DevOps Classroom Series – 17/Nov/2019

Ansible Parallelism

• Parallelism can be set in Ansible using an argument –fork.
• Default value is 5
• Scenario:
  • 100 nodes are in inventory and I run ansible-playbook test.yml, this execution happens in 20 batches
  • 100 nodes are in inventory and I run ansible-playbook -f 50 test.yml, this execution happens in 2 batches
  • 100 nodes are in inventory and I run ansible-playbook -f 500 test.yml, this execution happens in 1 batch

Ansible custom facts

• I have a playbook, inside this playbook I need to decide on some module execution depending on other module execution.
• Eg: only when you create a file, then only restart service and do further executions
• In any case if you want to store the value only till your playbook execution is completed, you can use a module called as set_fact
• For more info Refer Here

How to make non idempotent modules idempotent

• To make your non idempotent modules idempotent create some flags
• One easier implementation is create and check for file existence
• You can also try environmental variables

Register Variables

• Refer Here

---
- hosts: all
  become: yes
  tasks:
    - name: contents of file
      command: cat /home/ansible/Readme.txt
      register: readme_output
    - name: display contents
      debug:
        var: readme_output

Vault

• Configuration Management has to deal with passwords or sensitive contents
• We need a mechanism to encrypt the sensitive content and thats exactly what ansible-vault does.
• Refer Here
• Execution in ansible requires vault password or password file to passed

ansible-playbook -i hosts  --vault-password-file /var/lib/password  test.yaml

Ansible Tower

• Tower provides authentication and reporting features to ansible control server

• Tower is web based platform for doing what you have done with ACS (Command line), we would have an user interface and REST API.

• Terms

  • Projects => Where your ansible playbooks are?
  • Inventory
  • Job
  • Schedule Jobs
  • Job Template

• In this series we will use Ansible Trail Version on AWS

Dynamic Inventory

• If the node list is changing consistently, then how to handle this in inventory.
• Dynamic Inventory can be used.
• Dynamic Inventory is any script in any language which returns information about your nodes in json format specified by Ansible.
• Json structure should look like

{
    "group001": {
        "hosts": ["host001", "host002"],
        "vars": {
            "var1": true
        },
        "children": ["group002"]
    },
    "group002": {
        "hosts": ["host003","host004"],
        "vars": {
            "var2": 500
        },
        "children":[]
    }

}

• Refer Here for sample AWS scenario.

DevOps Scenarios in Ansible

• Consider the below Architecture
• To create this deployment, we can go with three roles
  • tomcat => tomcat on servers
  • mysql => mysql on server
  • redis => redis on server
• Then you write your playbook => contoso.yml

---
- hosts: appservers
  become: yes
  roles:
    - role: tomcat_deploy
- hosts: dbservers
  become: yes
  roles:
    - role: mysql_deploy
- hosts: cacheservers
  become: yes
  roles
    - role: redis_deploy

Realizing this with CI/CD

• Servers are already available with OS installed.
• Create a inventory with servers
• Create a Jenkins Pipeline/Freestyle Project with following steps
  • Get latest code from GIT
  • Build the code
  • Call ansible-playbook to deploy the configuration

Realizing this with CI/CD and Infra Provisioning

• Starting Point : Servers are not available
• Create a Jenkins Pipeline/Freestyle Project with following steps
  • Get latest code from GIT
  • Build the code
  • Call Terraform to create servers and then terraform calls ansible to do configuration management

Exercise

• Write ansible-playbook to deploy any of the sample application from here
• Try to make java -jar spring-petclinic.jar as a Linux service Refer Here