Puppet

Purpose: Puppet is another declarative configuration management tool that excels in system configuration and enforcement. Like Ansible, it's designed to maintain the desired state of a system's configuration but uses a client-server (master-agent) architecture by default.

Architectural Overview¶

Detailed¶

sequenceDiagram
    participant Gitea as Gitea Repo (Puppet Environment)
    participant r10k as r10k (Environment Deployer)
    participant PuppetMaster as Puppet Server (lab-puppet-01.bunny-lab.io)
    participant Agent as Managed Agent (fedora.bunny-lab.io)
    participant Neofetch as Neofetch Package

    %% PuppetMaster pulling environment updates
    PuppetMaster->>Gitea: Pull Puppet Environment updates
    Gitea-->>PuppetMaster: Send latest Puppet repository code

    %% r10k deployment process
    PuppetMaster->>r10k: Deploy environment with r10k
    r10k->>PuppetMaster: Fetch and install Puppet modules
    r10k-->>PuppetMaster: Compile environments and apply updates

    %% Agent enrollment process
    Agent->>PuppetMaster: Request to enroll (Agent Check-in)
    PuppetMaster->>Agent: Verify SSL Certificate & Authenticate
    Agent-->>PuppetMaster: Send facts about system (Facter)

    %% PuppetMaster compiles catalog for the agent
    PuppetMaster->>PuppetMaster: Compile Catalog
    PuppetMaster->>PuppetMaster: Check if 'neofetch' is required in manifest
    PuppetMaster-->>Agent: Send compiled catalog with 'neofetch' installation instructions

    %% Agent installs neofetch
    Agent->>Agent: Check if 'neofetch' is installed
    Agent--xNeofetch: 'neofetch' not installed
    Agent->>Neofetch: Install 'neofetch'
    Neofetch-->>Agent: Installation complete

    %% Agent reports back to PuppetMaster
    Agent->>PuppetMaster: Report status (catalog applied and neofetch installed)

Simplified¶

sequenceDiagram
    participant Gitea as Gitea (Puppet Repository)
    participant PuppetMaster as Puppet Server
    participant Agent as Managed Agent (fedora.bunny-lab.io)
    participant Neofetch as Neofetch Package

    %% PuppetMaster pulling environment updates
    PuppetMaster->>Gitea: Pull environment updates
    Gitea-->>PuppetMaster: Send updated code

    %% Agent enrollment and catalog request
    Agent->>PuppetMaster: Request catalog (Check-in)
    PuppetMaster->>Agent: Send compiled catalog (neofetch required)

    %% Agent installs neofetch
    Agent->>Neofetch: Install neofetch
    Neofetch-->>Agent: Installation complete

    %% Agent reports back
    Agent->>PuppetMaster: Report catalog applied (neofetch installed)

Breakdown¶

1. PuppetMaster Pulls Updates from Gitea¶

• PuppetMaster uses r10k to fetch the latest environment updates from Gitea. These updates include manifests, hiera data, and modules for the specified Puppet environments.

2. PuppetMaster Compiles Catalogs and Modules¶

• After pulling updates, the PuppetMaster compiles the latest node-specific catalogs based on the manifests and modules. It ensures the configuration is ready for agents to retrieve.

3. Agent (fedora.bunny-lab.io) Checks In¶

• The Puppet agent on fedora.bunny-lab.io checks in with the PuppetMaster for its catalog. This request tells the PuppetMaster to compile the node's desired configuration.

4. Agent Downloads and Applies the Catalog¶

• The agent retrieves its compiled catalog from the PuppetMaster. It compares the current system state with the desired state outlined in the catalog.

5. Agent Installs neofetch¶

• The agent identifies that neofetch is missing and installs it using the system's package manager. The installation follows the directives in the catalog.

6. Agent Reports Success¶

• Once changes are applied, the agent sends a report back to the PuppetMaster. The report includes details of the changes made, confirming neofetch was installed.

Deployment Steps:¶

You will need to perform a few steps outlined in the official Puppet documentation to get a Puppet server operational. A summarized workflow is seen below:

Install Puppet Repository¶

Installation Scope: Puppet Server / Managed Devices

# Add Puppet Repository / Enable Puppet on YUM
sudo rpm -Uvh https://yum.puppet.com/puppet7-release-el-9.noarch.rpm

Install Puppet Server¶

Installation Scope: Puppet Server

# Install the Puppet Server
sudo yum install -y puppetserver
systemctl enable --now puppetserver

# Validate Successful Deployment
exec bash
puppetserver -v

Install Puppet Agent¶

Installation Scope: Puppet Server / Managed Devices

# Install Puppet Agent (This will already be installed on the Puppet Server)
sudo yum install -y puppet-agent

# Enable the Puppet Agent
sudo /opt/puppetlabs/bin/puppet resource service puppet ensure=running enable=true

# Configure Puppet Server to Connect To
puppet config set server lab-puppet-01.bunny-lab.io --section main

# Establish Secure Connection to Puppet Server
puppet ssl bootstrap

# ((On the Puppet Server))
# You will see an error stating: "Couldn't fetch certificate from CA server; you might still need to sign this agent's certificate (fedora.bunny-lab.io)."
# Run the following command (as root) on the Puppet Server to generate a certificate
sudo su
puppetserver ca sign --certname fedora.bunny-lab.io

Validate Agent Functionality¶

At this point, you want to ensure that the device being managed by the agent is able to pull down configurations from the Puppet Server. You will know if it worked by getting a message similar to Notice: Applied catalog in X.XX seconds after running the following command:

puppet agent --test

Install r10k¶

At this point, we need to configure Gitea as the storage repository for the Puppet "Environments" (e.g. Production and Development). We can do this by leveraging a tool called "r10k" which pulls a Git repository and configures it as the environment in Puppet.

# Install r10k Pre-Requisites
sudo dnf install -y ruby ruby-devel gcc make

# Install r10k Gem (The Software)
# Note: If you encounter any issues with permissions, you can install the gem with "sudo gem install r10k --no-document".
sudo gem install r10k

# Verify the Installation (Run this as a non-root user)
r10k version

Configure r10k¶

# Create the r10k Configuration Directory
sudo mkdir -p /etc/puppetlabs/r10k

# Create the r10k Configuration File
sudo nano /etc/puppetlabs/r10k/r10k.yaml

---
# Cache directory for r10k
cachedir: '/var/cache/r10k'

# Sources define which repositories contain environments (Be sure to use the SSH URL, not the Git URL)
sources:
  puppet:
    remote: 'https://git.bunny-lab.io/GitOps/Puppet.git'
    basedir: '/etc/puppetlabs/code/environments'

# Lockdown the Permissions of the Configuration File
sudo chmod 600 /etc/puppetlabs/r10k/r10k.yaml

# Create r10k Cache Directory
sudo mkdir -p /var/cache/r10k
sudo chown -R puppet:puppet /var/cache/r10k

Configure Gitea¶

At this point, we need to set up the branches and file/folder structure of the Puppet repository on Gitea.

You will make a repository on Gitea with the following files and structure as noted by each file's title. You will make a mirror copy of all of the files below in both the Production and Development branches of the repository. For the sake of this example, the repository will be located at https://git.bunny-lab.io/GitOps/Puppet.git

# Node definition for the Fedora agent
node 'fedora.bunny-lab.io' {
  # Include the neofetch class to ensure Neofetch is installed
  include neofetch
}

forge 'https://forge.puppet.com'
mod 'puppetlabs-stdlib', '9.6.0'
mod 'puppetlabs-puppetdb', '8.1.0'
mod 'puppetlabs-postgresql', '10.3.0'
mod 'puppetlabs-firewall', '8.1.0'
mod 'puppetlabs-inifile', '6.1.1'
mod 'puppetlabs-concat', '9.0.2'
mod 'puppet-systemd', '7.1.0'

# Specifies the module path for this environment
modulepath = modules:$basemodulepath

# Optional: Specifies the manifest file for this environment
manifest = manifests/site.pp

# Optional: Set the environment's config_version (e.g., a script to output the current Git commit hash)
# config_version = scripts/config_version.sh

# Optional: Set the environment's environment_timeout
# environment_timeout = 0

# Node definition for the Fedora agent
node 'fedora.bunny-lab.io' {
  # Include the neofetch class to ensure Neofetch is installed
  include neofetch
}

# Node definition for the Puppet Server
node 'lab-puppet-01.bunny-lab.io' {

  # Include the puppetdb class with custom parameters
  class { 'puppetdb':
    listen_address => '0.0.0.0', # Allows access from all network interfaces
  }

  # Configure the Puppet Server to use PuppetDB
  include puppetdb
  include puppetdb::master::config

  # Ensure the required iptables rules are in place using Puppet's firewall resources
  firewall { '100 allow Puppet traffic on 8140':
    proto  => 'tcp',
    dport  => '8140',
    jump   => 'accept',  # Corrected parameter from action to jump
    chain  => 'INPUT',
    ensure => 'present',
  }

  firewall { '101 allow PuppetDB traffic on 8081':
    proto  => 'tcp',
    dport  => '8081',
    jump   => 'accept',  # Corrected parameter from action to jump
    chain  => 'INPUT',
    ensure => 'present',
  }
}

# Node definition for the Fedora agent
node 'fedora.bunny-lab.io' {
  # Include the neofetch class to ensure Neofetch is installed
  include neofetch
}

# Default node definition (optional)
node default {
  # This can be left empty or include common classes for all other nodes
}

class neofetch {
  package { 'neofetch':
    ensure => installed,
  }
}

Storing Credentials to Gitea¶

We need to be able to pull down the data from Gitea's Puppet repository under the root user so that r10k can automatically pull down any changes made to the Puppet environments (e.g. Production and Development). Each Git branch represents a different Puppet environment. We will use an application token to do this.

Navigate to "Gitea > User (Top-Right) > Settings > Applications - Token Name: Puppet r10k - Permissions: Repository > Read Only - Click the "**Generate Token" button to finish.

Now we want to configure Gitea to store the credentials for later use by r10k:

# Enable Stored Credentials (We will address security concerns further down...)
sudo yum install -y git
sudo git config --global credential.helper store

# Clone the Git Repository Once to Store the Credentials (Use the Application Token as the password)
# Username: nicole.rappe
# Password: <Application Token Value>
sudo git clone https://git.bunny-lab.io/GitOps/Puppet.git /tmp/PuppetTest

# Verify the Credentials are Stored
sudo cat /root/.git-credentials

# Lockdown Permissions
sudo chmod 600 /root/.git-credentials

# Cleanup After Ourselves
sudo rm -rf /tmp/PuppetTest

Finally we validate that everything is working by pulling down the Puppet environments using r10k on the Puppet Server:

# Deploy Puppy Environments from Gitea
sudo /usr/local/bin/r10k deploy environment -p

# Validate r10k is Installing Modules in the Environments
sudo ls /etc/puppetlabs/code/environments/production/modules
sudo ls /etc/puppetlabs/code/environments/development/modules

!!! success "Successful Puppet Environment Deployment If you got no errors about Puppetfile formatting or Gitea permissions errors, then you are good to move onto the next step.

External Node Classifier (ENC)¶

An ENC allows you to define node-specific data, including the environment, on the Puppet Server. The agent requests its configuration, and the Puppet Server provides the environment and classes to apply.

Advantages:

• Centralized Control: Environments and classifications are managed from the server.
• Security: Agents cannot override their assigned environment.
• Scalability: Suitable for managing environments for hundreds or thousands of nodes.

Create an ENC Script¶

sudo mkdir -p /opt/puppetlabs/server/data/puppetserver/scripts/

#!/usr/bin/env ruby
# enc.rb

require 'yaml'

node_name = ARGV[0]

# Define environment assignments
node_environments = {
  'fedora.bunny-lab.io' => 'development',
  # Add more nodes and their environments as needed
}

environment = node_environments[node_name] || 'production'

# Define classes to include per node (optional)
node_classes = {
  'fedora.bunny-lab.io' => ['neofetch'],
  # Add more nodes and their classes as needed
}

classes = node_classes[node_name] || []

# Output the YAML document
output = {
  'environment' => environment,
  'classes' => classes
}

puts output.to_yaml

# Ensure the File is Executable
sudo chmod +x /opt/puppetlabs/server/data/puppetserver/scripts/enc.rb

Configure Puppet Server to Use the ENC¶

Edit the Puppet Server's puppet.conf and set the node_terminus and external_nodes parameters:

[master]
node_terminus = exec
external_nodes = /opt/puppetlabs/server/data/puppetserver/scripts/enc.rb

Restart the Puppet Service

sudo systemctl restart puppetserver

Pull Puppet Environments from Gitea¶

At this point, we can tell r10k to pull down the Puppet environments (e.g. Production and Development) that we made in the Gitea repository in previous steps. Run the following command on the Puppet Server to pull down the environments. This will download / configure any Puppet Forge modules as well as any hand-made modules such as Neofetch.

sudo /usr/local/bin/r10k deploy environment -p
# OPTIONAL: You can pull down a specific environment instead of all environments if you specify the branch name, seen here:
#sudo /usr/local/bin/r10k deploy environment development -p

Apply Configuration to Puppet Server¶

At this point, we are going to deploy the configuration from Gitea to the Puppet Server itself so it installs PuppetDB automatically as well as configures firewall ports and other small things to functional properly. Once this is completed, you can add additional agents / managed devices and they will be able to communicate with the Puppet Server over the network.

sudo /opt/puppetlabs/bin/puppet agent -t