Spark + Cassandra cluster setup and usage examples

Set of Ansible deployment roles and Spark examples to run local virtual cluster or deploy C*+Spark cluster at your own servers.

Software you need

Tested with the next versions of tools:

• Ansible 1.8.2
• Vagrant 1.7.2
• SBT 0.13.6

Directory layout

• ./deployment - set of Ansible roles and deployment scripts to automate C*+Spark cluster roll-out. Tested with ubuntu/trusty in Vagrant.
• ./spark-examples - set of examples taken from DataStax Cassandra documentation

Launching the cluster

• For local virtual clusters it is recommended to launch deployment process with ./deployment/deploy.sh with one node, then pack it in Vagrant box to lower down the time of whole deployment. In case of full deploy C* and Spark binaries will be downloaded for each node
• You have to install the same Spark version on your local machine in case you want to launch spark-submit without ssh-ing to the cluster nodes