如何设置EMR类路径

Question

I am running a job on an AWS EMR cluster, and am having issues with a Jackson library conflict. Based on the article here I tried to add a bootstrap step to set my classpath with the following script:

#!/bin/bash
export HADOOP_USER_CLASSPATH_FIRST=true;
echo "HADOOP_CLASSPATH=s3n://bucket/myjar.jar" > /home/hadoop/conf/hadoop-user-env.sh

I have built my jar so that all its dependencies are included with it. The first problem I have when I do this is that my enable debugging step that I have dies with the following error:

Exception in thread "main" java.lang.RuntimeException: java.lang.ClassNotFoundException: Class com.amazon.ws.emr.hadoop.fs.EmrFileSystem not found
at org.apache.hadoop.conf.Configuration.getClass(Configuration.java:1895)
at org.apache.hadoop.fs.FileSystem.getFileSystemClass(FileSystem.java:2427)
at org.apache.hadoop.fs.FileSystem.createFileSystem(FileSystem.java:2440)
at org.apache.hadoop.fs.FileSystem.access$200(FileSystem.java:88)
at org.apache.hadoop.fs.FileSystem$Cache.getInternal(FileSystem.java:2479)
at org.apache.hadoop.fs.FileSystem$Cache.get(FileSystem.java:2461)
at org.apache.hadoop.fs.FileSystem.get(FileSystem.java:372)
at com.amazon.elasticmapreduce.scriptrunner.ScriptRunner.fetchFile(ScriptRunner.java:39)
at com.amazon.elasticmapreduce.scriptrunner.ScriptRunner.main(ScriptRunner.java:56)
at sun.reflect.NativeMethodAccessorImpl.invoke0(Native Method)
at sun.reflect.NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.java:57)
at sun.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:43)
at java.lang.reflect.Method.invoke(Method.java:606)
at org.apache.hadoop.util.RunJar.main(RunJar.java:212)
Caused by: java.lang.ClassNotFoundException: Class com.amazon.ws.emr.hadoop.fs.EmrFileSystem not found
at org.apache.hadoop.conf.Configuration.getClassByName(Configuration.java:1801)
at org.apache.hadoop.conf.Configuration.getClass(Configuration.java:1893)
... 13 more

So I have two questions, what is wrong with this regards to the enable debugging step also? Is it valid to give my classpath as a s3 location? If not what should the value of:

/path/to/my.jar

be in the example on the page indicated above?

Answer 1

Looking at your bootstrap action, it looks like there might be a mistake in your string. The line should look like the following:

#!/bin/bash
export HADOOP_USER_CLASSPATH_FIRST=true
echo "HADOOP_CLASSPATH=/path/to/my.jar" >> /home/hadoop/conf/hadoop-user-env.sh

Note the ' >> ' characters. A single ' > ' means that you're replacing the entire file with the output of the 'echo' command, whereas a double '>>' means you're appending that line at the end of the script. Additionally, a semi-colon isn't needed in a Bash script.

References : http://docs.aws.amazon.com/ElasticMapReduce/latest/DeveloperGuide/emr-hadoop-config_hadoop-user-env.sh.html

PS : Amazon's awesome support found this question and replied to my email; although this question was not asked by me. So this is the attribution to the author - AWS Support Engineer named Rendy O.