HDFS

There is a slogan in JAVA Write Once and Read any number of times, but In HADOOP, Write Once Read any number of times and DONT CHANGE THE CONTENT OF FILE (Streaming Access Pattern)

HDFS has 5 services:

• NameNode
• Secondary NameNode
• JobTracker
• DataNode
• TaskTracker

1. First three are the Master Services and rest two are the Slave/Demon Services. We cant see these services working as they all work internally.
2. Every Master Service can talk to each other and so do Slave Services.
3. If 'NameNode' is a master node then its corresponding slave node is 'DataNode'. If 'JobTracker' is a master  node then its corresponding slave node is 'TaskTracker'.
4. One master service can talk to its own slave service but cant talk to another slave service of another master service.
5. NameNode is act as a Manager which leads the data to store into which sector or which region of storage device.