[MAPREDUCE-6941] The default setting doesn't work for MapReduce job

• Type: Bug
• Status: Resolved
• Priority: Blocker
• Resolution: Not A Problem
• Affects Version/s: 3.0.0-beta1
• Target Version/s: 3.0.0-beta1
• Assignee: Unassigned
• Reporter: Junping Du
• Created: 16/Aug/17 20:24
• Updated: 25/Oct/19 20:26
• Resolved: 06/Sep/17 00:36

Description

On the deployment of hadoop 3 cluster (based on current trunk branch) with default settings, the MR job will get failed as following exceptions:

2017-08-16 13:00:03,846 INFO mapreduce.Job: Job job_1502913552390_0001 running in uber mode : false
2017-08-16 13:00:03,847 INFO mapreduce.Job:  map 0% reduce 0%
2017-08-16 13:00:03,864 INFO mapreduce.Job: Job job_1502913552390_0001 failed with state FAILED due to: Application application_1502913552390_0001 failed 2 times due to AM Container for appattempt_1502913552390_0001_000002 exited with  exitCode: 1
Failing this attempt.Diagnostics: [2017-08-16 13:00:02.963]Exception from container-launch.
Container id: container_1502913552390_0001_02_000001
Exit code: 1
Stack trace: ExitCodeException exitCode=1:
	at org.apache.hadoop.util.Shell.runCommand(Shell.java:994)
	at org.apache.hadoop.util.Shell.run(Shell.java:887)
	at org.apache.hadoop.util.Shell$ShellCommandExecutor.execute(Shell.java:1212)
	at org.apache.hadoop.yarn.server.nodemanager.DefaultContainerExecutor.launchContainer(DefaultContainerExecutor.java:295)
	at org.apache.hadoop.yarn.server.nodemanager.containermanager.launcher.ContainerLaunch.launchContainer(ContainerLaunch.java:455)
	at org.apache.hadoop.yarn.server.nodemanager.containermanager.launcher.ContainerLaunch.call(ContainerLaunch.java:275)
	at org.apache.hadoop.yarn.server.nodemanager.containermanager.launcher.ContainerLaunch.call(ContainerLaunch.java:90)
	at java.util.concurrent.FutureTask.run(FutureTask.java:266)
	at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1142)
	at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:617)
	at java.lang.Thread.run(Thread.java:745)

This is because mapreduce related jar are not added into yarn setup by default. To make MR job run successful, we need to add following configurations to yarn-site.xml now:

    <property>
      <name>yarn.application.classpath</name>
      <value>
        ...
        <HADOOP_HOME>/share/hadoop/mapreduce/*,
        <HADOOP_HOME>/share/hadoop/mapreduce/lib/*
        ...
      </value>

But this config is not necessary for previous version of Hadoop. We should fix this issue before beta release otherwise it will be a regression for configuration changes.

This could be more like a YARN issue (if so, we should move), depends on how we fix it finally.

Issue Links

• is related to: MAPREDUCE-6704 Update the documents to run MapReduce application