If I understand your problem correctly, one solution that worked for me is to use the -libjars flag when launching your hadoop job:
bin/hadoop jar -libjars <comma separated jars> yourMainClass.jar <args>... I used this solution on my 5-slave cluster. I needed to have the third party jar files to become available to all nodes without me manually distributing them from the master node where I launch my job. Kyle On Mon, 2008-10-13 at 12:11 -0700, Allen Wittenauer wrote: > On 10/13/08 11:06 AM, "Tarandeep Singh" <[EMAIL PROTECTED]> wrote: > > I want to push third party jar files that are required to execute my job, on > > slave machines. What is the best way to do this? > > Use a DistributedCache as part of your job submission. >