Setting up a single node Hadoop cluster on Mac

Question

After setting Java path and Hadoop distribution file, when I'm trying to check the usage document I'm getting errors. I'm confused even if have set the path right. This is how I have set the path.

The following is displayed in terminal

/Users/hadoop/hadoop-2.6.0/etc/hadoop/hadoop-env.sh: line 26: JAVA_HOME=${/Library/Java/JavaVirtualMachines/jdk1.8.0_40.jdk/Contents/Home }: bad substitution

/Users/hadoop/hadoop-2.6.0/etc/hadoop/hadoop-env.sh: line 35: HADOOP_CONF_DIR=${"/Users/hadoop/hadoop-2.6.0 "}: bad substitution

Answer 1

By doing ${...} you are trying to do a variable substitution, thus your error.

Change to JAVA_HOME=/Library/Java/JavaVirtualMachines/jdk1.8.0_40.jdk/Contents/Home instead.

---

Alternatively, I think this is better

export JAVA_HOME=`/usr/libexec/java_home`

And you can see it works like so

$ echo $JAVA_HOME                                               
/Library/Java/JavaVirtualMachines/jdk1.8.0_74.jdk/Contents/Home

---

Also, following the tutorial , it doesn't literally mean put the {/path/to/folder}

export HADOOP_PREFIX={your hadoop distribution directory}

It wants you to do this for wherever you downloaded the package to

export HADOOP_PREFIX=/Users/hadoop/Downloads/hadoop-2.6.0

---

Then adding to the PATH would look something like

export PATH=$JAVA_HOME:$PATH
export PATH=$HADOOP_PREFIX:$PATH