Automated deployment of the OSS infrastructure

*Kubernetes deployment based on the roles, taken from Kargo project

Requirements for deployment node

1. CentOS 7 / Ubuntu 16.04
2. Packages installed: python-pip, python-netaddr, Junja2 >=2.8

Requirements for target nodes

1. CentOS7 nodes, installed from iso, which was builded by OSS image builder
2. Access to nodes via ssh key for root user
3. Proper resolving by hostname on the machine, which you are using for deployment
4. Ansible, docker-py

How to deploy

1. Generate inventory file (minimum required nodes - 3):

utils/inventory-generator --nodes node1 node2 node3

2. Run bootstrap playbook for loading required packages (required internet connection)

ANSIBLE_HOST_KEY_CHECKING=False ansible-playbook -i inventory/bootstrap.cfg bootstrap-runner.yml

3. Run playbook (doesn't require internet connection)

ANSIBLE_HOST_KEY_CHECKING=False ansible-playbook -i inventory/automation.cfg automation-runner.yml

Overriding default options

You can override some default settings. For example:

• build_oss_images=false - skip building of oss images on CCP
• deploy_oss_images=false - skip deployment of oss images on CCP
• keepalived_ip=10.1.1.8 - specify virtual ip for docker private registry (it is generated automatically by default)

To apply this options - execute ansible-playbook with --extra-vars:

ANSIBLE_HOST_KEY_CHECKING=False ansible-playbook -i inventory/automation.cfg automation-runner.yml --extra-vars "keepalived_ip=10.1.1.8 build_oss_images=false"