okd/README_AWS.md

AWS Setup Instructions

Get AWS API credentials

1. AWS credentials documentation

Create a credentials file

1. Create a credentials file (eg ~/.aws_creds) that looks something like this (variables must have have these exact names). export AWS_ACCESS_KEY_ID='AKIASTUFF' export AWS_SECRET_ACCESS_KEY='STUFF'
2. source this file source ~/.aws_creds Note: You must source this file before running any Ansible commands.

Alternatively, you could configure credentials in either ~/.boto or ~/.aws/credentials, see the boto docs for the format.

(Optional) Setup your $HOME/.ssh/config file

In case of a cluster creation, or any other case where you don't know the machine hostname in advance, you can use .ssh/config to setup a private key file to allow ansible to connect to the created hosts.

To do so, add the the following entry to your $HOME/.ssh/config file and make it point to the private key file which allows you to login on AWS.

Host *.compute-1.amazonaws.com
  PrivateKey $HOME/.ssh/my_private_key.pem

Alternatively, you can configure your ssh-agent to hold the credentials to connect to your AWS instances.

(Optional) Choose where the cluster will be launched

By default, a cluster is launched with the following configuration:

• Instance type: m3.large
• AMI: ami-307b3658 (for online deployments, ami-acd999c4 for origin deployments and ami-10663b78 for enterprise deployments)
• Region: us-east-1
• Keypair name: libra
• Security group: public

Master specific defaults:

• Master root volume size: 10 (in GiBs)
• Master root volume type: gp2
• Master root volume iops: 500 (only applicable when volume type is io1)

Node specific defaults:

• Node root volume size: 10 (in GiBs)
• Node root volume type: gp2
• Node root volume iops: 500 (only applicable when volume type is io1)
• Docker volume size: 25 (in GiBs)
• Docker volume ephemeral: true (Whether the docker volume is ephemeral)
• Docker volume type: gp2 (only applicable if ephemeral is false)
• Docker volume iops: 500 (only applicable when volume type is io1)

If needed, these values can be changed by setting environment variables on your system.

• export ec2_instance_type='m3.large'
• export ec2_image='ami-307b3658'
• export ec2_region='us-east-1'
• export ec2_keypair='libra'
• export ec2_security_groups="['public']"
• export ec2_vpc_subnet='my_vpc_subnet'
• export ec2_assign_public_ip='true'
• export os_etcd_root_vol_size='20'
• export os_etcd_root_vol_type='standard'
• export os_etcd_vol_size='20'
• export os_etcd_vol_type='standard'
• export os_master_root_vol_size='20'
• export os_master_root_vol_type='standard'
• export os_node_root_vol_size='15'
• export os_docker_vol_size='50'
• export os_docker_vol_ephemeral='false'

Install Dependencies

1. Ansible requires python-boto for aws operations: RHEL/CentOS/Fedora yum install -y ansible python-boto OSX: pip install -U boto

Test The Setup

1. cd openshift-ansible
2. Try to list all instances (Passing an empty string as the cluster_id argument will result in all ec2 instances being listed) bin/cluster list aws ''

Creating a cluster

1. To create a cluster with one master and two nodes bin/cluster create aws <cluster-id>

Updating a cluster

1. To update the cluster bin/cluster update aws <cluster-id>

Terminating a cluster

1. To terminate the cluster bin/cluster terminate aws <cluster-id>

Specifying a deployment type

The --deployment-type flag can be passed to bin/cluster to specify the deployment type

1. To launch an online cluster (requires access to private repositories and amis): bin/cluster create aws --deployment-type=online <cluster-id> Note: If no deployment type is specified, then the default is origin.