[MAPREDUCE-6609] NativeTask::MCollectorOutputHandler::handleInput causes misaligned memory access coredumps

• Type: Sub-task
• Status: Open
• Priority: Major
• Resolution: Unresolved
• Affects Version/s: 2.7.1
• Component/s: task

Environment

Solaris SPARC

• Assignee: Alan Burlison
• Reporter: Alan Burlison
• Created: 18/Jan/16 13:26
• Updated: 18/Jan/16 13:26

Description

NativeTask::MCollectorOutputHandler::handleInput coredumps becasue of misaligned memory accesses. Details below:

# A fatal error has been detected by the Java Runtime Environment:
#
#  SIGBUS (0xa) at pc=0xffffffff5ea3c80c, pid=250, tid=25
#
# JRE version: Java(TM) SE Runtime Environment (7.0_85-b33) (build 1.7.0_85-b33)
# Java VM: Java HotSpot(TM) 64-Bit Server VM (24.85-b07 mixed mode solaris-sparc compressed oops)
# Problematic frame:
# C  [libnativetask.so.1.0.0+0x3c80c]  NativeTask::MCollectorOutputHandler::handleInput(NativeTask::ByteBuffer&)+0xa0

Stack: [0xffffffff5cf00000,0xffffffff5d000000],  sp=0xffffffff5cffdc10,  free space=1015k
Native frames: (J=compiled Java code, j=interpreted, Vv=VM code, C=native code)
C  [libnativetask.so.1.0.0+0x3c80c]  NativeTask::MCollectorOutputHandler::handleInput(NativeTask::ByteBuffer&)+0xa0
C  [libnativetask.so.1.0.0+0x3c078]  NativeTask::BatchHandler::onInputData(unsigned)+0x38
C  [libnativetask.so.1.0.0+0x3c18c]  Java_org_apache_hadoop_mapred_nativetask_NativeBatchProcessor_nativeProcessInput+0x24
j  org.apache.hadoop.mapred.nativetask.NativeBatchProcessor.nativeProcessInput(JI)V+0
j  org.apache.hadoop.mapred.nativetask.NativeBatchProcessor.nativeProcessInput(JI)V+0
j  org.apache.hadoop.mapred.nativetask.NativeBatchProcessor.sendData()V+12
j  org.apache.hadoop.mapred.nativetask.buffer.ByteBufferDataWriter.flush()V+4
j  org.apache.hadoop.mapred.nativetask.serde.KVSerializer.serializePartitionKV(Lorg/apache/hadoop/mapred/nativetask/buffer/DataOutputStream;ILorg/apache/hadoop/mapred/nativetask/util/SizedWritable;Lorg/apache/hadoop/mapred/nativetask/util/SizedWritable;)I+72
j  org.apache.hadoop.mapred.nativetask.handlers.BufferPusher.collect(Ljava/lang/Object;Ljava/lang/Object;I)V+33
j  org.apache.hadoop.mapred.nativetask.handlers.NativeCollectorOnlyHandler.collect(Ljava/lang/Object;Ljava/lang/Object;I)V+7
j  org.apache.hadoop.mapred.nativetask.NativeMapOutputCollectorDelegator.collect(Ljava/lang/Object;Ljava/lang/Object;I)V+7
j  org.apache.hadoop.mapred.nativetask.testutil.EnforceNativeOutputCollectorDelegator.collect(Ljava/lang/Object;Ljava/lang/Object;I)V+11
j  org.apache.hadoop.mapred.MapTask$OldOutputCollector.collect(Ljava/lang/Object;Ljava/lang/Object;)V+21
j  org.apache.hadoop.mapred.nativetask.combinertest.WordCountWithOldAPI$TokenizerMapperWithOldAPI.map(Ljava/lang/Object;Lorg/apache/hadoop/io/Text;Lorg/apache/hadoop/mapred/OutputCollector;Lorg/apache/hadoop/mapred/Reporter;)V+41
j  org.apache.hadoop.mapred.nativetask.combinertest.WordCountWithOldAPI$TokenizerMapperWithOldAPI.map(Ljava/lang/Object;Ljava/lang/Object;Lorg/apache/hadoop/mapred/OutputCollector;Lorg/apache/hadoop/mapred/Reporter;)V+9
j  org.apache.hadoop.mapred.MapRunner.run(Lorg/apache/hadoop/mapred/RecordReader;Lorg/apache/hadoop/mapred/OutputCollector;Lorg/apache/hadoop/mapred/Reporter;)V+39
j  org.apache.hadoop.mapred.MapTask.runOldMapper(Lorg/apache/hadoop/mapred/JobConf;Lorg/apache/hadoop/mapreduce/split/JobSplit$TaskSplitIndex;Lorg/apache/hadoop/mapred/TaskUmbilicalProtocol;Lorg/apache/hadoop/mapred/Task$TaskReporter;)V+204
j  org.apache.hadoop.mapred.MapTask.run(Lorg/apache/hadoop/mapred/JobConf;Lorg/apache/hadoop/mapred/TaskUmbilicalProtocol;)V+162
j  org.apache.hadoop.mapred.LocalJobRunner$Job$MapTaskRunnable.run()V+223
j  java.util.concurrent.Executors$RunnableAdapter.call()Ljava/lang/Object;+4
j  java.util.concurrent.FutureTask.run()V+42
j  java.util.concurrent.ThreadPoolExecutor.runWorker(Ljava/util/concurrent/ThreadPoolExecutor$Worker;)V+95
j  java.util.concurrent.ThreadPoolExecutor$Worker.run()V+5
j  java.lang.Thread.run()V+11