Missing Hive Execution Jar: /usr/local/hadoop/hive

Question

I have Hadoop 1.0.4 running on a single node cluster set up on my Ubuntu machine.

I did the following steps to download and install the hive release

> svn co http://svn.apache.org/repos/asf/hive/trunk hive    
> cd hive

My hadoop install folder is $HADOOP_HOME/hive and $HADOOP_HOME=/usr/local/hadoop. I set both the environment variables in .bashrc under /home/hduser which is my Hadoop user and the $JAVA_HOME as well.

export HADOOP_HOME=/usr/local/hadoop

export HIVE_HOME=$HADOOP_HOME/hive

export JAVA_HOME=/usr/lib/jvm/java-6-openjdk

I have also added the bin folder of both hadoop and hive in my $PATH variable as follows

export PATH=$PATH:$HADOOP_HOME/bin    
export PATH=$HIVE_HOME/bin:$PATH

But while running hive from CLI, I am getting the below error.

hduser@somnath-laptop:/usr/local/hadoop/hive$ bin/hive    
Missing Hive Execution Jar: /usr/local/hadoop/hive/lib/hive-exec-*.jar

Should I download this jar and add it to the /lib/ or is there some hive-specific environment variables that I need to configure. Any suggestion would be very helpful.

Answer 1

The question is about that the hive path, So you can check up all configuration file involving the hive path. Remember that you must confirm that the hadoop had been installed.

1, the environment parameter(/etc/profile or ~/.profile)

export HIVE_HOME=/usr/app/apache-hive-2.3.0-bin
export PATH=$HIVE_HOME/bin:$PATH

2, $HIVE_HOME/conf/hive-env.sh

export JAVA_HOME= ${Your_JAVA_HOME_directory}
export HADOOP_HOME= ${Your_HADOOP_HOME_directory}
export HIVE_HOME= ${Your_HIVE_HOME_directory}
export HIVE_CONF_DIR= ${Your_HIVE_HOME_directory}/conf

Hive is based on Hadoop, so you must configure the hadoop's path on the hive-env.sh.

Answer 2

i have the same issue,and i use the command "source ~/.bashrc"

problem resolved!

Answer 3

1. Try cross checking your environment variable path, in case you typed it wrong.

2. Try Reloading .bashrc by typing following command

   source ~/.bashrc

3. Try rebooting your machine

Answer 4

I resolved the problem myself but not sure what exactly happened.

By following the process I mentioned in my original question, I created the $HADOOP_HOME/hive but it was giving me a missing jar error.

So, what I did was: I downloaded hive-0.10.0.tar.gz and extracted it under $HADOOP_HOME. So the newly created folder was $HADOOP_HOME/hive-0.10.0.

I copied the entire lot of jars under $HADOOP_HOME/hive-0.10.0/lib to $HADOOP_HOME/hive/lib and when I executed next,

$HADOOP_HOME/hive> bin/hive

It worked! Please note my $HIVE_HOME=$HADOOP_HOME/hive and $HIVE_HOME/bin is added to path. Hope this helps somebody facing similar problem.

Answer 5

Just want to post what worked for me (in 2017).

Using Spark 2.0.2, I had to change my $HIVE_HOME variable (which in my case, constituted me just removing the variable from my .bash_profile.

Hope this helps someone else.