Set up Cassandra Cluster with Docker

Here I'm going to set up a Cassandra cluster over several virtual machines tunning Ubuntu operating system, furthermore they will be running under Docker containers.


Not going to discuss about the advantages or when is suitable to use Cassandra as there are many resources over the Web where this can be found. Just to point one of the most interesting ones from my point of view, which is the possibility to add new nodes easily to the cluster in case you are running out of space.

First step is to have at least 2 virtual machines running Ubuntu and docker as explained in my entry another entry Docker install


Lets say you have these IP addresses 192.168.1.74 (Node A) and 192.186.1.75 (Node B)

On Node A type

docker pull cassandra

This will get the latest image version of Cassandra to your Docker.

Then to run it just

docker run -d --name=cassandra1 -e CASSANDRA_BROADCAST_ADDRESS=192.168.1.74 -p 9042:9042 -p 7000-7001:7000-7001 -p 7199:7199 -p 9160:9160 cassandra

Main things there are the broadcast address (so other nodes can see this one) and expose the ports to be able to communicate. Rest ot options are the usual ones to run an image in Docker.

Now we have our first node running.

Lets move to second virtual machine

Once we pull Cassandra we run there

docker run -d --name=cassandra2 -p 9042:9042 -p 7000-7001:7000-7001 -p 7199:7199 -p 9160:9160 -e CASSANDRA_BROADCAST_ADDRESS="192.168.1.75" -e CASSANDRA_SEEDS="192.168.1.74" cassandra

We will have our second instace of Cassandra running

To check that our cluster in correct we can run in any of the nodes this


docker exec cassandra1 nodetool status

And here we have our cluster ready!

In case you want to add a new node to the cluster is a simple as we have dont with the second one.


Obstacles:

One important thing if you are running instances in separated virtual machines is that to make sure you state the bradcast and the seeds IP addresses, in some documentation you can see people just use
the option --net=host so instance just take the host network configuration

Also make sure you expose the ports



References:

https://hub.docker.com/_/cassandra/

https://github.com/spring-projects/spring-boot/tree/master/spring-boot-samples/spring-boot-sample-data-cassandra

https://yurisubach.com/2016/03/24/cassandra-docker-test-cluster/

http://blog.ditullio.fr/2016/06/10/docker-docker-basics-cassandra/