Import data : Flume and Sqoop play a special role in the Hadoop ecosystem. They transport data from sources like local file systems, HTTP, MySQL and Twitter which hold/produce data to data stores like HDFS, HBase and Hive. Both tools come with built-in functionality and abstract away users from the complexity of transporting data between these systems. 

Flume: Flume Agents can transport data produced by a streaming application to data stores like HDFS and HBase. 

Sqoop: Use Sqoop to bulk import data from traditional RDBMS to Hadoop storage architectures like HDFS or Hive. 

Course Objective:

Practical implementations for a variety of sources and data stores ..

• Sources : Twitter, MySQL, Spooling Directory, HTTP
• Sinks : HDFS, HBase, Hive

Flume features : 

Flume Agents, Flume Events, Event bucketing, Channel selectors, Interceptors

Sqoop features : 

Sqoop import from MySQL, Incremental imports using Sqoop Jobs

Audience:

• Engineers building an application with HDFS/HBase/Hive as the data store
• Engineers who want to port data from legacy data stores to HDFS

Prerequisite:

• Knowledge of HDFS is a prerequisite for the course
• HBase and Hive examples assume basic understanding of HBase and Hive shells
• HDFS is required to run most of the examples, so you'll need to have a working installation of HDFS