For this recipe, we will need an AWS account. If we do not already have an account, we can navigate to http://aws.amazon.com and sign up for one.
The code in this recipe builds on the micro folder that we've been working on in our previous recipe, Deploying a full system, so we will need to have this code available.
On sign up we will need to provide a payment method to AWS. The cost to run this recipe for an hour or so is in the order of a few dollars. However, please ensure that all resources are terminated after use in order not to incur additional costs.
Next, we will need to install the AWS command line tools. These are implemented in Python and are best installed using pip. Python and pip may be installed using a systems package manager such as homebrew or apt-get.
Pip is the Python equivalent of npm, and it stands for "Pip installs packages". Full details on how to install the AWS tools can be found here: http://docs.aws.amazon.com/cli/latest/userguide/cli-chap-getting-set-up.html.
Once we have installed and configured the tools, we can validate them by running the following:
$ aws ec2 describe-instances
This should complete without error. Note that this will operate against the default AWS region, which can be configured by setting the AWS_DEFAULT_REGION environment variable.
We will also need a copy of kubectl available locally. We should already have this from the previous recipes in this chapter involving Kubernetes.
Finally, we are going to be using a tool called kops to install our Kubernetes cluster. The kops command is an open source Kubernetes cluster management tool that can be found at https://github.com/kubernetes/kops. Instructions on how to install kops are available at https://kubernetes.io/docs/getting-started-guides/kops/.
Kops is short for Kubernetes Operations. Kops allows us to deploy and manage multiple Kubernetes clusters and supports a variety of cloud platforms.
Once we have kops installed, we are good to go!