Thanks Shekar & murali.

I have modified PATH and flume works now :)


On Thu, May 30, 2013 at 10:45 PM, murali adireddy <> wrote:
Hi Lenin,

You should add /bin too to the hadoop path(/usr/local/hadoop/hadoop)

Like in your case you should add "/usr/local/hadoop/hadoop/bin" to run hadoop command from any folder.

Your path should be some thing like.



On Thu, May 30, 2013 at 9:09 PM, Lenin Raj <> wrote:

I am trying to add Hadoop to PATH so Flume can access the necessary JAR files.

I have Hadoop installed in

root@li339-83:/usr/local/hadoop/hadoop# ls
bin        CHANGES.txt  docs                     hadoop-core-1.0.4.jar         hadoop-test-1.0.4.jar   ivy.xml  LICENSE.txt  README.txt  src
build.xml  conf         hadoop-ant-1.0.4.jar     hadoop-examples-1.0.4.jar     hadoop-tools-1.0.4.jar  lib      logs         sbin        webapps
c++        contrib      hadoop-client-1.0.4.jar  hadoop-minicluster-1.0.4.jar  ivy                     libexec  NOTICE.txt   share

And my PATH is
root@li339-83:/usr/local/hadoop/hadoop# echo $PATH

Since /usr/local/hadoop/hadoop is in PATH, I should be able to run "bin/hadoop" from any folder.

But, not able to run hadoop from other folders.

root@li339-83:~# bin/hadoop
-bash: bin/hadoop: No such file or directory

And which hadoop & which bin/hadoop returns nothing.

root@li339-83:~# which hadoop
root@li339-83:~# which bin/hadoop

According to this,
Hadoop JAR files should be shown when we execute  hadoop classpath from any folder.

How should I change my PATH variable to make flume work?