DevOps Classroom Series – 12/Sept/2020

Chef Infra Server Components

• Infra Components Refer Here

Creating Lab Environment

• In this scenario
  • PROD: will be Ubuntu 18 machines from aws
  • UAT: Will be Ubuntu 18 machines from aws
  • QA: Will be Ubuntu 18 machine from azure
• Upload qthms-app and qthms-db to the chef server

• Now lets bootstrap manually all the nodes from workstation to nodes
• Lets bootstrap chef server in the qa, uat and prod environments
• Lets try to come up with run_list for every node

QA : qthms-app::default,qthms-db::default

uat
	app: qthms-app::default
  
	db: qthms-db::default
  
prod:
	app: qthms-app::default
	
	db: qthms-db::default

• In a large kind of deployment, managing run_lists is not effecient by add recipes to every node.
• To manage this lets create two roles
  1. appserver
  2. dbserver
• Roles can be created from chef-manage.
• But it is a good practice to have version control of your roles. so lets create roles files in chef-repo
• For Roles Official docs Refer Here
• Lets create two roles in the roles folder with the following content for
  1. appserver.rb

  name  'appserver'
  description 'This is qthms appserver'
  run_list 'qthms-app::default'
  

  2. dbserver.rb

  name  'dbserver'
  description 'This is qthms  dbserver'
  run_list 'qthms-db::default'
  

• Now lets upload the roles to the chef server

knife role from file <path-to-rb>

• Now lets edit run lists to have the roles in the run_list rather than recipes
• Now the convergence time is 30 minutes for all the servers. Now lets learn how to change that
• Client Configurations can be changed from client.rb file in the chef repo refer Here or easier way is to add the following cookbook to the run_list refer Here

# first approach
Chef::Config[:interval] = <seconds>

• To do common configuration lets create a cookbook qt-general
• In this cookbook lets add dependency of chef-client and set the attribute

default['chef_client']['interval']= '1800'

• Now we need to add this recipe to all the nodes, we have roles, go and change the roles
• For app servers the interval should be ‘3600’ seconds for db servers the convergence should be ‘7200’
• The appserver.rb (role)

name  'appserver'
description 'This is qthms appserver'
run_list 'qt-general::default', 'qthms-app::default'
default_attributes(
    'chef_client' => {
        'interval' => '3600'
    }
)

• Role db server

name  'dbserver'
description 'This is qthms dbserver'
run_list 'qt-general::default','qthms-db::default'
default_attributes(
    'chef_client' => {
        'interval' => '7200'
    }
)

• Now we need to add a new version of the cookbook to include java ’11’ and also install apache server
• Now make changes and after successful testing of this cookbook if you upload changes it will be applied to all servers in appserver role which is not what we want.
• We want this changes to be applied only to QA, SO this is where environments come into play. \
• Like Roles environments can be created from manage, inside chef-repo Refer Here
• By default chef has a environment already called as _default
• Now lets create 3 environment files in environments folder in chef repo
  • qa.rb
  • uat.rb
  • prod.rb
• Refer to git for these files
• Lets upload the environments

knife environment from file <path-to-env>

• Lets apply environments to the nodes

• Now upload the cookbook

• Verify the QA Environment

• Verify in other Environment

• Now we have multiple places where we define attributes

  • recipe
  • attribute file
  • roles
  • environments

• Lets understand what gets applied.

• Chef has an extra layer which is attribute types

  • default
  • force_default
  • normal
  • override
  • force_override
  • automatic

• We have 4 different places and six different types where attributes can be defined

• Chef follows the following table for precedence

• Refer Here for offical docs

• Note: GitHub Link Refer Here

Next Steps

• How to deal sensitive Content in chef (data bags)
• How to avoid manual bootstrapping
• What is Chef Automate
• How to use chef for managing infra in AWS (OpsWorks) & Azure (VM Extensions)
• Exercises