run() line: 745

There is also a NPE in the FileSinkOperator: the FileSystem object fs is null:
// in recent hadoop versions, use deleteOnExit to clean tmp files.
if (isNativeTable) {
autoDelete = fs.deleteOnExit(fsp.outPaths[0]);

Daemon Thread [Executor task launch worker-1] (Suspended (exception NullPointerException))
FileSinkOperator.createBucketFiles(FileSinkOperator$FSPaths) line: 495
FileSinkOperator.closeOp(boolean) line: 925
FileSinkOperator(Operator<T>).close(boolean) line: 582
SelectOperator(Operator<T>).close(boolean) line: 594
SMBMapJoinOperator(Operator<T>).close(boolean) line: 594
DummyStoreOperator(Operator<T>).close(boolean) line: 594
FilterOperator(Operator<T>).close(boolean) line: 594
TableScanOperator(Operator<T>).close(boolean) line: 594
MapOperator(Operator<T>).close(boolean) line: 594
SparkMapRecordHandler.close() line: 175
HiveMapFunctionResultList.closeRecordProcessor() line: 57
HiveBaseFunctionResultList$ResultIterator.hasNext() line: 122
Wrappers$JIteratorWrapper<A>.hasNext() line: 41
Iterator$class.foreach(Iterator, Function1) line: 727
Wrappers$JIteratorWrapper<A>(AbstractIterator<A>).foreach(Function1<A,U>) line: 1157
RDD$$anonfun$foreach$1.apply(Iterator<T>) line: 760
RDD$$anonfun$foreach$1.apply(Object) line: 760
SparkContext$$anonfun$runJob$3.apply(TaskContext, Iterator<T>) line: 1118
SparkContext$$anonfun$runJob$3.apply(Object, Object) line: 1118
ResultTask<T,U>.runTask(TaskContext) line: 61
ResultTask<T,U>(Task<T>).run(long) line: 56
Executor$TaskRunner.run() line: 182
ThreadPoolExecutor.runWorker(ThreadPoolExecutor$Worker) line: 1145
ThreadPoolExecutor$Worker.run() line: 615
Thread.run() line: 745