Starting and configuring 10s or 100s of virtual machines in a robust and predictable fashion is not a trivial problem. In this talk I will go over several patterns that simplify this task and I will show you a demo based on CloudStack and Hadoop.
When deploying Apache Hadoop on cloud infrastructure the first step is to provision a large number of virtual machines and do a minimal amount of configuration to ensure the environment matches some basic assumptions.
Starting and configuring 10s or 100s of virtual machines in a robust and predictable fashion is not a trivial problem. You have to make sure you use the API in an appropriate manner and you should design each step of this process to be idempotent.
In this talk I will go over several patterns that simplify this task and I will show you a demo based on CloudStack and Hadoop.