Question

What are the features of Fully-Distributed mode in Hadoop?

Answer 1

In this mode, all daemons execute in separate nodes forming a multi-node cluster. Thus, we allow separate nodes for Master and Slave.

We use this mode in the production environment, where ‘n’ number of machines forming a cluster. Hadoop daemons run on a cluster of machines. There is one host onto which NameNode is running and the other hosts on which DataNodes are running. Therefore, NodeManager installs on every DataNode. And it is also responsible for the execution of the task on every single DataNode.

The ResourceManager manages all these NodeManager. ResourceManager receives the processing requests. After that, it passes the parts of the request to corresponding NodeManager accordingly.

🔗Source: Hadoop Interview Questions and Answers

Answer 2

In Fully Distributed mode, the clusters range from a few nodes to 'n' number of nodes. It is used in production environments, where we have thousands of machines in the Hadoop cluster. The daemons of Hadoop run on these clusters. We have to configure separate masters and separate slaves in this distribution, the implementation of which is quite complex. In this configuration, Namenode and Datanode runs on different hosts and there are nodes on which task tracker runs. The root of the distribution is referred as HADOOP_HOME.