For those looking to learn Hadoop, the easiest method I found for experimenting is the Hortonworks Sandbox. It is an all in one Virtual Machine that you can use to learn Hadoop instead of spending your time on trying to install and configure a Hadoop cluster.
I installed the VM on a 64 bit laptop running 16 MB of RAM on Ubuntu 12 Desktop. Ubuntu allows you to use more memory because of its small footprint on start up.
It is available from the Hortonworks site here. Included in the sandbox is plenty of tutorials to get you started understanding PIG, HIVE, and how the Map Reduce jobs work to process big data.