[MAPREDUCE-3112] Calling hadoop cli inside mapreduce job leads to errors

• Type: Bug
• Status: Closed
• Priority: Major
• Resolution: Fixed
• Affects Version/s: 0.20.205.0, 0.23.0
• Fix Version/s: 0.20.205.0, 0.23.0
• Component/s: contrib/streaming

Environment

Java, Linux

Release Note

Removed inheritance of certain server environment variables (HADOOP_OPTS and HADOOP_ROOT_LOGGER) in task attempt process.

• Assignee: Eric Yang
• Reporter: Eric Yang
• Created: 28/Sep/11 05:42
• Updated: 19/Oct/11 00:26
• Resolved: 04/Oct/11 00:52

Description

When running a streaming job with mapper

bin/hadoop --config /etc/hadoop/ jar contrib/streaming/hadoop-streaming-0.20.205.0.jar -mapper "hadoop --config /etc/hadoop/ dfs -help" -reducer NONE -input "/tmp/input.txt" -output NONE

Task log shows:

Exception in thread "main" java.lang.ExceptionInInitializerError
	at org.apache.hadoop.util.ToolRunner.run(ToolRunner.java:57)
	at org.apache.hadoop.util.ToolRunner.run(ToolRunner.java:79)
	at org.apache.hadoop.fs.FsShell.main(FsShell.java:1895)
Caused by: org.apache.commons.logging.LogConfigurationException: User-specified log class 'org.apache.commons.logging.impl.Log4JLogger' cannot be found or is not useable.
	at org.apache.commons.logging.impl.LogFactoryImpl.discoverLogImplementation(LogFactoryImpl.java:874)
	at org.apache.commons.logging.impl.LogFactoryImpl.newInstance(LogFactoryImpl.java:604)
	at org.apache.commons.logging.impl.LogFactoryImpl.getInstance(LogFactoryImpl.java:336)
	at org.apache.commons.logging.impl.LogFactoryImpl.getInstance(LogFactoryImpl.java:310)
	at org.apache.commons.logging.LogFactory.getLog(LogFactory.java:685)
	at org.apache.hadoop.conf.Configuration.<clinit>(Configuration.java:142)
	... 3 more
java.lang.RuntimeException: PipeMapRed.waitOutputThreads(): subprocess failed with code 1
	at org.apache.hadoop.streaming.PipeMapRed.waitOutputThreads(PipeMapRed.java:311)
	at org.apache.hadoop.streaming.PipeMapRed.mapRedFinished(PipeMapRed.java:545)
	at org.apache.hadoop.streaming.PipeMapper.close(PipeMapper.java:132)
	at org.apache.hadoop.mapred.MapRunner.run(MapRunner.java:57)
	at org.apache.hadoop.streaming.PipeMapRunner.run(PipeMapRunner.java:36)
	at org.apache.hadoop.mapred.MapTask.runOldMapper(MapTask.java:436)
	at org.apache.hadoop.mapred.MapTask.run(MapTask.java:372)
	at org.apache.hadoop.mapred.Child$4.run(Child.java:261)
	at java.security.AccessController.doPrivileged(Native Method)
	at javax.security.auth.Subject.doAs(Subject.java:396)
	at org.apache.hadoop.security.UserGroupInformation.doAs(UserGroupInformation.java:1059)
	at org.apache.hadoop.mapred.Child.main(Child.java:255)

Upon inspection, there are two problems in the inherited from environment which prevent the logger initialization to work properly. In hadoop-env.sh, the HADOOP_OPTS is inherited from the parent process. This configuration was requested by user to have a way to override HADOOP environment in the configuration template:

export HADOOP_OPTS="-Djava.net.preferIPv4Stack=true $HADOOP_OPTS"

-Dhadoop.log.dir=$HADOOP_LOG_DIR/task_tracker_user is injected into HADOOP_OPTS in the tasktracker environment. Hence, the running task would inherit the wrong logging directory, which the end user might not have sufficient access to write. Second, $HADOOP_ROOT_LOGGER is override to: -Dhadoop.root.logger=INFO,TLA by the task controller, therefore, the bin/hadoop script will attempt to use hadoop.root.logger=INFO,TLA, but fail to initialize.

Issue Links

• blocks: MAPREDUCE-3080 dfs calls from streaming fails with ExceptionInInitializerError