Clustered Services With Apache Whirr: From Ops Down

Apache Whirr is an incubator project focused on simplifying management of distributed services such as Hadoop, ElasticSearch, and Cassandra. Using Whirr, you can (in a single line) startup a cluster from scratch in your cloud provider of choice, or even from a list of machines in a text file. This discussion will show how you can create support for a new service without writing code. We'll review the an interaction module based on JSON formatted node lists and Puppet modules, in context of a working Cassandra configuration. You'll understand pros and cons of this approach, and see where it makes sense to use master vs masterless Puppet mechanisms.