HDFS is a basic, low level, os related strage model. YARN is a resource(hardware) manager or negotiator.

Just as the title says, it talks mainly on HDFS and YARN since we don’t use hadoop very often.

A key-value database based on HDFS which relies on zookeeper to manage its cluster. Oh, it stoges data by column, which is better to timely insert and random query, especially for big, big tables.

You can learn Architecture of Hbase Cluster(Region, HMaster…), CRUD Logic from this post.