Before starting to learn Hadoop, install a Linux operating system and use Ubuntu for a server distribution because installing Hadoop in Linux is easy. Try to learn basis Linux commands. Learn Java or Scala programming languages. Also, learn SQL because packages like Hive, Pig, Hbase use SQL like queries.
Having theoretical knowledge is necessary for Hadoop. First, start with the Hadoop architecture. After that, learn Hbase, Hive, Pig, Scoop, Flume. Practice Hadoop on the VM provided by Cloudera. There are a lot of active bloggers who write about Hadoop. Read those blogs on Hadoop to keep you updated. Take up a good course, follow their curriculum, and get a certification.
You can enroll in this Big Data Hadoop training course if you are interested.
For more information about Hadoop and how to learn Hadoop, you can watch this video: