What are the advantages disadvantages of Hadoop Dockerization

Question

I am working on a Hadoop cluster which is using Hue, Flume & Cassandra. I have heard about Docker & have an idea about how it works. Before actually deploying the cluster in a real time environment, I want to consider the advantages & disadvantages of using docker container for Hadoop?

I guess the portability is one of the major benefit of using docker, but I am interested in knowing and comparing more. Can anyone help me out on this?

Answer 1

As you are already having a Hadoop cluster, you can understand it is difficult to reproduce this environment. Then next important thing is, docker helps you to isolate the environment which will not conflict any dependencies with other applications present in your host machine.

Talking from the perspective of Hadoop, to easily setup a multi node cluster. You can setup one docker Hadoop container. Then replicate the container and the change the setting. So, it will be very easy to setup a multi node cluster.

But again I would like to add, if you have no pain in setting a multi node cluster or you have no issue with the dependencies, So, you do not have to use the docker container because it is a hot topic. It only depends on your need and your ease.