Hadoop Cluster modes

Hadoop Cluster modes

There are 3 different types of Hadoop Cluster. They are

1.       Standalone mode or Local mode

2.       Pseudo-distributed mode

3.       Fully distributed mode

This basically indicates how the Hadoop Cluster is set up. The following diagram shows the minimal setup

 

Hadoop Components

JDK

Operating System (UNIX, CentOS etc.)

 

1.       Standalone mode or Local mode

a.       This mode is purely used for development purpose

b.      Once mapreduce programs are developed and tested, we can migrate it to the PROD environment

c.       There will not be any daemons running

d.      Hadoop configuration files are not set up (i.e.) configuring name node etc. is not happening here

2.       Pseudo-Distributed mode

a.       This mode is also configured for development purpose

b.      All the daemons are running in the machine  (Name node, Data node, Node manager , Resource manager are all configured)

c.       Hadoop configuration files are configured

3.       Fully distributed mode

a.       This is real production environment system

b.      It is configured with multiple machines

c.       Many machines will be configured based on Master/Slave architecture