100% found this document useful (2 votes)

584 views

Apress Ansible For Kubernetes by Example

Uploaded by

anton

Available Formats

Download as PDF, TXT or read online on Scribd

100% found this document useful (2 votes)

584 views

Apress Ansible For Kubernetes by Example

Uploaded by

anton

Available Formats

Download as PDF, TXT or read online on Scribd

You are on page 1/ 300

Ansible for

Kubernetes by
Example
Automate Your Kubernetes Cluster
with Ansible
—
Luca Berton
Ansible for Kubernetes by
Example
Automate Your Kubernetes Cluster
with Ansible

Luca Berton
Ansible for Kubernetes by Example: Automate Your Kubernetes Cluster with Ansible
Luca Berton
BOURNEMOUTH, UK

ISBN-13 (pbk): 978-1-4842-9284-6 ISBN-13 (electronic): 978-1-4842-9285-3

https://doi.org/10.1007/978-1-4842-9285-3

Copyright © 2023 by Luca Berton

This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part of the
material is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation,
broadcasting, reproduction on microfilms or in any other physical way, and transmission or information
storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology now
known or hereafter developed.
Trademarked names, logos, and images may appear in this book. Rather than use a trademark symbol with
every occurrence of a trademarked name, logo, or image we use the names, logos, and images only in an
editorial fashion and to the benefit of the trademark owner, with no intention of infringement of the
trademark.
The use in this publication of trade names, trademarks, service marks, and similar terms, even if they are not
identified as such, is not to be taken as an expression of opinion as to whether or not they are subject to
proprietary rights.
While the advice and information in this book are believed to be true and accurate at the date of publication,
neither the authors nor the editors nor the publisher can accept any legal responsibility for any errors or
omissions that may be made. The publisher makes no warranty, express or implied, with respect to the
material contained herein.
Managing Director, Apress Media LLC: Welmoed Spahr
Acquisitions Editor: Divya Modi
Development Editor: James Markham
Coordinating Editor: Divya Modi
Copy Editor: Kezia Endsley
Cover designed by eStudioCalamar
Cover image designed by Freepik (www.freepik.com)
Distributed to the book trade worldwide by Springer Science+Business Media New York, 1 New York Plaza,
Suite 4600, New York, NY 10004-1562, USA. Phone 1-800-SPRINGER, fax (201) 348-4505, e-mail orders-ny@
springer-sbm.com, or visit www.springeronline.com. Apress Media, LLC is a California LLC and the sole
member (owner) is Springer Science + Business Media Finance Inc (SSBM Finance Inc). SSBM Finance Inc
is a Delaware corporation.
For information on translations, please e-mail [email protected]; for reprint,
paperback, or audio rights, please e-mail [email protected].
Apress titles may be purchased in bulk for academic, corporate, or promotional use. eBook versions and
licenses are also available for most titles. For more information, reference our Print and eBook Bulk Sales
web page at http://www.apress.com/bulk-sales.
Any source code or other supplementary material referenced by the author in this book is available to
readers on GitHub (github.com/Apress/Ansible-for-Kubernetes-by-Example-by-Luca-Berton). For more
detailed information, please visit http://www.apress.com/source-code.
Printed on acid-free paper
For my son Filippo, the joy of my life.
Table of Contents
About the Author�� xi

About the Technical Reviewer�� xiii

Acknowledgments��xv

Introduction��xvii

Chapter 1: Modern IT Infrastructure and Hello App�� 1

Modern IT Infrastructure (DevOps and IaC)�� 1
The Move to Containers�� 4
Ansible by Red Hat�� 7
The Cloud Native Computing Foundation�� 8
Kubernetes Support�� 9
Kubernetes Distributions: OpenShift, Rancher, EKS, AKS, and GCP�� 9
Containers and Pods�� 11
Creating a Hello App�� 12
Linux Base Images�� 13
Enterprise Linux-Based Images�� 15
Container Security�� 17
The Hello Dockerfile�� 18
The Hello Application�� 19
Building the Hello App�� 22
Running Hello in Docker�� 22
Deploying Hello in Kubernetes�� 24
Deploying Hello in Operator�� 25
Key Takeaways�� 25

v
Table of Contents

Chapter 2: Ansible Language Code�� 27

What Is Ansible?�� 27
Provisioning�� 28
Configuration Management�� 29
Application Deployment�� 29
Ansible Open-Source vs Commercial Options�� 29
Ansible’s Architecture�� 30
UNIX Target Node�� 31
Windows Target Node�� 32
Ansible Installation�� 34
Getting Started with Ansible�� 36
Running Your First Ansible Ad Hoc Command�� 36
Creating a Basic Inventory�� 37
Ansible Code Language�� 37
Ansible Inventory�� 38
Ansible Playbook�� 40
Ansible Roles�� 51
Ansible Collection�� 54
Ansible Execution Environment�� 55
Key Takeaways�� 61

Chapter 3: Ansible for Containers�� 63

Ansible for Containers�� 63
Install Docker on Linux and Windows�� 63
Install Docker in Debian Linux�� 65
Install Docker in Red Hat Linux�� 67
Install Docker on Windows�� 70
Flatpak in Linux�� 71
Snap in Linux�� 74
Deploy a Web Server in a Container�� 77
Apache with Docker for Debian-like Systems�� 78
Apache with Podman for Red Hat-like Systems�� 80

vi
Table of Contents

Use Vagrant and Packer�� 82

Vagrant�� 83
Packer�� 84
Key Takeaways�� 85

Chapter 4: Ansible for K8s Tasks�� 87

Kubernetes Objects�� 90
Control Plane vs Data Plane�� 91
kubectl�� 92
GitOps Continuous Deployment�� 93
Jenkins�� 94
VMWare Tanzu Application Platform�� 95
Set Up Your Laboratory�� 95
Virtual Machines�� 96
Raspberry Pis�� 97
Kubespray�� 99
OpenShift Local�� 100
hetzner-ocp4�� 103
Create a Cluster with Minikube�� 104
Kubeadm�� 106
K3s Lightweight Kubernetes�� 107
Kubernetes Upgrade�� 107
Create a Cluster with kOps�� 109
Configure Ansible for Kubernetes�� 109
Ansible Troubleshooting�� 113
401 unauthorized�� 113
Kubernetes�� 114
OpenShift�� 116
x509 error�� 118
kubeconfig�� 118
Configure a Python Virtual Environment�� 120

vii
Table of Contents

Configure an Ansible Execution Environment�� 123

Create a Namespace�� 126
Report Namespaces�� 129
Report Deployments in Namespace�� 131
Create a Pod�� 136
Create a Secret�� 139
Use a Service to Expose Your App�� 141
Kubernetes Networking�� 143
Scale Your App�� 148
Auto-scaling�� 153
Update Your App�� 154
Assign Resources to Kubernetes K8s Pods�� 157
Metrics�� 158
CPU Resources�� 158
Memory Resources�� 161
Namespace Resources�� 162
GPU Resources�� 163
Configure a Pod to Use a Volume for Storage�� 163
Apply Multiple YAML Files at Once on Kubernetes�� 166
Key Takeaways�� 167

Chapter 5: Ansible for K8s Data Plane�� 169

Configuring a Java Microservice�� 171
The Demo Java Web Application�� 172
Stateless: Deploying PHP Guestbook Application with Redis�� 180
Kustomize: Do More with Less�� 184
Stateful: Deploying WordPress and MySQL with Persistent Volumes�� 186
Security Namespace (Pod Security Admission)�� 191
Security Pod Resources (AppArmor)�� 192
Security Pod Syscalls (seccomp)�� 193

viii
Table of Contents

Ansible Dynamic Inventory�� 194

Key Takeaways�� 199

Chapter 6: Ansible for K8s Management�� 201

The Helm Package Manager�� 202
Helm Repositories�� 204
Helm Packages�� 207
Helm Plugins�� 212
Deploy a Monitoring Tool�� 216
kube-prometheus�� 217
Ansible Collections�� 218
Helm Chart�� 219
Fetch Logs from Resources�� 220
Apply a JSON Patch Operation�� 222
Copy Files and Directories to and from a Pod�� 224
Manage Services on Kubernetes�� 225
Taint Nodes�� 227
Drain, Cordon, or Uncordon Nodes�� 229
Kubernetes Dynamic Inventory�� 232
Roll Back Deployments and DaemonSets�� 232
Set a New Size for a Deployment, ReplicaSet, Replication Controller, or Job�� 233
Security�� 233
AAA�� 235
OpenID Identity Provider�� 236
Calico�� 237
Key Takeaways�� 237

Chapter 7: Ansible for Kubernetes Cloud Providers�� 239

Cloud Architecture�� 241
Amazon Web Services (AWS)�� 244
Google Cloud Platform (GCP)�� 251

ix
Table of Contents

Microsoft Azure Cloud Services�� 255

Other Vendors�� 259
Key Takeaways�� 260

Chapter 8: Ansible for Enterprise�� 261

The Ansible Automation Platform�� 261
Event-Driven Ansible�� 268
IT Trends�� 272
Ansible Trusted Code�� 273
What’s Next?�� 275
Thank You�� 276

Index�� 277

x
About the Author
Luca Berton is an Ansible automation expert who works for
JPMorgan Chase & Co. and previously worked with the Red
Hat Ansible Engineer Team for three years. He is the creator
of the Ansible Pilot project. With more than 15 years of
experience as a system administrator, he has strong expertise
in infrastructure hardening and automation. An enthusiast
of open-source, Luca supports the community by sharing his
knowledge at different public access events. He is a geek by
nature, Linux by choice, and Fedora, of course.

xi
About the Technical Reviewer
Nikhil Jain is an Ansible expert with over 12 years of DevOps
experience. He has been using Ansible and contributing to
it from its inception. He currently works closely with Ansible
engineering.
He is an open-source enthusiast and is part of the
Ansible Pune Meetup Organizing team. He has presented
multiple Ansible sessions at various global and local events.
Apart from sitting in front of his computer automating things
using Ansible, he loves watching sports and is a regular part
of the local cricket team.

xiii
Acknowledgments
To my son, family, and friends, who make life worth living and whose support and
encouragement make this work possible.
To all whom I’ve worked with over the years and shared any ideas for this book:
Thank you for the knowledge you’ve shared.

xv
Introduction
This book guides you through the process of automating a Kubernetes cluster using the
Ansible open-source technology.
If you are an IT professional who wants a jargon-free explanation of the Ansible
technology, this book is for you. This includes Kubernetes, Linux, and Windows systems
administrators, DevOps professionals, thought leaders, and infrastructure-as-code
enthusiasts, as well as information technology team members providing leadership to
their businesses.
Infrastructure and operations (I&O) leaders look at the Ansible platform to
implement Infrastructure as Code (IaC), Configuration as a Code (CaC), Policy as a Code
(PaC), code pipelines, orchestration (K8s), DevSecOps, self-healing infrastructures, and
event-driven automation methodologies.
This book is a powerful resource for computer engineers and leaders who believe
that innovation, automation, and acceleration are the drivers of the booming business of
tomorrow.
A successful infrastructure is a matter of improvements that are made little by little
and good habits that you develop using automation during the journey.
In this book, you learn how to become more productive and effective using the
Ansible open-source automation technology to deploy world-scale, cloud-native
applications and services. Containers have revolutionized the IT industry, but it takes
more than container runtime engines like Docker and Podman to handle these modern,
huge workloads. This is where Kubernetes come in to play.
Kubernetes applications are designed as microservices, packaged in containers,
and deployed in a hybrid cloud in multiple cloud platforms worldwide to serve your
customers.
It’s also possible to upgrade applications and services without impacting user
performance in many ways: by rolling, canary, and blue/green environments.
The scriptable infrastructure of the cloud providers enables you to elastically meet
the demands of traffic with a virtually unlimited compute powerhouse. Engineers have a
great deal of power and responsibility for their business’s success.

xvii
Introduction

What Is In This Book?

This book provides in-depth content of the following topics:

• The Ansible code language for beginners and experienced users, with
examples

• Ansible installation on the latest releases of Ansible

• Ansible for container management (Docker and Podman)

• Ansible for Kubernetes infrastructure examples

• Troubleshooting common errors

Development Environment
You do not need a specific IDE to benefit from this book. You simply need a base
environment consisting of the following:

• A text editor: Terminal-based (VIM, Emacs, Nano, Pico, and so on) or

GUI-based (VS Code, Atom, Geany, and so on)

• A workstation with the ansible or ansible-core packages installed

• Kubernetes or OpenShift cluster

Conventions Used in the Book

This is a practical book, with a lot of examples and terminal commands. There are also
some Ansible language code samples.
The Ansible language code used in the Ansible Playbook examples mostly uses the
YAML and INI formats.
Commands and code samples throughout the book are either inline (for example,
ansible [command]) or in a code block (with or without line numbers), such as this:

---
# YAML file example

The Ansible language is written in the YAML format, which is a human-readable,

data-serialization language that is extremely popular for configuration files. The code
xviii
Introduction

follows the latest YAML 1.2.2 specification. It uses Python-style indentation and a more
compact format for lists and dictionary statements. It’s very close to JSON and can be
used as an XML alternative. The YAML code was validated using the YAMLlint popular
validator and tested in the most commonly used Ansible versions in the market.
The INI format used in Ansible inventories examples is a well-known format for
configuration files since the MS-DOS operating system and uses key-value pairs for
properties.
The terminal commands use the standard POSIX conventions and are ready for use
in UNIX-like systems such as Linux, macOS, or BSD. Each command is assumed to be
used by a standard user account when the prefix is the $ symbol (dollar) or by the root
user when the prefix is the # symbol (number sign).
You are going to find this code in installation, code execution, and troubleshooting
examples. The commands were tested in the most modern operating system versions
available.
The Ansible code included in this book was tested by the author and the technical
reviewer in a wide variety of modern systems and uses the Ansible best practices
regarding playbooks, roles, and collections. It was verified using the latest release of the
Ansible Linter.
Some code may intentionally break a specific Ansible best practice rule only to
demonstrate a troubleshooting session and reproduce a fatal error in a specific use case.

Chapters at a Glance
This book aims to guide you through your journey of automating the Ansible platform
for the Kubernetes (K8s) infrastructure. There are eight chapters, and the book is full of
code samples and terminal commands to verify the results.

• Chapter 1 explores how modern IT infrastructure is organized inside

an organization and outside to its customers and partners. You are
going to learn about the history of Kubernetes, from the initial Google
project to CNCF. You also learn about its most famous flavor: Red
Hat OpenShift, Amazon EKS, and other public cloud providers. A
practical example of the Hello App written in Python language is
deployed as a container (Fedora, Ubuntu, Alpine Linux), built in
Docker and running in Kubernetes.

xix
Introduction

• Chapter 2 dives deep into the Ansible language code. It explains

the architecture, how to set up a target node, inventories, as well as
playbooks, variables, filters, conditionals, handlers, loops, magic
variables, vaults, templates, plugins, roles, and collections.

• Chapter 3 is all about containers—how to automate the installation of

Docker in your operating systems, how to use the Flatpack and Snap
technologies in Linux, and how to deploy the popular web server
Apache as a containerized application in Debian/Ubuntu. To do
that, you use Docker and Red Hat Enterprise Linux with the Podman
runtime engine.

• Chapter 4 goes deep into Kubernetes, explaining the difference

between the control plane and the data plane and how to use the
kubectl utility. You learn how to set up your laboratory using virtual
machines, Raspberry Pis, Minikube, Kubespray, Kubeadm, and
OpenShift Local. You configure Ansible to interact with Kubernetes,
and troubleshoot and fix the most popular errors. There is a
lot of Ansible playbook code to create and report namespaces,
deployments, pods, secrets, and persistent volumes. You'll apply
multiple YAML files at once to Kubernetes. You also explore how to
scale up and down your applications and run rolling updates.

• Chapter 5 explains how to build a web, RESTful Java microservice

application using Spring technology and deploy it as a containerized
image with the Tomcat web server. This chapter includes more use
cases that explain how to automate the deployment of two examples
of stateless (a PHP Guestbook application with Redis) and stateful
(WordPress and MySQL with persistent volumes) applications in
Kubernetes. You learn how to apply security to namespaces, pod
resources (AppArmor), and pod syscalls (seccomp). You also learn
how to save time using the Kustomize tool to generate Kubernetes
YAML resources. Ansible Dynamic Inventory is handy for listing all
the cluster resources that interact with Kubernetes API.
• Chapter 6 covers how to manage the control plane, which is a
more useful tool for cluster administrator use cases, and discusses
how to apply Ansible to the deployment of the DEV, SIT, UAT, and

xx
Introduction

PROD infrastructures. You learn how to deploy a monitoring tool

(Prometheus) via the Helm package manager, fetch logs from
resources for troubleshooting, and copy files and directories to
and from a pod. The chapter also covers the processes of tainting,
draining, cordoning, and uncordoning nodes. It also discusses
rolling back deployments and DaemonSets and applying Security
Authentication, Authorization, and Accounting (AAA).

• Chapter 7 is dedicated to the public could providers, including how

you can take advantage of the hybrid/multi-cloud providers, what
the shared responsibility model is, and some specific infrastructure
implementations of Amazon Web Services (AWS), Google Cloud
Platform (GCP), and Microsoft Azure Cloud Services. It also covers how
the Certified Kubernetes Conformance Program by CNCF is helping
with portability of applications and services between different clusters.

• Chapter 8 is all about Ansible for enterprise, with the completed

container-oriented Shiny Ansible Automation Platform providing
the Web-UI RESTful API and Ansible Certified Collections. You
learn about the future information technology trends of event-
driven Ansible, Ansible Trusted Code, and Ansible Trusted Ansible
Collections and Signed Ansible Projects. You also learn how artificial
intelligence can propel your automation journey.

The sky is the limit with the OpenContainer standard, Kubernetes orchestration, and
Ansible automation; you can deploy the cloud-native applications of tomorrow, taking
advantage of unlimited infrastructure capacity and worldwide coverage.

Source Code
All source code used in this book can be downloaded from github.com/Apress/
Ansible-for-Kubernetes-by-Example-by-Luca-Berton.

Disclaimer
Any opinions or personal views I express in this book are my own and not those of Red
Hat Inc. or JPMorgan Chase & Co.
xxi
Introduction

This work is not a product of JPMorgan Chase & Co. or its affiliates. Neither
JPMorgan Chase & Co. nor any of its affiliates makes any explicit or implied
representation or warranty regarding the contents of this paper and neither JPMorgan
Chase & Co. nor any of its affiliates accepts any liability in connection with this work,
including, without limitation, with respect to the completeness, accuracy, or reliability
of the information contained herein, whether the contents of the paper violates or
misappropriates the intellectual property of another, and the potential legal, compliance,
tax, or accounting effects thereof.

xxii
CHAPTER 1

Modern IT Infrastructure
and Hello App
Information technology infrastructure is evolving year after year, faster than ever.
Following Moore’s Law, servers have become more and more powerful and are able to
execute more operations in a fraction of a second. Similarly, network infrastructure is
evolving faster than ever.
The IT infrastructure is the backbone of a successful business. Companies
efficiently manage their IT real estate with powerful tools that save time. The secret to
success is a holistic cooperation between the IT Development and Operations teams.
Automate, accelerate, and innovate are the mantras of every corporation in the post-
pandemic world.
Kubernetes and Ansible are Infrastructure as Code (IaC) platforms that manage a
fleet of services and make crucial decisions promptly.

Modern IT Infrastructure (DevOps and IaC)

All IT departments worldwide face challenges in efficiently delivering top-level customer
IT services. In the same vein, it’s important to obtain the most you can from your
infrastructure (your return on investment). Just to be clear, the IT infrastructure is the
whole data center: the servers, network devices, cabling, storage, and so on. Every good
system administrator knows how important every piece in the infrastructure puzzle is.
Traditionally, infrastructure teams in large organizations or system administrators in
small ones carried the burden of manually managing the IT infrastructure, in order to
deliver the best service for their internal and external stakeholders. A large organization
typically operates multiple on-premise data centers and cloud providers.
Deploying an application in this traditional infrastructure required a combination of
operating systems and underlying system libraries, as shown in Figure 1-1.

1
© Luca Berton 2023
L. Berton, Ansible for Kubernetes by Example, https://doi.org/10.1007/978-1-4842-9285-3_1
Chapter 1 Modern IT Infrastructure and Hello App

App App App

Libraries

Operating System

Hardware
Figure 1-1. Traditional deployment

The main drawback of this architecture was that the applications were tightly
connected to the underlying operating system and the system libraries. It was sometimes
necessary to upgrade the underlying operating system just to update an application, or
it was impossible to do so, because another application used a different version of the
shared libraries.
The job of the infrastructure team included installing and updating the full lifecycle
of the software, changing configurations, and managing services on servers, virtual
machines, and nodes. The quality of the service provided could be measured using key
indicators.
Some of you are familiar with key indicators, which reveal the uptime of software
or hardware. Uptime measures the percentage of time the service is available to receive
interactions by users. An uptime of 100 percent means the service is always available.
Unfortunately, this was an unrealistic scenario because software and hardware failures
negatively affects this in real life. Customers then experience a lower quality of service.
You can apply this analysis to any web service, API service, website, web application,
hardware, server, network device, or cloud provider. Some businesses also have penalties
or insurance associated with poor uptime performance in their contracts.
For all those reasons, it’s crucial to boost the performance of your infrastructure to
accommodate business needs or to compete better in the marketplace.
Modern processor chips integrate virtualization technology, enabling resource
splitting to accommodate increased demand. Static resource partition technology, such
as VMware vSphere, allows many organizations to run multiple operating systems in
one hardware setup. The ability to run various operating systems using virtualization,
resource partition, Linux “cgroup” technologies, and Linux Containers (LXC) delivers
virtual machines as base components of resource allocation. See Figure 1-2.

2
Chapter 1 Modern IT Infrastructure and Hello App

Virtual machine Virtual machine Virtual machine

App App App App

Libraries Libraries Libraries

Operating Operating Operating

System System System

Container Runtime

Operating System

Hardware

Figure 1-2. Virtualized deployment

Business flourished for well-known companies like VMware vSphere, Xen, and so on.
Cloud providers such as Amazon Web Services (AWS), Azure Cloud, and Google Cloud
Platform (GCP) also benefitted from the rise of virtualization. Their popular products are
called Amazon EC2, Azure Virtual machines, and Google Compute Engine.
The most significant disadvantages of virtual machines are as follows:

• Static allocation of resources (CPU and RAM)

• More resources to manage by the IT infrastructure team

• Maintenance and overhead required of multiple operating systems

(host and guest)

• Single hardware point of failure (mitigated by cluster premium

features and the Veeam backup suite)

3
Chapter 1 Modern IT Infrastructure and Hello App

The Move to Containers

Due to those disadvantages, more organizations began to adopt a multi-cloud (multiple
cloud vendors) or a hybrid-cloud (data center and one or more cloud vendors) strategy
to offer better service to their customers. They could avoid the failure of one single
vendor. Containers enable enterprises the flexibility to choose between on-premise,
private, and public cloud providers or a mix of all these infrastructure technologies
based on universal standards (SLA, quality of services, and cost). The portability of the
application is ensured by the microservices software architectural design pattern, where
applications are broken into their smallest components, independent from each other.
See Figure 1-3.

Container Container Container

App App App

Libraries Libraries Libraries

Container Runtime

Operating System

Hardware

Figure 1-3. Container deployment

To deliver the best value in the least amount of time, IT engineers needed more
powerful tools than bare metal or virtual machines. Deploying an application in a
container is entirely independent of the underlying operating systems, as shown in
Figure 1-3.
The Docker company was founded in 2011 by Kamel Founadi, Solomon Hykes, and
Sebastien Pah, after the Y Combinator in the summer of 2010. Docker debuted to the
public in Santa Clara at PyCon in 2013. Docker specifies how a container image is built
and how the image is loaded and runs. Docker created a set of platform-as-a-service
(PaaS) products that use OS-level virtualization to deliver containers, for example, the
command-line Docker utility.
4
Chapter 1 Modern IT Infrastructure and Hello App

Podman is an open-source alternative to Docker (its full name is the POD MANager).
Its main advantage is that you don’t need a service to be running like Docker on your
machine; Podman is “daemonless.” The command syntax is very similar to Docker’s. If
you are concerned about security in your applications or planning to use Kubernetes,
Podman is a better alternative. This book uses Docker because it’s more well-known and
widespread in the IT industry. But for most of the commands with the Docker prefix, you
can substitute docker with podman and obtain the same result.
The template code of the container is called a Dockerfile, which allows you to build a
binary version of the container, called an image. A more technology-agnostic alternative
to Dockerfile is a Containerfile, by the Open Container Initiative (OCI), part of The Linux
Foundation. Both are configuration files that automate the steps of creating a container
image. The OCI promotes open industry standards around container formats and
runtimes. The OCI releases three specifications: the Runtime Specification (runtime-
spec), the Image Specification (image-spec), and the Distribution Specification
(distribution-spec). Dockerfiles and Containerfiles use a combination of system tools,
system libraries, code, runtimes, and settings to specify how an application runs. A single
image can be run as many times as you want in a local Docker or Kubernetes cluster.

Container
Registry
Dockerfile build run
Containerfile container
Image

Image

Figure 1-4. Container lifecycle

Once built, the container image is stored in a Container Registry (see Figure 1-4).
The first Container Registry is Docker Hub, but many Container Registries are available
in the marketplace nowadays. The first alternative that comes to my mind is Quay from
Red Hat. Container Images are stored inside with metadata. That makes it easy for
Kubernetes clusters to search within them and find them.

5
Chapter 1 Modern IT Infrastructure and Hello App

You can maintain multiple versions of your container in a Container Registry. You
can also organize images in tags and branches that match the DEV (Development), SIT
(System Integration Test), UAT (User Acceptance Test), and PROD (Production) DevOps
principles.
Tags simply add a marker to a specific version of the image and make it easy for the
users to find it.
If you need more than one container, you must move to a Docker Compose tool,
which defines and runs multi-container Docker applications.
Modern business applications require one or more of the following features:

• High availability

• Multi-cloud compatibility

• Multi-tier storage

• Elastic/auto-scaling

• Self-healing

• Security by design (DevSecOps)

Containers can be useful for the rescue administrator workflow. In general, it’s easier
to manage storage, network, and configurations; containers help make developers more
flexible and more productive.
First introduced by Docker, containers are a standard way to deliver applications and
required libraries without the underlying operating system. This way, you can quickly
relocate to another machine or promote the development to production.
These are the benefits of using containers:

• They are the standard way to package an application and library into
one single object (container).

• They are much faster to spin up rather than using a virtual machine.

• They are easy to scale.

• They are portable because you can host anywhere.

• They enable microservices.
Running a container in a virtual machine is possible, but requires extra work to
maintain the operating system of the running host and the guest container manually in a
virtual machine. It is possible, but a sub-optimal solution.
6
Chapter 1 Modern IT Infrastructure and Hello App

Orchestration tools such as Kubernetes allow you to easily manage containers

at scale.
Global trends of the DevOps and DevSecOps developer-centric methodologies
deliver application-as-service, microservices, and serverless applications. They are
ramping up faster and faster. In this scenario, containers created by the Continuous
Integration and Continuous Delivery (CI/CD) toolset are tested for quality assurance
and promoted to production faster than ever. Apache Mesos and Docker Swarm
technologies are also often involved.

Ansible by Red Hat

Ansible is the leading open-source infrastructure automation technology. Infrastructure
teams and system administrators worldwide realized that they needed a better way to
scale their systems’ management scripts to keep up with business demands. The hosted
web applications increase the complexity, email flow, and new releases of the operating
system. Manual work becomes unsustainable when the number of target devices grows.
API-driven server management and configuration management tools like Ansible helped
make things manageable.
The main Ansible use case is as follows:

• Provision

• Config management

• Application deployment

Red Hat Inc. (part of IBM since 2019) is leading the development of the Ansible
project worldwide, guaranteeing the open-source quality of code.
The Ansible platform is available for a huge variety of modern operating systems
and requires only an OpenSSH connection and a Python interpreter on the Linux
target node. It supports a wide range of operating systems via different connection
technologies. For example, you can use WinRM for Windows target nodes and Network
API for network devices.
The learning curve is very rapid. You see this for yourself in Chapter 2, where you
learn all about the Ansible programming language.
Red Hat also organizes the yearly AnsibleFest worldwide, usually at the end of
October, to educate people about Ansible and share new features and success stories.

7
Chapter 1 Modern IT Infrastructure and Hello App

The Cloud Native Computing Foundation

In the early stage of application containerization, runtime systems such as Docker and
Podman were extremely useful for manually deploying and testing applications.
However, when the workload became higher and managing failover was crucial, the
world needed a more automated way of doing this.
Kubernetes (also called K8s) is the best way to orchestrate and manage a container
workload. Kubernetes take care of deploying, scaling, and managing container
applications. It is open-source and completely automatic.
Kubernetes has many releases per year; at the moment, there are three
releases a year.
While Docker is focused on how a single container image is created, loaded, and run,
Kubernetes is a way to schedule and manage a large number of containers.
You can deploy your container in Kubernetes to produce applications at a scale.
You can seamlessly move containers and applications at scale. Similarly to Docker and
Podman, you can run your container in Kubernetes. Kubernetes run anywhere—on-
premise, in a private data center, in a public cloud, and even in embedded hardware
such as the latest tiny Raspberry Pi devices. Kubernetes supports multiple processor
architectures and modern operating systems.
Updating, upgrading, and managing in Kubernetes is relatively easy and guarantees
security for your organization.
The Cloud Native Computing Foundation (CNCF) leads the development of
the Kubernetes project, offering a vendor-neutral vision. CNCF is part of the Linux
Foundation. CNCF also organizes community events, including KubeCon and
CloudNativeCon, where you can meet experts and share stories.
The most inspiring customer stories are from innovation leaders such as BlackRock,
Netflix, Zalando, Uber, The New York Times, ING, Squarespace, BlaBlaCar, Huawei,
Amadeus, PNC Bank, Seagate, Verizon Media, and much more.
In Kubernetes, a cluster is made up of at least one Kubernetes master; many
Kubernetes worker machines are called nodes. A cluster is fundamental for Kubernetes:
all containerized workloads and services run within a cluster.

8
Chapter 1 Modern IT Infrastructure and Hello App

Kubernetes Support
Kubernetes relies on a vibrant community, external experts, and consultant services.
Kubernetes is officially a self-support model, but more private organizations offer
support and customization services.
Canonical, the organization leading the Ubuntu project, offers support for Kubeadm,
MicroK8s, and Charmed Kubernetes on VMWare, OpenStack, bare metal, AWS, Azure,
Google, Oracle Cloud, IBM Cloud, and Rackspace.

ubernetes Distributions: OpenShift, Rancher, EKS,

K
AKS, and GCP
There are currently several Kubernetes distribution options. They all integrate the core
Kubernetes features, but they differ by the included modules and managing software.
Kubernetes is the main framework for developing commercial products and cloud
services.

Note This book focuses on the generally available Kubernetes, while most of the
code also works in some cloud-native Kubernetes services.

OpenShift by Red Hat

OpenShift Container Platform (OCP) is a key product of the Red Hat multi-cloud
strategy. It is getting traction in the market because it’s easy to set up and maintain.
Red Hat packages the core Kubernetes experience with a selection of valuable utilities
to simplify the new user onboarding process. OpenShift Container Platform requires
having a OpenShift subscription to use it, which includes Red Hat support. For this
reason, it supports only Red Hat operating systems underneath, which is Red Hat
Enterprise Linux (RHEL) and Red Hat CoreOS (discontinued).
Red Hat has a strong cloud-provider partnership for fully managed OpenShift
services with products like Red Hat OpenShift Service on AWS (ROSA), Azure Red Hat
OpenShift, and Red Hat OpenShift on Google Cloud Marketplace.

9
Chapter 1 Modern IT Infrastructure and Hello App

The main differences between OpenShift and Kubernetes are that OpenShift:

• Requires a paid subscription

• Has limited operating system support (Red Hat Enterprise Linux

(RHEL) and Red Hat CoreOS

• Uses the oc command-line tool instead of kubectl.

• Has stricter security policies than default Kubernetes

• Forbids running a container as root

• Uses Role Based Access Control (RBAC) security

• Routes object instead of Ingress of Kubernetes based on HAproxy

• Uses ImageStreams for managing container images

• Uses OpenShift projects vs Kubernetes namespaces

OpenShift uses Helm as a software package manager, which simplifies the packaging
and deployment of applications and services to OpenShift Container Platform clusters.
OpenShift mainly uses Kubernetes operators; however, Helm is very popular in many
Kubernetes applications. For Kubernetes-native applications, consider using Kubernetes
Operator instead, because Operator continuously checks the application’s status and
determines if the application is running according to the configuration defined by the
software developer.
Red Hat OpenShift Local (formerly Red Hat CodeReady Containers) is the fastest
way to spin up an OpenShift cluster on your desktop/laptop for development purposes.
This happens in minutes. It supports Linux (x86), Windows (x86), and macOS, even Intel
and Apple Silicon (M1 and M2 processors). It is a simplified version relying on a virtual
machine (~20GB) to set up, run, test, and emulate the cloud development environment
locally. It uses the crc command-line utility to spin up a complete lab with an API and
web interface. The requirements of the latest 2.10 release are four physical CPU cores,
9 GB of free memory, and 35 GB of storage space.
Red Hat also releases an open-source version of OpenShift called OKD. It’s free to
use and includes most of the features of its commercial product without support.

10
Chapter 1 Modern IT Infrastructure and Hello App

Kubernetes in the Public Cloud

Many cloud provider vendors build their Container Orchestration and Management
products. The most popular cloud-native services are as follows:

• Amazon Elastic Container Service (Amazon ECS) by Amazon Web

Services (AWS)

• Google Container Engine (GKE) by Google Cloud Platform (GCP)

• Azure Container Service by Microsoft Azure

• IBM Bluemix Cloud Kubernetes Container Service by IBM Cloud

• Oracle Container Cloud Service (OCCS) by Oracle Cloud

• Alibaba Cloud Container Registry

More services are available in the IT industry, but I consider these the major actors
in the marketplace nowadays. See Chapter 7 for more details.

Amazon EKS
Amazon Elastic Kubernetes Service (Amazon EKS) is a service of Amazon Web Services
(AWS). It adds the eksctl command-line tool for working with EKS clusters and uses the
kubectl command underneath.
Amazon also provides Amazon Elastic Container Registry (Amazon ECR), an AWS-
managed container image registry service that’s cross-region and completely integrated
within the AWS ecosystem. Specified users or Amazon EC2 instances defined in the IAM
policies can access the ECR container repositories and images. Amazon ECR manages
private image repositories consisting of Docker and Open Container Initiative (OCI)
images and artifacts. Amazon Linux container images are already available in ECR, built
with the same software components included in the Amazon Linux AMI for applications
in Amazon EC2. See Chapter 7 for more details.

Containers and Pods
A container is a single unit for delivering software. But the real world is more complex;
sometimes you need more than one container to provide the complete solution.

11
Chapter 1 Modern IT Infrastructure and Hello App

A pod contains

• One group of one or more containers with shared storage and

network resources

• One specification for how to run the containers

The name comes from the fact that a group of whales is called a pod. Pods enable
microservice deployment. The advantage is that they are loosely coupled, and their
execution is autonomous and independent.
Pods are the smallest deployable objects in Kubernetes. The pod’s contents are
colocated, coscheduled, and run in a shared context. The best practice is to insert only
containers that are tightly coupled in a pod for an application-specific task or feature.
As well as a container, a pod might contain an init container that runs during the
pod’s startup.

Note For debugging purposes, you can inject ephemeral contains if your

cluster has enabled this feature.

Creating a Hello App

The “Hello” app is a simple implementation in a container of the “Hello world” message
that prints the text message "Hello world" onscreen and terminates its execution.
First of all, you need to define a Dockerfile. You can rename the Dockerfile as
Containerfile to be fully OCI-compliant. This simple task requires you to execute the
UNIX command echo into a Dockerfile. Before being able to execute the command,
you need to add some lines of code to instruct Docker (or Podman) about the system
requirements, some metadata, and finally, the command to execute.
The simple example in Listing 1-1 shows the full Dockerfile to print the Hello World
text on the screen using the shell echo command.

Listing 1-1. A simple Dockerfile

#A simple Docker container image.
FROM busybox
LABEL org.opencontainers.image.title="Hello World image"

12
Chapter 1 Modern IT Infrastructure and Hello App

LABEL maintainer="Luca Berton"

#Run a command
CMD echo "Hello World"

Every line prefixed with a # (number/pound symbol/hashtag) is considered a

comment, so it’s ignored by the program. It is useful for documentation or sharing
developer information.
The FROM instruction tells Docker that you want to begin a new image using, in this
case, another image as your starting point. It is starting from the image called busybox.
BusyBox is a very popular tool in many embedded Linux systems. In a ~5 MB of on-
disk storage space, BusyBox delivers a tiny version of many common UNIX utilities into a
single small executable. It provides the base functionality for the UNIX most commonly
used utilities. The modern Linux distributions use the GNU full-feature alternative
tools packaged in GNU fileutils, shellutils, and so on. The utilities in BusyBox generally
have fewer options than their full-featured GNU alternatives. Despite the limited size,
the included options provide the expected functionality and behave very much like
their GNU counterparts. This is why BusyBox is so well known and used outside the
embedded domain.
The LABEL instruction adds metadata to the image. The metadata enables the
Container Registry to search and organize the images. The label instructions might be
multiple and follow a key=value format, where the key is the parameter’s name and the
value is the associated value.
The CMD instruction tells Docker that what follows is a shell command and is the
process to run upon the container start.
This is a very minimal Dockerfile; there are more instructions available but these are
the most basic.

Linux Base Images

There are a huge amount of base images available in the Container Registries for your
application development that you can use in the FROM instruction. Deriving an image
from another is actually the best way to reuse code and obtain security updates from the
original image maintainer. You don’t need to reinvent the wheel!

13
Chapter 1 Modern IT Infrastructure and Hello App

Fedora Linux
Imagine you want to use the Fedora or Ubuntu images as base images.
For the Fedora Linux distribution, you could specify the following:

FROM fedora:latest

In this way, you select the latest release of the Fedora Linux distribution. As I’m
writing this book, the latest release is Fedora 37. The latest release is typically updated
every six months, based on the current release cycle. This acts like a rolling-release
pointer.

FROM fedora:37

You select a specific Fedora version release, 37 in this case.

FROM fedora:rawhide

In this way, you have selected the rawhide release channel, which is suitable for
developers, is sometimes unstable, and updates very often. This should be selected only
by users who know what they are doing.

Ubuntu Linux
Many popular Linux distributions are available as containers. The Ubuntu Linux
distribution is also available in a container image.
The Ubuntu Linux distribution is available in Docker Registry:

FROM ubuntu:latest

In this way, you select the latest LTS release of the Ubuntu Linux distribution. When
I was writing this book, the latest release was Ubuntu 22.04.1 LTS Jammy Jellyfish,
released on August 10, 2022. It is an LTS release with long-term support, which means
five years of free security and maintenance updates, guaranteed until April 2027. Note
that the latest release always points to the latest release, so it’s updated every six months
considering the current release cycle. This acts like a rolling-release pointer.

FROM ubuntu:22.04 or FROM ubuntu:jammy

14
Chapter 1 Modern IT Infrastructure and Hello App

You can specify one Ubuntu-specific version using the release numbers or name.
In the specific, the version number is 22.04 and the name is jammy, which refers to the
Jammy Jellyfish release name.

FROM ubuntu:rolling

You can specify the “rolling release,” which is the latest release available, using the
rolling keyword. Currently, I’m writing this book referring to the 22.10 release, Kinetic
Kudu name, with a release date of October 21, 2022. Note the rolling release is supported
only for nine months until July 2023.

Alpine Linux
Another Linux distribution popular as a base for container images is Alpine Linux.
Alpine Linux is a security-oriented, lightweight Linux distribution based on a different
version of the system libc called musl against the usual GNU libc and BusyBox. It is
famous because of its very small footprint and reduced attack surface. The image is
~5 MB in size and has access to the Alpine package repository via the Alpine Package
Keeper (apk) package manager.

Enterprise Linux-Based Images

Enterprise releases of Linux are available as containers as well. They are very powerful,
especially for partners and system integrators that develop their solutions on top
of them.
Red Hat was the first company to introduce the Red Hat Universal Base Image (UBI)
in May 2019 for Red Hat Enterprise Linux (RHEL) versions 7 and 8. RHEL 9, released on
May 2022, is also available as a Universal Base Image (UBI). UBIs are a lighter-weight
versions of the Linux operating system’s userland stripped down to the bare essentials
and distributed by the official release channels.
UBIs are designed to be a foundation of cloud and web application use cases. Images
are available via Red Hat Container Registry at https://catalog.redhat.com/.
UBIs are tested for Open Container Initiative (OCI) compliance, performance, and
security. UBIs are also freely redistributable and deployable on non-Red Hat platforms.
Many Red Hat partners distribute solutions on top of UBIs and distribute them as
new images.

15
Chapter 1 Modern IT Infrastructure and Hello App

Red Hat Universal Base Images are available for each released operating system
version in four sizes: Standard, Minimal, Micro, and Multi-service for AMD64 and
Intel 64 (amd64), ARM (arm64), Power ppcle64, and IBM zSeries s390x processor
architectures.

• The standard ubi UBI images (compressed ~76 MB) are general
purpose and designed for every application that runs on RHEL,
including the standard YUM or DNF Package Manager, OpenSSL
crypto stack, and basic OS tools (tar, gzip, vi, and so on).

• The multi-service ubi-init UBI images (compressed ~80 MB) are on

top of the standard (ubi) UBI and contain the systemd initialization
system services, useful for building images to deploy services on top
of it. For example, a web application or a file server.

• The minimal ubi-minimal UBI images (compressed ~38 MB) are

very minimalistic images with a very small footprint; they use the
microdnf Package Manger (a basic package manager to install,
update, and remove packages extensible via modules).

• The micro ubi-micro UBI image (compressed ~11 MB) is the

smallest possible UBI image, without a package manager and all of
its dependencies, reducing the attack surface of container images
for highly regulated and security environments. UBI micro was
introduced in June, 2021.

Other useful Red Hat Universal Base Images include the following:

• Language runtime images (nodejs, ruby, python, php, perl,

and so on)

• A repository full of packages for the most common applications

There are other Enterprise Linux distributions available as container images.

Since August 2022, the SUSE Linux Enterprise Server Base Container Images (BCI)
via the SUSE Container Registry (https://registry.suse.com) images for AMD64 and
Intel 64 (x86_64), ARM (aarch64), Power (ppcle64), and IBM zSeries (s390x) processor
architectures.

16
Chapter 1 Modern IT Infrastructure and Hello App

At the moment, the following five SUSE Enterprise Linux BCIs are available:

• Base bci-base (compressed ~47 MB): SUSE Linux Enterprise

general-purpose image with zypper and a minimal set of packages

• Init bci-init (compressed ~58MB): Similar to bci-base with a

systemd service manager for managing services

• Minimal bci-minimal (compressed ~15 MB): For deployment

container RPM-only with manual package dependencies resolution

• Micro bci-micro (compressed ~10 MB): For static-binary

deployment without a package manager

• BusyBox bci-busybox (compressed ~5 MB): The UNIX coreutils tools

provided by the BusyBox project instead of GNU coreutils for size
reduction

Other Enterprise Linux distribution images include these:

• Rocky Linux image rockylinux tags 8, 8-minimal, 9, 9-slim

• Oracle Linux image oraclelinux tags 7, 7-slim, 8, 8-slim, 9, 9-slim

Container Security
Container images are created by software that might be affected by security flaws
that could harm your organization. The advantage of using maintained images is that
the creator constantly monitors the security of the images through the static analysis
of vulnerabilities and releases a new version when necessary. Some examples of
impactful security threats that pop up are Log4Shell (CVE-2021-44228), Heartbleed
(CVE-2014-0160), and Shellshock (CVE-2014-6271).
Once you create your container image, you can use a vulnerability analysis engine
to inspect the image layer-by-layer for known security flaws. Most Container Registries
integrate some container image static analysis tools.

• Docker Hub Vulnerability Scanning has been powered natively by

Docker, instead of a third party, since February 27, 2023.

• Red Hat Quay uses the open-source Clair security framework.

17
Chapter 1 Modern IT Infrastructure and Hello App

These tools scan each container image layer and notify you of vulnerabilities
that might be a threat. They use the popular Internet thread databases to search for
security patterns: NIST Common Vulnerabilities and Exposures database (CVE), Red
Hat Security Bulletins (RHSB), Ubuntu Security Notices (USN), and Debian Security
Advisories (DSA).
In the market, there are many open-source and commercially available products for
this task nowadays. Popular tools include Anchore, Datadog Cloud SIEM, Sophos Cloud
Native Security, Bitdefender GravityZone, Sysdig Secure, Red Hat Advanced Cluster
Security for Kubernetes, and Aqua Security. Integrating one or more of these products
into your DevSecOps process is a good way to address and mitigate security risks.

The Hello Dockerfile

Now that you know more about the Dockerfile format, you can move forward with your
Hello App.
The Hello App is a Python 3 base on the Flask Restful framework. Flask is a popular
web application framework for Python, implemented on Werkzeug and Jinja2. Flask-
RESTful is a popular Flask extension for quickly building REST APIs. It’s a fast way to
integrate a web server into a few lines of code.
The two Hello App source files are shown in Listing 1-2.

Listing 1-2. Hello Dockerfile

– Dockerfile

FROM python:3
LABEL org.opencontainers.image.title="Hello World image"
LABEL maintainer="Luca Berton"
ADD helloworld.py /
RUN pip install flask
RUN pip install flask_restful
EXPOSE 8080
CMD [ "python", "./helloworld.py"]

The FROM instruction informs Docker that you want to derive from the python:3
image. At the moment of writing the book, tag 3 of the python image points to the latest
release of the Python image. The same result could be obtained using the tags 3.11.0,
3.11, 3, and latest. The Python image is the official distribution of the Python computer
18
Chapter 1 Modern IT Infrastructure and Hello App

programming language in a slim release. Docker Hub Container Registry is available

on top of the latest Debian 11.5 “bullseye” Linux distribution, released on September
10, 2022.
The ADD instruction adds files to the image from the current directory. You need to
specify the source filename and the destination.
The LABEL instruction adds metadata to the image—specifically an image title and
the maintainer’s name.
Line number 4 uses the LABEL instruction:

ADD helloworld.py /

This line copies the helloworld.py file into the destination directory / (root), which
is the base directory of every UNIX file system.
The RUN instruction executes some commands inside the image. Specifically, line
number 5 and line number 6 execute the pip command.

RUN pip install flask

RUN pip install flask_restful

The pip command is the Python Package manager that is used to install additional
content inside the container. The advantage of this approach, rather than using the
system packages, is that it fetches the latest release of the package and compiles it for
your system. The pip command also takes care of the dependencies and ensures that the
software is up-to-date on your system. In this example, it added the flask and flask_
restful programs.
The EXPOSE instruction specifies the network port of the service of your container. In
this case, port 8080 is required by the web application.
The CMD instruction tells Docker to execute the shell command and is the process to
run upon the container start. It runs the concatenation of the two strings of the list. So
the python system command and the ./helloworld.py application.

The Hello Application

The Hello application is a simple Python program that you will execute inside a
container. Listing 1-3 details the contents of this helloworld.py application.

19
Chapter 1 Modern IT Infrastructure and Hello App

Listing 1-3. helloworld.py

– helloworld.py

#!/usr/bin/env python3
from flask import Flask, request
from flask_restful import Resource, Api
app = Flask(__name__)
api = Api(app)
class Hello (Resource):
    def get(self):
        return 'Hello World'
api.add_resource(Hello, '/') # Route_1
if __name__ == '__main__':
    app.run('0.0.0.0','8080')

A few lines of this code run a web server on port 8080 and serve the default address /
with the Hello World message.
You can test this code in a browser with the target address of localhost and the
local port 8080 by just typing localhost:8080 in your favorite browser. When the
helloworld.py application runs, it returns the Hello World text on the screen, as shown
in Figure 1-5.

Figure 1-5. The hello app in a browser

In order to run the helloworld.py file, you must manually install all the necessary
dependencies. In this case, the installation of the packages flask and flask_restful.
This process is boring, repetitive, and error-prone; you need to automate it with a
container image. The best way to test the code is to create a Python virtual environment
using the Python venv module and then use the pip command to install the required
packages.

20
Chapter 1 Modern IT Infrastructure and Hello App

To install the required dependencies (all the commands listed in the Python
import), use:

pip install flask flask_restful

If you want to share the virtual environment with a friend, you could generate a
requirements.txt file with the full dependency tree that’s usable on every machine
with Python installed. Listing 1-4 shows the requirements file generated on my machine
at the moment of writing this book (versions of the dependencies might vary on your
system).

Listing 1-4. requirements.txt

– requirements.txt

aniso8601==9.0.1
click==8.1.3
Flask==2.2.2
Flask-RESTful==0.3.9
itsdangerous==2.1.2
Jinja2==3.1.2
MarkupSafe==2.1.1
pytz==2022.6
six==1.16.0
Werkzeug==2.2.2

You could easily install all the required dependencies and run helloworld.py typing
these commands:

pip install -r requirements.txt

python helloworld.py

The first command installs all the necessary software dependencies and the
second runs your application. The expected output for successfully listening for a new
connection is the following:

* Serving Flask app 'helloworld'

* Running on http://127.0.0.1:8080

21
Chapter 1 Modern IT Infrastructure and Hello App

Building the Hello App

Before being able to execute your container, you need to build the binary image, as
illustrated in Figure 1-6. The process requires some time but is completely automated.
Once you practice, it will become straightforward.

Dockerfile build Image

Containerfile

Figure 1-6. The container build process

Using the build parameter of the docker command in the terminal, you request the
building of your container, specifying the image name using the -t parameter and the
name hello-world-python:

$ docker build -t hello-world-python .

Under the hood, Docker downloads the base image name Python and selects tag 3.
After that, Docker adds the helloworld.py file to the file system of the new container
and downloads the required Python dependencies.
The result is a container named hello-world-python in your workstation. In case
you’re wondering, it’s ~1GB on my workstation.

Running Hello in Docker

Once the container image is successfully built, you can execute it as many times as you
want or upload it to a Container Registry (Docker Hub, Quay.io, Google’s Container
Registry, and so on). Figure 1-7 illustrates the container run process.

22
Chapter 1 Modern IT Infrastructure and Hello App

run
Image Container

Figure 1-7. Container run process

You can run your containers by specifying the port mapping using the 8080 port of
the localhost mapped to the port 8080 on the container.

$ docker run -p 8080:8080 hello-world-python

You can check the status of the Docker execution using the ps parameter of the
docker command:

$ docker ps
CONTAINER ID IMAGE
2c66d0efcbd4 hello-world-python

The computer-generated string 2c66d0efcbd4 represents the container ID of the

running container in your workstation.
You can verify the execution using the command-line browser cURL:

$ curl http://localhost:8080

If you prefer to use your favorite browser software, just type localhost:8080 in the
title bar; Figure 1-8 shows the result.

Figure 1-8. The hello app in a browser

23
Chapter 1 Modern IT Infrastructure and Hello App

As you can see, the result of the execution using Docker is the same as the manual
Python execution of the Hello App.
The main advantage is that you can execute the container as many times as you want
or share it with others via the Container Registry.

Deploying Hello in Kubernetes

In order to deploy your image to Kubernetes, you need to create a manifest file called
hello.yml. Manifest files are written in YAML programming code and are very easy for
humans to understand. See Listing 1-5.

Listing 1-5. The hello.yml Manifest File

- hello.yml
apiVersion: v1
kind: Pod
metadata:
   name: redis-pod
spec:
   containers:
      - name: hello-container01
        image: hello-world-python:latest
        ports:
           - containerPort: 8080

Once you create the hello.yml manifest file, you can deploy the Hello App to your
Kubernetes cluster. The cluster by itself fetches the image and starts running, waiting for
new connections:

$ kubectl create -f hello.yml

pod/hello-container01 created

You can then verify that the Kubernetes cluster has successfully initialized all the
Hello App resources using the following command:

$ kubectl get pods

NAME READY STATUS RESTARTS AGE
hello-world-python 1/1 Running 1 11h

24
Chapter 1 Modern IT Infrastructure and Hello App

Deploying Hello in Operator

Operator enables you to verify and deploy new applications in your Kubernetes cluster
automatically. You need some coding skills, because the code is going to rely on the
Kubernetes Operator SDK, which is a Minikube environment for running Kubernetes
locally, as well as the Go language (Golang). At the moment, the app is so tiny that it
seems overdoing it to create an operator to display a text message, but it’s good practice.
The seven steps needed to generate a Kubernetes Operator are as follows:
1. Generate the boilerplate code using the Kubernetes Operator SDK
and Minikube:

$ minikube start init

$ operator-sdk init

2. Create the APIs and a custom resource:

$ operator-sdk create api --version=v1alpha1 --kind=Traveller

3. Download any dependencies.

4. Create a deployment.

5. Create a service.

6. Add a reference in the controller for the deployment and service.

7. Deploy the service, install the CRD, and deploy a CRD instance.

Operators extend Kubernetes APIs and create custom objects to customize the
cluster according to your needs.

Key Takeaways
IT infrastructure is more critical than ever. It serves the business mission and delivers
the best possible service to stakeholders both inside your organization and outside, to
customers and partners.
The Kubernetes platform orchestrates your container fleet precisely and efficiently.
K8s, combined with Ansible, enables incredible powerful horsepower that allows you to
use DevSecOps, Infrastructure as Code, and self-healing infrastructures.

25
CHAPTER 2

Ansible Language Code

Ansible is an IT automation tool that enables Infrastructure as Code (IaC). Michael
DeHaan started the Ansible project. The first release of Ansible was made public on the
February 20, 2012. Michael took inspiration from several tools he had written, along
with some hands-on experience with the state of configuration management at the time.
Some of Ansible’s unique attributes, such as its module-based architecture and agentless
approach, quickly attracted attention in the open-source world. The logo is reminiscent
of the Start Trek Communicator device of instantaneous communications between
galaxies.

What Is Ansible?
Ansible is classified as an infrastructure automation tool; it allows you to quickly
automate your system administrator tasks. Infrastructure as Code manages and provides
computer data centers through machine-readable definition files rather than physical
hardware configuration or interactive configuration tools. With Ansible, you can deploy
your Infrastructure as Code on-premises and on well-known public cloud providers.
Ansible is used to apply the DevOps principles in worldwide organizations. DevOps
is a set of practices that combines software development (Dev) and IT operations (Ops).
As DevOps is a cross-functional mode of working, those who practice the methodology
use different tools, referred to as “toolchains,” rather than a single tool. These toolchains
fit into one or more categories, reflective of key aspects of the development and delivery
process.
These are the six benefits to using Ansible and counting:

• Simple: The code is in the YAML language, a human-readable

data serialization language. It is well-known and easy to learn; it is
commonly used for configuration files and in applications where data
is being stored or transmitted.

27
© Luca Berton 2023
L. Berton, Ansible for Kubernetes by Example, https://doi.org/10.1007/978-1-4842-9285-3_2
Chapter 2 Ansible Language Code

• Powerful: Ansible is robust and battle-tested in configuration

management, workflow orchestration, and application deployment.

• Cross-platform: Ansible is agentless, which means that it supports

all major operating systems, as well as physical, virtual, cloud, and
network providers. Agentless support for all primary OS, physical,
virtual, cloud, and networks.

• Works with existing tools: Ansible can work with existing tools,
making it easy to homogenize the current environment.

• Batteries included: Ansible is bundled with many modules to

automate the most common tasks.

• Community-powered: Every month, Ansible has more than 250,000

downloads, an average of 3,500 contributors, and more than 1,200
users on IRC.

Ansible can be easily extended by using additional resources and plugins. The
most common way to distribute these is via an Ansible collection that comes with a
standardized format. The more than 750 modules include cloud modules for clustering,
executing commands, crypto, database, managing files and directories, identity
management, inventory, messaging, monitoring, net tools, network, notification,
packaging, remote management, source control and versioning, storage, system, utilities,
web infrastructures, and interacting with Windows operating systems.
Ansible’s three prominent use cases are provisioning, configuration management,
and application deployment. After using this technology, you may just invent more ways
to use it!

Provisioning
Provisioning is the process of setting up the IT infrastructure: all system administrators
know how important it is to manage a uniform fleet of machines. Some people still rely
on software to create workstation images. But there is a drawback to that; with imaging
technology, you’re only taking a snapshot of the machine at a certain time. You have to
reinstall software every time because of modern critical activation systems or to update
the latest security patches. Ansible is very powerful in automating this process.

28
Chapter 2 Ansible Language Code

Configuration Management
Configuration management is the process of maintaining systems and software in a
desired and consistent state. It maintains an up-to-date and consistent fleet, including
coordinating rolling updates and scheduling downtime. With Ansible, you can verify the
status of your managed hosts and take action in a small group of them. A wide variety
of modules is available for the most common use cases, not to mention the typical use
case of checking for the compliance of your fleet to international standards and apply
resolution plans.

Application Deployment
Application deployment is the process of publishing software between testing, staging,
and production environment. For example, Ansible can automate your web application’s
continuous integration/delivery workflow pipeline. Your DevOps team will be delighted!

Ansible Open-Source vs Commercial Options

Ansible is a community-driven project with fast-moving innovations. It’s open-source
with only command-line tools.
Ansible Inc., originally AnsibleWorks Inc., was the company set up to support and
sponsor the project commercially by Michael DeHaan, the founder of the Ansible
project. In October of 2015, Red Hat acquired Ansible Inc. and evaluated Ansible as a
“powerful IT automation solution” designed to help enterprises move toward frictionless
IT. Since 2016, AnsibleFest has been an annual conference of the Ansible community of
users and contributors; previous editions were hosted in Europe and the United States.
Since 2023, the event has been part of the Red Hat Summit main event planned in May
each year.
Red Hat Ansible Automation Platform is the commercial framework designed by
Red Hat. You can interact using a web user interface, an API, command-line interfaces,
and Ansible collections. It provides configuration management (CM), CM tools, a GUI,
Role-Based Access Control (RBAC), and third-party authentications and integrations to
manage your infrastructure. Red Hat Ansible Automation Controller was formerly known
as Ansible Tower. Red Hat Ansible Automation Platform has loads of products, including
Ansible Automation Controller, Automation Hub, Automation Analytics, and more.

29
Chapter 2 Ansible Language Code

Enterprise needs more services and some stable releases. For example, they need
an SLA for support. Red Hat offers this service to companies under the Ansible Tower
umbrella, now rebranded as Ansible Automation Controller.
Ansible Tower is a REST API, web service, and web-based console designed to make
Ansible more usable for IT teams with members of different technical proficiencies and
skill sets. It is a hub for automation tasks. The Tower is a commercial product supported
by Red Hat Inc. but derived from the AWX upstream project, which has been open-
source since September 2017.
Red Hat maintains Ansible Core (formerly Ansible Engine) with the explicit intent
of being used in the open-source community and as an enterprise IT platform. With
Ansible Core, organizations can access the tools and innovations available from the
underlying Ansible technology in a hardened, enterprise-grade manner.

Ansible’s Architecture
Ansible typically requires two or more hosts—one that executes the automation, called
the Ansible Control Node, and one (or more) that receives the action, called the Target
Node. Figure 2-1 illustrates the architecture.

web1.example.com

Ansible Control Node web2.example.com

web3.example.com

Figure 2-1. Ansible Architecture

The Ansible Control Node directs the automation and effectively requires Ansible to
be fully installed inside. The Ansible Target Node requires only a valid login to connect.
The Ansible Control Node usually uses Ansible Playbook and Inventories (see
Figure 2-2) for the execution. The Ansible Playbook is the automation blueprint and has
a step-by-step list of tasks to execute against the target hosts.

30
Chapter 2 Ansible Language Code

Ansible Control Node

Ansible Playbook
—
- name: example
hosts: all
tasks:
- name: hello
ansible.builtin.debug:
msg: Hello
…

Ansible Inventory

web1.example.com
web2.example.com
web3.example.com

Figure 2-2. Ansible Playbook and Inventory

Ansible Playbook is coded in the YAML language, so easy to write and human-
readable.
The Ansible Inventory is the list of target hosts in INI, YAML, or JSON format. For
simplicity, you can group the hosts using a common name. In this way, you can easily
specify the group name for the execution.
The next section reviews how to install and configure a control node for Ansible for
the most common operating systems.

UNIX Target Node

You can configure Ansible in the following UNIX target node: Linux distributions,
macOS, and UNIX (FreeBSD, OpenBSD, etc.), as shown in Figure 2-3.

31
Chapter 2 Ansible Language Code

Linux
Ansible Control Node ss
macOS
h
FreeBSD, etc.

Figure 2-3. Ansible UNIX target nodes

The supported Target Node includes all the current releases of UNIX-compliant
desktops and servers.
Ansible requires only OpenSSH and Python to be installed. OpenSSH is used
for connection and one login user. Using SSH keys instead of passwords is strongly
encouraged. The local Python interpreter in the target node will execute the Ansible
commands.

Windows Target Node

The supported Target Node includes all the modern Windows desktop and server
releases. The list includes Windows 7, 8.1, 10, and 11 and Windows Server 2008, 2008 R2,
2012, 2012 R2, 2016, 2019, and 2022 at the moment of writing this book, but Windows
support is future-proof. See Figure 2-4.

Windows
Ansible Control
Node
Windows
Server

Figure 2-4. Ansible Windows target nodes

The Windows host requires PowerShell 3.0 or newer and at least .NET 4.0. You only
need to upgrade old Windows 7 and Windows Server 2008 nodes.
The WinRM listener receives and executes the commands between Ansible
Controller and the target node. Ansible 2.8 added an experimental SSH connection for
Windows-managed nodes for Windows 10+ clients and Windows Server 2019+.

32
Chapter 2 Ansible Language Code

The easier, more straightforward configuration is using the WinRM connection

method with the basic authentication.
You can specify more advanced authentication methods: Basic, Certificate, NTLM,
Kerberos, and CredSSP. You can do this for local user accounts, Active Directory
accounts, credential delegations, and HTTP encryption, as shown in Table 2-1.

Table 2-1. WinRM Authentication Methods

Option Local Accounts Active Directory Credential HTTP
Accounts Delegation Encryption

Basic Yes No No No
Certificate Yes No No No
Kerberos No Yes Yes Yes
NTLM Yes Yes No Yes
CredSSP Yes Yes Yes Yes

Basic authentication is the simplest authentication option but also the most insecure
because it transmits the password encrypted with base64 encoding.
The following example shows the WinRM host variables configured for basic
authentication in the Ansible Inventory:

ansible_connection: winrm
ansible_winrm_transport: basic
ansible_user: LocalUsername
ansible_password: Password

WinRM Certificate authentication uses certificates as public/private key pairs in

PEM format. The following example shows how to configure WinRM host variables for
certificate authentication in the Ansible Inventory:

ansible_connection: winrm
ansible_winrm_transport: certificate
ansible_winrm_cert_pem: /path/to/certificate/public/key.pem
ansible_winrm_cert_key_pem: /path/to/certificate/private/key.pem

33
Chapter 2 Ansible Language Code

For all the other installation scenarios and comprehensive configuration details,
refer to the documentation on the official Ansible website: https://docs.
ansible.com/.

Ansible Installation
The Ansible package must be installed only on the Ansible Control Node (see
Figure 2-5). Ansible is available for a large variety of modern operating systems. You can
configure the Ansible Control Node in a large variety of UNIX operating systems: Linux
distributions, macOS, UNIX (FreeBSD, OpenBSD, and so on).

Ansible Control Node

ansible package

Figure 2-5. Ansible Control Node

The supported control node includes the current UNIX-compliant desktop and
server releases.
Ansible Engineering Team distributes a minimalist language and runtime called
ansible-core (it was called ansible-base in version 2.10), and the Ansible Community
Team has a community package called ansible.
Choose the Ansible style and version that matches your particular needs or is
available in your operating system.
Ansible Core (ansible-core package) focuses primarily on developers and users
who want to install only the basic Ansible language and runtime and add any additional
Ansible Collections. It contains only the ansible.builtin modules and plugins,
allowing other Ansible Collections to be installed. The Ansible Core is similar to Ansible
2.9 without any content; it has since moved into external Collections, resulting in a
smaller package on disk. The current release cycle at the moment of writing this book of
Ansible Core is two major releases per year.

34
Chapter 2 Ansible Language Code

The ansible Community package includes the Ansible Core plus 85+ community-
curated Collections that contain thousands of modules and plugins. It re-creates and
expands on the functionality included in Ansible 2.9 in a bigger package on disk. The
current release cycle of the Ansible Community package at the moment of writing this
book is two major versions per year, on a flexible release cycle that trails the release of
the ansible-core package.
The easiest way to install Ansible is to install the ansible or ansible-core in your
package manager or use the pip tool for Python.
Use the following command on Red Hat Enterprise Linux, Fedora, CentOS, Rocky
Linux, Alma Linux, Oracle Linux, or Amazon Linux systems with root privileges to install
Ansible:

yum install ansible

On Debian and Ubuntu Linux systems, use this command (root privileges required,
often sudo is used):

apt-get install ansible

The PyPI is the largest archive of Python modules, and you can interact with it
using the popular pip command-line utility. Using the pip command-line utility, you
can download and install the latest release available from the PyPI website on your
system. System administrators often configure a Python virtual environment in order to
separate the Ansible installation and the library that might be necessary for additional
components, such as roles and collection from the system libraries.
Using the Python pip tool, you can install Ansible using the following command
(root privileges may or may not be required, depending on your environment):

pip install ansible

The pip utility is available in many operating systems inside the python-pip or
python3-pip system package. Some variants are available for Python 3: pip3 or specific
versions of Python, for example, Python 3.10: pip3.10.
After successfully installing Ansible on your system, you can use the ansible --version
command to check the current version of Ansible on your system and, of course, also check if
the installation was successful.

35
Chapter 2 Ansible Language Code

The first line of the output of the ansible --version command shows the full
version of Ansible installed:

ansible [core 2.13.4]

In Ansible jargon, the full version reads 2.13.4 with 2 as a major, 13 as a minor, and 4
as the revision.

Getting Started with Ansible

In this section, you learn how to execute a simple Ansible ad hoc command against a list
of hosts.

Running Your First Ansible Ad Hoc Command

Every command is executed by the ansible command-line utility (installed with Ansible
on your system). Try running the Ansible ping command. A command is called a module
in Ansible jargon. The purpose is to test the connection between the controller and the
target node.
The expected result is printing the screen’s message "ping: pong." If you see this
message, the connection was successful, and the execution of the Ansible code went
fine. When the connection is unsuccessful, an error message is displayed onscreen.
The expected output after running the ping command is shown here:

$ ansible localhost -m ping

   localhost | SUCCESS => {
       "changed": false,
       "ping": "pong"
   }

Sometimes, you’ll need to execute the command as a root or administrative user.

Ansible integrates a proper privilege escalation mechanism.
Adding additional parameters to the command line allows you to specify the
connection user devops and indicate that you need root privileges using the sudo
mechanism (default).

ansible localhost -m ping -u devops --become

36
Chapter 2 Ansible Language Code

This command executes under the root user’s permission, supplying the sudo
password. You must verify the configuration and the policies of your system in order
to be able to execute the sudo command via the current user. This is the result of the
execution. Some modules have different results for the normal user and the root user.
The ping module has the same result, but modules that install packages or perform
system-wide configurations have a different result.

Creating a Basic Inventory

Every execution needs a list of target hosts. Ansible needs to know which nodes to
target the execution against. You can specify the list of target hosts in the command line
to an inventory file. The inventory is a simple list of target hosts in the default Ansible
inventory file. If no Ansible inventory is specified in the command line, Ansible relies on
the default /etc/ansible/hosts file. It’s simply a text file with a list of host names or IP
addresses. It follows the INI format.
The inventory file stores the list of target nodes. For example, you can specify only
one host named localhost and the content will look like the following:

localhost anisble_connection=local
web1.example.com ansible_connection=ssh ansible_user=devops ansible_ssh_
private_key_file=~/.ssh/id_rsa

There are two hosts in this inventory. The first is localhost, so you can execute the
automation against the control node and another host web1.example.com using the
username devops connecting via SSH using the key file ~/.ssh/id_rsa.

Ansible Code Language

Ansible is a declarative language, which means that it is focused on the result of the
execution rather than the steps to achieve the results, like in traditional programming
languages. The code is organized into Ansible Playbook when you want to execute more
than one command, like in Ansible Ad Hoc.
Ansible is not a complete computer programming language but relies on multiple
common computer programming statements for each code segment. Every task is
executed in sequence when a condition is verified and in iteration and uses a module
construct. Think about a module like a Lego blueprint for your automation; they could

37
Chapter 2 Ansible Language Code

combine together as needed to achieve great results. One important property of some
modules is idempotency, which means the ability to perform some action on the target
host only if needed and to skip the execution whenever it is not necessary.
Ansible relies on Python language underneath, so it has a vast plethora of libraries
and tools. The Ansible language, called Playbook, is coded in YAML and uses many of
the Jinja2 functions and filters.
Ansible can be extended using plugins that extend and propel the integration to
many cloud providers and more functions, modules, and so on. Code reuse greatly
emphasizes Ansible, and the standard way to distribute and consume the Ansible code is
with Ansible Roles and Ansible Collections.

Ansible Inventory
An Ansible Inventory is the list of managed hosts that target your Ansible automation.
You can also organize your inventory in groups or patterns to select the hosts or groups.
The ansible-inventory command-line tool is included in every Ansible installation,
and its purpose is to show the current Ansible inventory information. It’s advantageous
to verify the current status of your Ansible Inventory. It accepts Ansible Vault and Ansible
Inventory in INI, YAML, and JSON formats. You can get the complete lists of parameters
using --help. A handy feature is to display the hosts in a list using the --list parameter
or in a tree view using the --graph parameter.
The INI inventory is the simplest inventory type. You can insert all your target hosts
or IP addresses. The default location of the inventory is /etc/ansible/hosts, but most
command-line tools accept a customized path with the -i parameter.
Moreover, it’s possible to define multiple groups inside, specifying the group name
between brackets in INI format [group], or under the children node in YAML format.
Obviously, hosts might be present in multiple groups inside an inventory.
The particular keyword all includes all the hosts of the inventory. The only
exception is localhost, which you need to specify. The all keyword is important and
targets all your Ansible example inventory.
One particular host is localhost. You might sometimes want to execute your Ansible
code with the local machine as a target. You can achieve this result by specifying the
ansible_connection “local” property. By default, Ansible assumes it should use the
SSH connection for any target nodes; in this way, you force it to use the local connection
instead.

38
Chapter 2 Ansible Language Code

INI Inventory
The most straightforward Ansible inventory format is INI. The INI format has been
around for years and is very popular for Windows and Linux configuration files. The INI
inventory file can have some name=value fields inside. Obviously, name is the name of
the property, and value is the property’s value. The example in Listing 2-1 shows one
ungrouped host called bastion.example.com and two hosts inside the [web] group
containing web1.example.com and web2.example.com.

Listing 2-1. The example.ini File

bastion.example.com ansible_user=ansible
[web]
web[1:2].example.com
[web:vars]
ansible_user=devops

You can use your target host directly with the hostname or the additional [web]
group. Note that this example specifies some hostnames in a range between brackets,
specifying the first and last elements with an alternation of numbers and letters.
You can also specify some variables for each host. This example scenario is common
because it defines different connections with different hosts. For example, use the local
link for the localhost and ssh, the default, for all the other hosts. For each host, you can
customize the login user. This example uses the ansible user for bastion.example.com
and the devops user for web1.example.com and web2.example.com. You might like to
move the host variable to the host_vars/bastion.example.com file, which contains the
variables for the host bastion.example.com. In a similar way, you can store the group
variables in the group_vars/web file containing the variables for the web group.

YAML Inventory
The YAML format is a simple human-readable format that is often used for Ansible
inventories instead of the INI format. In this format, the indentation of every single line
is essential. The list of hosts or IP addresses descends from the root all node. Each host
must be under the hosts field. You can specify the groups under the children keywords.

39
Chapter 2 Ansible Language Code

The same example-ini inventory could be expressed, as shown in Listing 2-2.

Listing 2-2. The example.yaml File

---
all:
  hosts:
    bastion.example.com:
  vars:
    ansible_user: ansible
  children:
    web:
      hosts:
        web1.example.com
        web2.example.com
      vars:
        ansible_user: devops

The YAML inventory is entirely equivalent to the INI format. As you can see, you can
specify a group name, some hosts underneath, and some host and group variables.

Ansible Playbook
An Ansible Playbook is a set of plays to be executed against an inventory. The code
is written in the YAML text syntax format, making the files easy to write and human-
readable. Each YAML file begins with three dashes (---) and ends with three dots (...).
The three dots (...) are often omitted. Be very careful about the indentation; most
of the time, improper indentation is the cause of malfunction. You can use the sharp/
hashtag symbol (#) for comments, as you can in many languages.
Ansible uses a lot of string variables; you can specify the value directly or with a
single or double quote. We recommend using a double quote as a general rule. Using
the pipe (|), you can define multi-line strings that keep newlines and a major (>) that
doesn’t.

40
Chapter 2 Ansible Language Code

The first Ansible Playbook in this book displays a simple "Hi" message on the screen.
Listing 2-3 shows the complete Ansible code.

Listing 2-3. The hi.yml File

---
- name: Hi example
  hosts: all
  tasks:
    - name: Hi message
      ansible.builtin.debug:
        msg: "Hi!"
...

Here is a line-by-line explanation of the hi.yml Ansible Playbook:

1. Start of the YAML file (---)

2. Name of the play ("Hi example")

3. Target hosts of execution (all of the inventory)

4. Beginning of tasks

5. The first task is named Hi message

6. The ansible.builtin.debug module displays messages onscreen

7. The msg argument of the debug module is the message that you
want to display

8. End of file (...)

You can execute this code using the ansible-playbook command-line utility
included in every Ansible installation. The full command requires specifying the
playbook name after it, as shown in Figure 2-6.

$ ansible-playbook hi.yml

41
Chapter 2 Ansible Language Code

Ansible Controller

ansible-playbook
Type 1 native Hypervisor
connection target host
inventory

Guest operating
system play book

Figure 2-6. The ansible-playbook process

The execution produces the “Hi” message on the screen and an OK status for the
execution. The Ansible playbook could contain multiple plays targeting different hosts,
for example, and executing different tasks.
A good security practice is to execute as much code as possible as standard user
permission. However, some operations are only possible at the root/administrative
permision level. Ansible has embedded the ability to switch between two users (usually
the normal and administrative users) and supports many of the most popular systems
for privilege escalation.
You can enable privilege escalation in Ansible Playbook using the following Ansible
statements:

• become specifies that privilege escalation is necessary

• become_method selects the escalation method

• become_user specifies the destination user (default root)

The Ansible playbook shown in Listing 2-4 installs the popular fail2ban program,
accessing the target host as a normal user then switching to root.

Listing 2-4. The fail2ban.yml File

---
- name: install fail2ban
  hosts: all
  become: true
  become_method: sudo
  become_user: root

42
Chapter 2 Ansible Language Code

  tasks:
    - name: install fail2ban
      ansible.builtin.apt:
        name: fail2ban
        status: present
        update_cache: true

The --check parameter of the ansible-playbook command allows you to perform a

read-only operation on the target node, often used as a dry run for this playbook.
Ansible doesn’t have a proper debug system, but you can combine the Ansible debug
module with the verbosity parameter (1-4). The debug module allows you to print text on
the screen during the execution, and you can specify the verbosity parameter of each
message. The verbosity parameter of the ansible-playbook command is expressed
by the number -v for level one, -vv for level two, -vvv for level three, and -vvvv for
level four.

Ansible Variables
Like in every language, variables allow you to make your code reusable in the future.
These are the variables available in the Ansible language:

• User-defined variables

• Extra variables

• Host and group variables (see the previous Ansible Inventory section)

• Registered variables

The user-defined variables are declared with the vars statement followed by the
list of each variable and value. Variable types are the same as in Python: string, number,
Boolean, list, dictionary, and so on. You can retrieve the variable value in any part of your
Ansible playbook by specifying the name between double brackets, like this: {{ name }}.
The extra variable is a way to pass value from the command line. It’s advantageous
when you’re trying to integrate Ansible into your current toolchain or simply want to
parameterize your playbook from the command line. It overrides any value in your
Ansible Playbook specific to the -e parameter of the ansible-playbook command.
Array variables are instrumental when you want to specify some properties that
apply to multiple objects. For example, Listing 2-5 contains some cars and properties.

43
Chapter 2 Ansible Language Code

Listing 2-5. An Array of Variable Snippets

cars:
  bmw:
    model: M1
    fuel: petrol
  tesla:
    model: Model S
    fuel: electric

You can access each property using dot (.) or bracket notation [], like in Python. For
example, cars.bmw.fuel and cars[bmw][fuel] both refer to the value petrol.
The registered variables save the output of any commands inside registered
variables. It is very useful when you want to combine the output of a module with
another module.

Ansible Filters
You can perform some alterations of Ansible variables using the Ansible filters. These
are native functions from the Jinja2 filters. They extend the functionality of the Ansible
language. Ansible filters manipulate data at a variable level. The most common filters are
as follows:

• Assign default mandatory values: {{ variable_name |

default(5) }}

• Make variables optional: {{ variable_name | default(omit) }}

• Assign ternary value: {{ status | ternary('restart',
'continue') }}

• Manage data types: {{ variable_name | items2dict }}

• Format data to JSON and YAML: {{ variable_name | to_json }}

or {{ variable_name | to_nice_yaml }}

• Work with regex: {{ "ansible" | regex_replace('^.', 'A') }}

44
Chapter 2 Ansible Language Code

Conditionals
Like in every programming language, conditional statements give developers the ability
to check a condition and change the behavior of the program accordingly. The Ansible
form is the when statement.
The when statement defines when a task will be executed or not based on the
Boolean result of the following:

• Complex expression using parentheses ()

• Comparison operators ==, >=, <=, !=

• Ansible facts (ansible_facts['os_family'] == "Debian")

The example in Listing 2-6 reloads the ssh service when the variable reload_ssh
is true.

Listing 2-6. The reload_ssh.yml File

---
- name: ssh reload
  hosts: all
  vars:
    reload_ssh: false
  tasks:
    - name: reload ssh
      ansible.builtin.service:
        name: nginx
        state: reloaded
      when: reload_ssh

Running the code produces one skipped status for the task where the statement is
false during the execution. When you switch the variable reload_ssh to true, the task
status is going to be executed.

45
Chapter 2 Ansible Language Code

Handler
The Ansible handler is a special conditional statement that runs operations only when
the first task reports a change in status. Handlers execute only when the previous task
reports a changed status. If the previous task report has an ok status, the handler is not
executed.
In Listing 2-7, the handler code is executed only if a change is detected. Note that the
notify statement mentions the handler’s name to run. This playbook checks the latest
version of the HTTPD web server on all hosts; if an update is available, the yum module
executes the upgrade and displays the message on the screen. If an upgrade is not
necessary, the handler code is not executed.

Listing 2-7. The httpd.yml File

---
- name: handler demo
  hosts: all
  become: true
  tasks:
    - name: latest package
      ansible.builtin.yum:
        name: httpd
        state: latest
      notify: httpd update
  handlers:
    - name: httpd update
      ansible.builtin.debug:
        msg: "Webserver updated!"

Loop
Computers are great for repetitive tasks. Loop statements automate repetitive tasks in
Ansible Playbooks. The loop variable item is available during the iteration to express the
current value in each iteration.

46
Chapter 2 Ansible Language Code

These are the most popular Ansible loop statements:

• loop: The loop statement is added to the task and takes as a value
the list of items over which the task should be iterated.

• with_items statement: Like loop for simple lists, lists of strings, or a

list of hashes/dictionaries. Flatter to list if lists of lists are provided.

• with_file statement: This keyword requires a list of control node

filenames. The loop variable item holds the contents of the file.

• with_sequence statement: This keyword requires parameters to

generate a list of values based on a numeric sequence.

• with_fileglob statement: This statement lists files matching a

pattern. For example, you can specify the *.txt parameter to list
all the files with txt extensions on the target node, as shown in
Listing 2-8.

Listing 2-8. The checkservices.yml File

---
- name: Check services
  hosts: all
  tasks:
  - name: services running
    ansible.builtin.service:
      name: "{{ item }}"
      state: started
    loop:
      - apache2
      - sshd

This example shows you how to iterate through a list of elements and pass as a single
item to the service module to process one at the time using the loop Ansible statement.

47
Chapter 2 Ansible Language Code

Ansible Facts
Variables related to remote systems are called facts. They are powerful because you can
obtain a comprehensive vision of the current host, the operating system, the distribution
used, the IP address, the networking configuration, the storage configuration, and so on.
With Ansible facts, you can use the behavior or state of one system as a configuration on
other systems.
Note that this requires gather_facts be enabled in the Play section (see Listing 2-9).
The gather_facts variable is enabled by default and adds an extra “Fact Gathering” task
at the beginning of every Ansible playbook execution. You can disable using the gather_
facts: false statement when you don’t want to acquire any Ansible facts from target
machines and use them inside Ansible Playbook. It’s a good habit that explicitly declares
the gather_facts variable in your Ansible Playbooks.

Listing 2-9. The facts.yml File

---
- name: print facts
  hosts: all
  gather_facts: true
  tasks:
  - name: print facts
    ansible.builtin.debug:
      var: ansible_facts

The facts.yml playbook prints all the Ansible facts of the target system. The
ansible_date_time fact, as an example, shows the current date and time on the screen.

Ansible Magic Variables

Magic variables are Ansible’s internal variables that come in handy at times. These are
the five most common magic variables:

• hostvars: Accesses variables defined for any host

• groups: Lists all the groups in the inventory

• group_names: Lists which groups are the current host

48
Chapter 2 Ansible Language Code

• inventory_hostname: Hostname configured in the inventory

• ansible_version: Ansible version information

Refer to the Ansible manual for the full list and uses.

Ansible Vault
Ansible Vault allows you to store encrypted variables and files and then use them,
specifying a password in Playbooks or Roles. The AES 256 cipher protects files with
strong encryption in the latest versions of Ansible. You can manage the Ansible vault
using the ansible-vault command in the terminal, which is included in all the Ansible
installations.
Creating a new Ansible Vault is very straightforward:

$ ansible-vault create secret.yml

You can use Ansible Vault in any Ansible Playbook by specifying the --vault-id
@prompt parameter in the command line.

Ansible Templates
The Ansible templates are helpful for applying variable values to configuration files.
Ansible templates work by taking advantage of the Jinja2 programming language, the
Ansible built-in template module.
The example shown in Listing 2-10 populates the hi.txt.j2 template with the value
of the example variable and saves it in the /tmp/hi.txt file.

Listing 2-10. The templates/hi.txt.j2 File

---
- name: template demo
  hosts: all
  vars:
    example: Hi
  tasks:
    - name: apply template
      ansible.builtin.template:
        src: templates/hi.txt.j2
        dest: /tmp/hi.txt
49
Chapter 2 Ansible Language Code

Listing 2-11. The template.yml File

The result is a file with the following contents:

Ansible Plugins
Plugins extend Ansible’s functionality to more services and application domains. Every
Ansible plugin executes on the Ansible Control Node. The full list of Ansible plugin types
include the following:

• action: Executes in the background before the module executes

• become: Extends privilege escalation systems

• cache: Buffers gathered facts or inventory source data

• callback: Extends behaviors when responding to events

• connection: Extends connection possibilities to the target hosts

• docs: Fragments to document common parameters of multiple

plugins or modules

• filter: data manipulation—extracts a value, transforms data

types and formats, performs mathematical calculus, splits and
concatenates strings, inserts dates and times, and so on

• httpapi: Interacts with a remote device’s HTTP-based API

• inventory: Data sources for Ansible target hosts

• lookup: Extension to the Jinja2 templating language

• module: The most common, basic Ansible plugin

• module_utils: Helper to write Ansible modules

• shell: Expands how Ansible executes tasks

• strategy: Controls the flow of play execution

50
Chapter 2 Ansible Language Code

• test: Evaluates expressions and return true or false

• vars: Adds variable data to Ansible

Some plugin types are specific for the Ansible for Network use case:

• cliconf: CLI interface for network devices

• netconf: Interface to network devices

• terminal: Initializes a network device’s SSH shell

Refer to the official “Ansible Working with Plugins” guide for more details. The most
commonly used are the lookup plugins that enable you to extend Jinja2 to access data
from outside sources within your playbooks.
Famous Ansible lookup plugins use cases are as follows:

• Reading from Windows INI style files (ini)

• Reading from CSV files (csvfile)

• Listening for files matching shell expressions (fileglob)

• Reading lines from stdout (lines)

• Generating a random password (password )

• Reading from a UNIX pipe (pipe)

• Returning content from an URL via HTTP or HTTPS (url)

Ansible Roles
The Ansible Role enables code reuse to Ansible, creating a standard format to distribute
playbook code. The standard structure splits the Ansible Playbook statements of Ansible
into folders in the local file system.
You can manage Ansible’s roles using the ansible-galaxy command-line utility
that’s included in every Ansible installation.
First of all, create the role-example Ansible role using the ansible-galaxy
command and the following parameters:
• role, because you want to interact with an Ansible role

• init, because you want to initialize a new role

• role-example, the name of the new role

51
Chapter 2 Ansible Language Code

Here’s the full command to create role-example:

$ ansible-galaxy role init role-example

The following directories were created:

• defaults, the main.yml file in this directory, contains the default

values of role variables that can be overwritten by the user when the
role is used. These variables have low precedence and are intended
to be changed and customized in plays.
• files. This directory contains static files that are referenced by
role tasks.

• handlers. The main.yml file in this directory contains the role’s

handler definitions.

• meta. The main.yml file in this directory contains information about

the role, including author, license, platforms, and optional role
dependencies.

• tasks. The main.yml file in this directory contains the role’s task
definitions.

• templates. This directory contains Jinja2 templates that are

referenced by role tasks.

• tests. This directory can contain an inventory and test.yml

playbook that can be used to test the role.

• vars. The main.yml file in this directory defines the role’s variable
values. Often these variables are used for internal purposes within
the role. These variables have high precedence and are not intended
to be changed when used in a playbook.

Not every role will have all of these directories. The most important file is main.yml
in the tasks directory. The contents of this file is the contents of the tasks of one of the
Ansible playbooks.
In the most straightforward Ansible role, you can print text on the screen. You can
write the following code inside the main.yml file under the tasks directory in the role-
example role, as shown in Listing 2-12.

52
Chapter 2 Ansible Language Code

Listing 2-12. The role.yml File

---
name: role-example demo
ansible.builtin.debug:
msg: "role-example"

You can include your roles in an Ansible Playbook using the roles statement, as
shown in Listing 2-12.

Listing 2-13. The Contents of the tasks/main.yml File

---
- name: role example
  hosts: all
  roles:
    - role-example

As you can see in Listing 2-12, you can take advantage of Ansible code reuse without
knowing the implementation details of the Ansible role.
This concept of sharing code with other developers is the stepstone behind the
Ansible Galaxy. It is a public directory where you can find available Ansible resources.
There is a search engine that quickly finds the proper Ansible role or collection. Every
resource shows the author, the download count, the supported operating systems, and
much more information about usage and internal variables.
You can easily download and use Ansible roles from Ansible Galaxy in two ways:
1. By installing roles manually.

The following command downloads and installs the latest version

of the geerlingguy.redis Ansible role from the Ansible Galaxy
website:

$ ansible-galaxy role install geerlingguy.redis

2. By installing roles via requirements.yml.

You can create a requirements.yml file when you want to install

more than one role or execute all the operations automatically.
See Listing 2-14.

53
Chapter 2 Ansible Language Code

Listing 2-14. The requirements.yml File

---
roles:
- src: geerlingguy.redis

You can execute your requirements.yml role file using this ansible-galaxy
command:

$ ansible-galaxy role install -r roles/requirements.yml

With the following parameters:

• role install parameter, which installs a role

• -r specifies the usage of the requirements.yml file under the roles

2. By installing a collection via requirements.yml.

You can create a requirements.yml file when you want to install

more than one role or execute all the operations automatically
(see Listing 2-15).

Listing 2-15. The requirements.yml File

---
collections:
- name: community.general
source: https://galaxy.ansible.com

You can execute the requirements.yml collection file using the ansible-galaxy
command:

$ ansible-galaxy collection install -r collections/requirements.yml

Include the following parameters:

• collection install parameter installs the collection

• -r specify the use of the requirements.yml file in the collections

Ansible Execution Environment

The latest approach is to create a container for your Ansible code is called Ansible
Execution Environment. This is a solution to a growing problem. The more an
organization adopts Ansible, the more the dependency tree grows, and the more
dependency every piece of code has. Maintaining each piece of code becomes complex.
Moreover, the developer environment might be misaligned with the production
environment and lead to unexpected results.
The main advantage of the Ansible Execution Environment is that developers can
use the same container used locally in production with the same version and library
dependencies.
The ansible-builder and ansible-runner command-line tools help in the creation
and execution of the Ansible Execution Environments (EE).

55
Chapter 2 Ansible Language Code

The ansible-builder command-line tool was added in 2022 to simplify the creation
of Ansible Execution Environments (EE). It is available in the Ansible toolchain package
with the Red Hat Ansible Automation Subscription:

$ dnf install ansible-builder

The ansible-builder command-line tool is also available in the PyPI repository (via
the pip tool):

$ pip install ansible-builder

Imagine that you want to create a custom my_ee Ansible Execution Environment with
all the tools to use Ansible for Amazon Web Services (AWS).
The following code creates an Ansible Execution Environment to run Amazon
Web Services (AWS) collections and dependencies. A collection is a group of Ansible
resources that connect to Amazon Web Services. The Ansible collection to interact with
Amazon Web Services is community.aws and it requires the boto3 and botocore Python
libraries to be installed in the system.
The main file is the execution-environment.yml file and it provides the description
of the steps to build an Ansible Execution Environment using the ansible-builder
command-line tool, as shown in Figure 2-7 and Listing 2-16.

56
Chapter 2 Ansible Language Code

execution-
environment.yml

ansible-builder my_ee

requirements.yml

bindep.txt

Figure 2-7. The ansible-builder process

Listing 2-16. The execution-environment.yml File

---
version: 1
dependencies:
  galaxy: requirements.yml
  python: requirements.txt
  system: bindep.txt
additional_build_steps:
  prepend: |
    RUN pip3 install --upgrade pip setuptools
  append:
    - RUN ls -al /

57
Chapter 2 Ansible Language Code

The Ansible collection dependencies are specified in the requirements.yml file,

parsed by the internal pip command. See Listing 2-17.

Listing 2-17. The requirements.yml File

---
collections:
- name: community.aws

The Python library dependencies are specified in the requirements.txt file, parsed
by the internal pip command. See Listing 2-18.

Listing 2-18. The requirements.txt File

botocore>=1.18.0
boto3>=1.15.0
boto>=2.49.0

Additional operating system dependencies can be specified inside the bindep.

txt file. Note that you can specify different package names for the RPM (Red Hat-like
systems) or DPKG (Debian-like systems). See Listing 2-19.

Listing 2-19. The bindep.txt File

git [platform:rpm]
git [platform:dpkg]

The Ansible Execution Environment is created only when the ansible-builder

command is executed:

$ ansible-builder build -t my_ee -v 3

With these variables:

–– build, to specify the build operation

–– -t my_ee, to specify the desired Execution Environment name,

my_ee in this case
–– -v 3, to specify the verbosity level of the build, in this case, level 3 so
pretty verbose

58
Chapter 2 Ansible Language Code

This tool might require access to the Red Hat Container Registry when you use an
RHEL UBI8 base image with a Red Hat Ansible Automation subscription (username and
password for the Red Hat Portal).

$ podman login registry.redhat.io

Running the ansible-builder command confirms successful execution with the

name of the container:

Successfully tagged localhost/my_ee:latest

If a problem arises, the result shows the specific error.

Under the hood, the ansible-builder tool performs the following steps:

1. Generates a Containerfile (the generalized standard version of the

Dockerfile that works with different Container engines) inside the
context directory.

2. Runs the docker build or podman build in the context to build

the ansible-builder Container and build the final image.

3. Fetches the base image from registry.redhat.io (usually ee–

minimal-rhel8:latest).

4. Downloads the Ansible collection(s) and dependencies specified

in the requirements.yml file using the ansible-galaxy tool.

5. Installs the package(s) and the dependencies specified in the

bindep.txt file using the microdnf package manager in Red Hat
Enterprise Linux.

6. Installs the Python package(s) and the dependencies specified

in the requirements.txt file using the Python Package
Manager (PIP).

7. Builds the final image with only the necessary files.

After executing the ansible-builder tool, you obtain the context directory with all
the building instructions and the my_ee container to execute your Ansible code.
Once the container has been built, you can execute your Ansible Playbook inside the
Ansible Execution Environment using the ansible-runner command (see Figure 2-8).

59
Chapter 2 Ansible Language Code

web1.example
.com

ansible- my_ee
ping.yml web2.example
builder .com

inventory

Figure 2-8. The ansible-runner process

The ansible-runner command-line tool was added in 2022 to simplify the creation
of Ansible Execution Environments (EE). It is available in the Ansible toolchain package
with the Red Hat Ansible Automation Subscription:

$ dnf install ansible-runner

The ansible-builder command-line tool is also available in the PyPI repository (via
the pip tool):

$ pip install ansible-runner

The Ansible Runner enables you to run the Ansible Execution Environment as a
container in the current machine. Use the following command:

$ ansible-runner run -p ping.yml --inventory inventory --container-

image=my_ee .

With these variables:

–– run, to specify the run operation

–– -p ping.yml, to specify the Ansible Playbook to execute

–– --inventory inventory, to specify the inventory name, in this case,

the inventory file

60
Chapter 2 Ansible Language Code

–– --container-image=my_ee, to specify the name of the Ansible

Execution Environment

–– . (dot), to specify the directory of the base path, in this example, the
current directory

The ansible-runner tool mimics the behavior of the ansible-playbook command

and returns the Ansible output onscreen.
Under the hood, ansible-runner runs your Ansible Playbook inside the Ansible
Execution Environment in your favorite Container Engine (Docker or Podman).
The main advantage of using the Ansible Execution Environment is that you can
deploy to a Kubernetes Control Plane and scale the resources as needed.

Key Takeaways
Ansible is a potent computer language for Infrastructure as Code (IaC). This chapter
started with a quick overview of the architecture and configuration for all the operating
systems (macOS, Windows, and Linux in bare metal), as well as virtual machines, private
and public cloud computing, and containers. Ansible includes the expected typed
variables, conditionals, and loops statements, plus some extra features for DevOps use
cases, such as the handlers, facts, host and group, and magic variables. In this chapter,
you about learned the foundation of the language; in the next chapter, you’ll apply your
knowledge to the Container and Kubernetes environments.

61
CHAPTER 3

Ansible for Containers
Managing containers in your organization might be a pretty daunting task. Especially
when you perform manually and deal with repetitive tasks. Sometimes you need to run a
container in your workstation or in your server fleet.
Ansible can simplify this workflow and automate the boring and daunting tasks. You
can also create some new ways to distribute your application in platform-independent
formats. Moreover, a common use case for security is to run your web server as a
container, so the content is secured from any tampering.

Ansible for Containers
Ansible is a great ally to simplify your journey of running containers. In the following
pages, you are going to discover how to simplify the installation of the Docker Container
Engine and how to automate the flatpak and snap in your workstations.
As a Container Engine, Podman is the preferred way on Linux because it has several
advantages compared to Docker, especially because it can run rootless. However,
Podman on Mac requires an additional machine to run your code, and Windows
requires the virtualized Windows System for Linux (WSL).

Install Docker on Linux and Windows

Docker is the preferred Container Engine for many platforms. You can start your Ansible
journey by automating the installation of Docker. This is very useful when you spin up a
new virtual machine. In the following sections of this book, I focus on the Docker Engine,
which is part of running a container.
For your reference, there is also a product called Docker Desktop that has a nice
GUI that runs on most operating systems. Table 3-1 lists the Docker Desktop-supported
platforms.

63
© Luca Berton 2023
L. Berton, Ansible for Kubernetes by Example, https://doi.org/10.1007/978-1-4842-9285-3_3
Chapter 3 Ansible for Containers

Table 3-1. Docker Desktop Supported Platforms

Platform x86_64 / amd64 arm64 / aarch64

Debian Yes
Fedora Yes
Ubuntu Yes
Binaries Yes
Windows Yes
MacOS Yes Yes

This chapter focuses on the Container Engine, the Docker Engine, only the Docker
runtime, called docker-engine, and supported according to Table 3-2.

Table 3-2. Docker Engine Supported Platforms

Platform x86_64 / amd64 arm64 / aarch64 arm32 s390x

CentOS Yes Yes

Debian Yes Yes Yes
Fedora Yes Yes
Raspbian Yes
RHEL (1) Yes
SLES (2) Yes
Ubuntu Yes Yes Yes Yes
Binaries Yes Yes Yes

(1) RHEL: Red Hat Enterprise Linux

(2) SLES: SUSE Linux Enterprise Linux

Most Linux distributions, especially the Enterprise Linux distributions (for example
Red Hat Enterprise Linux or SUSE Linux Enterprise Linux) prefer Podman as the
Container Engine. It usually is included in the system repository or out-of-the-box.

64
Chapter 3 Ansible for Containers

Tip The community.docker.docker connection Ansible connection plugin

turns a Docker container into an Ansible target. Just use the container name as the
entry in the Ansible inventory. The Docker container must be running on the Ansible
controller, but connecting remotely to a Docker engine on another host is possible.
It requires the community.docker Ansible collection and the Docker Python
library be installed.

Install Docker in Debian Linux

The following steps work in all the Debian and Ubuntu derived Linux distributions. You
are trying to automate the following three steps using Ansible:

–– Add a Docker GPG key

–– Add a Docker repository

–– Update apt cache and install Docker

The first step is to download the GPG signature key for the repository. Use the
ansible.builtin.apt_key Ansible module.
This encrypted key verifies the authenticity of the packages and the repository
and guarantees that the software is the same as the Docker releases. For the ansible.
builtin.apt_key Ansible module, you use two parameters: url and state. The
url parameter specifies the URL of the repository GPG signature key, and the state
parameter verifies that it is present in your system after the execution.
The second step is to add the Docker repository to the distribution. It’s an extra
website where apt, your distribution package manager, looks for software. You are going
to use the ansible.builtin.apt_repository Ansible module. For this Ansible module,
you use two parameters: repo and state. The repo parameter specifies the repository
parameters, and the state parameter verifies that it is present in your system after the
execution.
The third step is to update the apt cache for the available packages and install
Docker (docker-ce) using the ansible.builtin.apt Ansible module. For this Ansible
module, you use three parameters: name, state, and update_cache. The name parameter
specifies the package name (Docker in this use case), and the state parameter verifies

65
Chapter 3 Ansible for Containers

that it is present in your system after the execution. Before installing the package, the
update_cache parameter performs an update of the apt-cache to ensure that the latest
version of the package is downloaded.

Tip You can use the Ansible Magic Variable ansible_distribution, which
automatically contains the most current operating system. Possible values are
Alpine, Altlinux, Amazon, Archlinux, ClearLinux, Coreos, Debian, Gentoo, Mandriva,
NA, OpenWrt, OracleLinux, Red Hat, Slackware, SMGL, SUSE, and VMwareESX.

Listing 3-1 shows the full Ansible Playbook code.

Listing 3-1. install_docker_deb.yml

---
- name: install Docker
  hosts: all
  become: true
  tasks:
    - name: Install apt-transport-https
      ansible.builtin.apt:
        name:
          - apt-transport-https
          - ca-certificates
          - lsb-release
          - gnupg
        state: latest
        update_cache: true

- name: Add signing key

      ansible.builtin.apt_key:
        url: "https://download.docker.com/linux/{{ ansible_distribution |
lower }}/gpg"
        state: present

66
Chapter 3 Ansible for Containers

- name: Add repository into sources list

      ansible.builtin.apt_repository:
        repo: "deb [arch={{ ansible_architecture }}] https://download.
docker.com/linux/{{ ansible_distribution | lower }} {{ ansible_
distribution_release }} stable"
        state: present
        filename: docker

- name: Install Docker

      ansible.builtin.apt:
        name:
          - docker-ce
          - docker-ce-cli
          - containerd.io
          - docker-compose-plugin
        state: latest
        update_cache: true

As usual, you can execute the code using the ansible-playbook command-line
utility included in every Ansible installation.
A common way to verify successful installation of Docker is by executing the
following command on the target host:

$ docker run hello-world

When the installation of Docker is successful, it downloads the hello-world image

from the library, downloads the latest tag of it, validates the sha256 digest file, and prints
the onscreen message: Hello from Docker!

Install Docker in Red Hat Linux

The following steps work in all the Fedora or Red Hat derived Linux distributions.
The automation steps are very similar to Debian, but you use different Ansible
modules to handle rpm packages:
–– Add a Docker GPG key

–– Add a Docker repository

–– Update yum cache and install Docker

67
Chapter 3 Ansible for Containers

The first step is to download the GPG signature key for the repository. You are going
to use the ansible.builtin.rpm_key Ansible module. For this Ansible module, you use
two parameters: key and state. The key parameter specifies the URL or the key ID of the
repository GPG signature key, and the state parameter verifies that it is present in your
system after execution. This encrypted key guarantees that the code was not altered and
that it is the same as the Docker releases.
The second step is to add the Docker repository to the distribution. It’s an extra website
where YUM or DNF, your distribution package manager, looks for software. You are going to
use the ansible.builtin.yum_repository Ansible module. For this Ansible module, you
use four parameters: name, baseurl, gpgcheck, and gpgkey. The name parameter specifies
the repository parameters and the baseurl parameter is the URL of it. The gpgcheck
parameter enables the GPG verification with the URL specified in the gpgkey parameter.
The third step is to update the yum cache for the available packages and install
Docker using the ansible.builtin.yum Ansible module.
For this Ansible module, you use three parameters: name, state, and update_cache.
The name parameter specifies the package name (Docker in this use case), and the state
parameter verifies that it is present in your system after the Ansible execution. Before
installing the package, the update_cache updates the yum cache to ensure that the latest
version of the package will be downloaded.
The full Ansible Playbook is shown in Listing 3-2.

Listing 3-2. install_docker_rpm.yml

---
- name: install Docker
  hosts: all
  become: true
  tasks:
    - name: set mydistribution
      ansible.builtin.set_fact:
        mydistribution: "{{ 'rhel' if (ansible_distribution == 'Red Hat
Enterprise Linux') else (ansible_distribution | lower) }}"

- name: Add signing key

      ansible.builtin.rpm_key:
        key: "https://download.docker.com/linux/{{ mydistribution }}/gpg"
        state: present

68
Chapter 3 Ansible for Containers

- name: Add repository into repo.d list

      ansible.builtin.yum_repository:
        name: docker
        description: docker repository
        baseurl: "https://download.docker.com/linux/{{ mydistribution
}}/$releasever/$basearch/stable"
        enabled: true
        gpgcheck: true
        gpgkey: "https://download.docker.com/linux/{{
mydistribution }}/gpg"

- name: Install Docker

      ansible.builtin.yum:
        name:
          - docker-ce
          - docker-ce-cli
          - containerd.io
        state: latest
        update_cache: true

- name: Start Docker

      ansible.builtin.service:
        name: "docker"
        enabled: true
        state: started

In the same way as the previous code, you can execute the Ansible Playbook using
the ansible-playbook command-line utility included in every Ansible installation.
Run the command-line Docker on the target host:

$ docker run hello-world

This command is expected, as in the previous Debian code, to return the same
onscreen message: Hello from Docker!

69
Chapter 3 Ansible for Containers

Install Docker on Windows

You can also automate the installation of the Docker Desktop in your Windows 10 and 11
systems with the Ansible Playbook and the Chocolatey Package Manager. Chocolatey is
the largest archive of packages for Windows; at the time of writing this book, it includes
9583 community maintained packages.
It also provides a package manager utility that uses the NuGet packaging
infrastructure and Windows PowerShell to simplify the process of downloading and
installing software.
This section focuses on the Ansible module called win_chocolatey to automate the
software installation process. Its full name is chocolatey.chocolatey.win_chocolatey,
which means that it is part of the collection distributed directly by chocolatey. The best
part is that if the target system doesn’t have chocolatey installed, it installs the package
manager and then performs the installation.
The parameter list of the win_chocolatey Ansible module is pretty comprehensive,
but these are the most important options for this use case:

• In the name parameter, you specify the package’s name or a list of

packages.

• If you want to install a specific version, you can specify it in the

version parameter.

• The state specifies the action that you want to perform. In this
case, installing is present or to have the latest version and upgrade
eventually with latest.

Listing 3-3 automates the installation of Docker in Windows-like systems with

Ansible Playbook.

Listing 3-3. install_docker_win.yml

---
- name: install Docker
  hosts: all
  become: false
  gather_facts: false
  tasks:
    - name: Install Docker

70
Chapter 3 Ansible for Containers

      chocolatey.chocolatey.win_chocolatey:
        name: "docker-desktop"
        state: present

A successful Windows installation is shown in Figure 3-1.

Figure 3-1. Docker Desktop on Windows installed via Ansible

You can now execute any Docker container in your Windows target host.

Flatpak in Linux
The purpose of container technology is to simplify the development and maintenance of
applications in Linux, making it simple to deliver the application to every Linux user and
distribution. Traditionally, the different libraries between distributions make it difficult
to deliver applications to users, to control the update and the bug report process. Flathub
is the archive where you can find hundreds of applications available for any Linux
operating system.
Flatpak was developed in 2015 by Red Hat, Endless Computers, and Collabora.
At the moment of writing this book, Flatpak runs on 36 distributions, and there are
1889 applications available. The main target of the application is desktop-oriented
applications. Note that some flatpaks are available only for some architecture.

71
Chapter 3 Ansible for Containers

Here, you’ll set up the Podman Desktop Flatpak. The Flatpak installation in a target
system is effortless using the flatpak Ansible module, which manages Flatpaks in the
target system.
The full name is community.general.flatpak; it’s part of the community.general
Ansible collection maintained by the Ansible Community.
Before jumping to the flatpak Ansible module, you need to prepare your system for
managing Flatpak containers.
First of all, you need to ensure that the Flatpak binary is installed correctly in your
system. The flatpak command-line utility is available in all the major distributions in
the system repository. The easiest way to install a package with Ansible is by using the
package module with the name of the package in the name parameter (flatpak in this
case), the state set to present for installation, and latest set to an upgrade to get the
latest version.
Secondly, you need to configure the source of the software, Flathub. This archive is
seen by an additional repository using the flatpak command-line utility.
The Ansible module community.general.flatpak_remote takes care of this
boring task, verifying that the flathub repository is successfully configured for your
workstation.
Thirdly, you can use community.general.flatpak to install your application.
The only required parameter is name, where you specify the Flatpak name. In the
example, io.podman_desktop.PodmanDesktop was obtained from the Flathub website.
The state parameter specifies the action to perform: whether you want to perform
the installation action (the present option) or the remove an action (the absent option).
The method parameter specifies whether you want to install Flatpak system-wide
(system value) or only for the current user (user value).
More advanced parameters are available:

–– remote, which specifies the name of the source of the software; the
default is flathub

–– no_dependencies, which ignores the dependency installation (needs

to be solved manually)

–– executable, the flatpak executable path, which is by default the

Flatpak program

The full Ansible Playbook is shown in Listing 3-4 and Figure 3-2.

72
Chapter 3 Ansible for Containers

Listing 3-4. flatpak.yml

---
- name: flatpak module demo
  hosts: all
  become: true
  tasks:
    - name: flatpak present
      ansible.builtin.package:
        name: flatpak
        state: present
    - name: flathub repo
      community.general.flatpak_remote:
        name: flathub
        state: present
        flatpakrepo_url: https://flathub.org/repo/flathub.flatpakrepo
        method: system
    - name: install Podman Desktop via flatpak
      community.general.flatpak:
        name: io.podman_desktop.PodmanDesktop
        state: present
        method: system

Figure 3-2. Podman Desktop Flatpak on the latest Fedora Desktop

73
Chapter 3 Ansible for Containers

Once a Flatpak is successfully installed in your system, you are going to continue
using the same version. Flatpak doesn’t have an auto-update mechanism.
You can automate the update process of all your Flatpaks in the system using the
Ansible command module. The Ansible command module is used when there is no native
module available; it literally executes your command to the target host. As you might
understand, this is possibly dangerous because you are exiting somehow from the
Ansible sandbox mechanism. Actually, the shell module also executes the command
via a full shell. The difference is that the shell Ansible module supports the full variables
like $HOME and operations like <, >, |, ; and &, whereas the command Ansible module does
not. The Ansible Playbook to update your Flatpak is shown in Listing 3-5.

Listing 3-5. flatpak_update.yml

---
- name: flatpak update demo
  hosts: all
  become: true
  tasks:
    - name: update flatpak(s)
      ansible.builtin.command: "flatpak update --noninteractive"

Snap in Linux
Snap is another Linux distribution format that distributes applications in a container
promoted by Canonical, the company behind Ubuntu. The amount of documentation is
pervasive, so it’s fairly easy to on board your application in a snap.
An exciting feature is the ability to publish your application for Internet of Things
(IoT) use cases. In fact, it supports any class of Linux application, such as desktop
applications, server tools, IoT apps, system services, and printer drivers. The snaps
auto-update automatically. Under the hood, the snapd daemon checks for updates four
times a day using a so-called over-the-air (OTA) refresh mechanism. The significant
advantage is that the old snap is not replaced until the new snap is successfully installed.
The file is distributed using the SquashFS compressed file system, so it has a pretty small
footprint. Snapcraft is a tool for developers to package their applications.

74
Chapter 3 Ansible for Containers

Snap is supported by 41 Linux distributions. One of the biggest benefits is the

adoption of an application sandbox, so applications run isolated, minimizing
security risks.
Applications are available in a global Snap Store, and the application archive is
hosted and managed by Canonical, but it’s free to download. At the moment of writing
this book, there are 7885 snaps currently in the Snap Store, but around 800 appear to be
test or hello-world snaps.
Snap confinement allows setting the degree of isolation from the system. There are
three options—Strict, Classic, and Devmode. Strict is the most restrictive and runs in
complete isolation. Classic is the middle way and allows accessing the system resources
like your operating system. Devmode is intended only for developers.
The community.general.snap Ansible module enables smooth installation of any
Snap application. For example, you can install microk8s, a small, fast, secure, certified
Kubernetes distribution for workstations and appliances.
In Listing 3-6, the Ansible Playbook installs a snap in the Ubuntu distribution.

Listing 3-6. snap_ubuntu.yml

---
- name: snap module demo
  hosts: all
  become: true
  tasks:
    - name: snapd present
      ansible.builtin.apt:
        name: snapd
        state: present
    - name: install microk8s via snap
      community.general.snap:
        name: microk8s
        state: present
        classic: true

You can execute the Ansible Playbook using the ansible-playbook command
included in every Ansible installation.

75
Chapter 3 Ansible for Containers

Installing snap in the Fedora-like Linux operating system requires a bit more effort
because you need to take care of the SquashFS dependency and create the symlink
for the /snap directory. The other part of the code is exactly like in the Ubuntu Ansible
Playbook shown in Listing 3-7.

Listing 3-7. snap_fedora.yml

---
- name: snap module demo
  hosts: all
  become: true
  tasks:
    - name: snapd present
      ansible.builtin.yum:
        name:
          - snapd
          - fuse
          - squashfs-tools
          - squashfuse
          - kernel-modules
        state: present
    - name: symlink /snap
      ansible.builtin.file:
        src: "/var/lib/snapd/snap"
        dest: "/snap"
        state: link
    - name: load squashfs module
      community.general.modprobe:
        name: "squashfs"
        state: present
    - name: install microk8s via snap
      community.general.snap:
        name: microk8s
        state: present

The result of the execution is a full microk8s ready to be used in your Fedora
machine (see Figure 3-3).

76
Chapter 3 Ansible for Containers

Figure 3-3. microk8s in Fedora Desktop

As mentioned, the snap has the auto-update feature so the snapd daemon takes care
of the update and the install mechanism behind the scenes.

Deploy a Web Server in a Container

One of the best characteristics of containers is their ability to separate the application
from the data. Let’s apply this isolation principle to your services. Running, for example,
your webserver in a container lowers the attack surface and enables the service to be
updated smoothly, keeping the data in the file system.
The good news is that Apache is available as a container named httpd in the most
famous Container Registry. It supports all the architectures: amd64, arm32v5, arm32v6,
arm32v7, arm64v8, i386, mips64le, ppc64le, and s390x. At the moment of writing this
book, the standard image is based on Debian Linux 11 “bullseye” (latest or bullseye
tags) or alpine Linux (alpine tag).
If your team prefers to use the web server, Nginx, you can simply switch to the nginx
container image.
Let’s suppose you want to serve files from the /webroot directory.

77
Chapter 3 Ansible for Containers

Apache with Docker for Debian-like Systems

The full process of running the Apache webserver container requires six steps that you
can automate with different Ansible modules:

1. First, you need to install some Python packages and dependencies

using the ansible.builtin.apt Ansible module.

2. Then you need to install the Docker module for Python using the
ansible.builtin.pip Ansible module.

3. Then you need to pull the image for the Docker hub registry using
the community.docker.docker_image Ansible module.

4. Now create the document root with the right permission using the
ansible.builtin.file module.

5. Then create the custom index.html file using the ansible.

builtin.copy Ansible module. You can upgrade this step using
the template module.

6. Finally, you can run the webserver container, setting the right
port and volume settings with the community.docker.docker_
container Ansible module.

This Ansible module requires the community.docker Ansible Collection to be

installed on the system. If it’s not present, you can install it using this command:

$ ansible-galaxy collection install community.docker

The full Ansible Playbook is shown in Listing 3-8.

Listing 3-8. httpd_debian.yml
---
- name: deploy httpd on container
  hosts: all
  become: true
  vars:
    webroot: "/webroot/"
  tasks:
    - name: system packages present

78
Chapter 3 Ansible for Containers

      ansible.builtin.apt:
        name:
          - python3-pip
          - virtualenv
          - python3-setuptools
        state: latest
        update_cache: true
    - name: Docker Module for Python
      ansible.builtin.pip:
        name: docker
    - name: pull image
      community.docker.docker_image:
        name: httpd
        source: pull
        tag: latest
    - name: webroot present
      ansible.builtin.file:
        path: "{{ webroot }}"
        state: directory
    - name: custom index.html
      ansible.builtin.copy:
        dest: "{{ webroot }}index.html"
        content: |
          Custom Web Page
    - name: run httpd container
      community.docker.docker_container:
        name: webserver
        image: httpd
        state: started
        detach: true
        exposed_ports:
          - 80
        ports:
          - 8080:80
        volumes: "{{ webroot }}:/usr/local/apache2/htdocs/"

79
Chapter 3 Ansible for Containers

Apache with Podman for Red Hat-like Systems

You can obtain the same result as the previous scenario using the Podman Container
Engine in Red Hat systems. The full process requires four steps that you can automate
with different Ansible modules:

1. First, you need to verify that podman and its dependencies are
successfully installed on the target system using the ansible.
builtin.yum Ansible module.

2. Second, you need to create the custom index.html file with

the ansible.builtin.copy Ansible module. You can upgrade
this step using the template module. Note that modern Red
Hat systems come with SELinux enabled, so you need to set the
container_share_t SELinux Context; otherwise, the files aren’t
shared between the host and the guest.

3. Thirdly, you need to pull the image for the container hub registry
using the containers.podman.podman_image Ansible module.

4. Finally, you can run the webserver container, setting the right port
and settings using the containers.podman.podman_container
Ansible module.

This module requires additional containers.podman Ansible Collections that you

can install using this command:

$ ansible-galaxy collection install containers.podman

The full Ansible Playbook is shown in Listing 3-9.

Listing 3-9. httpd_redhat.yml
---
- name: deploy httpd container
  hosts: all
  become: true
  gather_facts: false
  vars:
    webroot: "/webroot"

80
Chapter 3 Ansible for Containers

  tasks:
    - name: podman installed
      ansible.builtin.yum:
        name: podman
        state: latest

- name: pull image

      containers.podman.podman_image:
        name: docker.io/library/httpd
        pull: true
        tag: latest

- name: webroot present

      ansible.builtin.file:
        path: "{{ webroot }}"
        state: directory
        owner: "root"
        group: "root"
        mode: '0777'
        setype: "container_share_t"

- name: custom index.html

      ansible.builtin.copy:
        dest: "{{ webroot }}/index.html"
        content: |
          Custom Web Page
        setype: "container_share_t"

- name: run httpd container

      containers.podman.podman_container:
        name: webserver
        image: httpd
        state: started
        detach: true
        expose:
          - 80

81
Chapter 3 Ansible for Containers

        ports:
          - 8080:80
        volume:
          - "{{ webroot }}:/usr/local/apache2/htdocs/:exec"

You can execute your Ansible Playbook using the ansible-playbook command line
included in every Ansible installation. See Figure 3-4.

Figure 3-4. Webserver in a container in Fedora Linux

Tip The containers.podman.podman Ansible connection plugin turns a

Podman container into an Ansible target. Just use the container name as the entry
in the Ansible inventory. The Podman container must be running on the Ansible
Controller. It requires the containers.podman Ansible collection be installed.

Use Vagrant and Packer

Vagrant and Packer are two tools developed by HashiCorp and distributed open-source.
They are designed to simplify your journey with virtual machine images when creating
a complex environment for the development of integration tests. HashiCorp also
developed the container management software called Nomad, which is open-source in
the Cloud Native Computing Foundation (CNCF). The Nomad project is very simple to
install and it’s portable because it has a lightweight server-client design.

82
Chapter 3 Ansible for Containers

Vagrant
Vagrant is a tool for creating environments for quick use, reuse, and cleanup. It interacts
with a hypervisor (VirtualBox, KVM, Hyper-V, libvirt, Docker containers, VMware,
Parallels, and AWS) to quickly create temporary virtual machines from an image.
Vagrant uses a “box” template image to spin up your virtual machine. You can search
for common boxes on the public Vagrant Cloud website. The Vagrant box usually starts
with the minimal software installed. You can define a Vagrantfile to tell Vagrant what
box to use, including CPU, memory, networks, forwarded ports, and shared directories.
Vagrant supports defining multiple machines in the same Vagrantfile, coded in the Ruby
programming language. In a Vagrantfile, you can also execute Ansible as a provisioner
to execute an automation script automatically at the first boot. Vagrant generates
Ansible inventory automatically for all of the defined machines. You can also define
Ansible groups. This allows you to create a realistic integration test environment on your
workstation machine. The Vagrantfile shown in Listing 3-10 creates two virtual machines
(host01 in the group database and host02 in the group web) using the VirtualBox
hypervisor with different CPU and memory resources and provisions them using the
playbook.yaml Ansible Playbook.

Listing 3-10. Vagrantfile

# -- mode: ruby --

# vi: set ft=ruby :
cluster = {
  "host01" => { :ip => "192.168.0.10", :cpus => 2, :mem => 2048 },
  "host02" => { :ip => "192.168.0.11", :cpus => 4, :mem => 1024 }
}
groups = {
  "database" => ["host01"],
  "web" => ["host02"],
}
Vagrant.configure("2") do |config|
  config.vm.box = "ubuntu/jammy64"
  cluster.each do |name, data|
    config.vm.define name do |host|
      host.vm.hostname = name
      host.vm.network "private_network", ip: "#{data[:ip]}"
83
Chapter 3 Ansible for Containers

host.vm.provider :virtualbox do |vb, override|

        vb.cpus = data[:cpus]
        vb.memory = data[:mem]
      end
      host.vm.provision :ansible do |ansible|
        ansible.playbook = "playbook.yaml"
        ansible.groups = groups
      end
    end
  end
end

You can execute the Vagrantfile using the vagrand up command, which interacts
with VirtualBox to create the virtual machine and provision with the Ansible Playbook.

Packer
Packer is a tool for building your virtual machine “box” image from physical or virtual
machines. Packer uses a configuration language called HCL to connect to a provider via
SSH, run a provisioned Ansible, and capture the virtual machine’s state in an Amazon
Machine Image (AMI) or in the Vagrant box image format. It’s useful to build an image
as part of the continuous integration process and deploy your application to a test
environment.
Packer has a Docker builder that can manage Docker Containers and create
container images. It creates a container, runs provisioners, and saves the result as a new
image. You can use Ansible as a provisioner to build a Docker image, as illustrated in
Listing 3-11.

Listing 3-11. example.pkr.hcl

variable "ansible_connection" {
  default = "docker"
}
source "docker" "example" {
  image = "ubuntu:latest"
  commit = true
}

84
Chapter 3 Ansible for Containers

build {
  sources = ["source.docker.example"]
  provisioner "ansible" {
    playbook_file = "playbook.yaml"
  }
}

After you create the Packer HCL configuration file, you need to use the initialize
parameter to read the file and the build parameter to create and push your Docker
image to the registry. The full commands are similar to these:

$ packer init .
$ packer build .

Key Takeaways
Ansible is a great tool for automating any container workload. You can apply automation
to the container and distribute your application in a more secure way. In this chapter,
you saw how to apply automation to a web server using the popular distribution-
independent formats Flatpak and Snap. In the following chapter, you’re going to dive
deep into the Kubernetes world and learn how to automate it with Ansible.

85
CHAPTER 4

Ansible for K8s Tasks

Kubernetes (also known as “K8s”) is an orchestration platform that automates container
management. Deploying, scaling, and rolling service and application updates is
easy. There is a vibrant community around Kubernetes because it’s an open-source
project. Kubernetes was originally developed by Google and is now maintained by the
Cloud Native Computing Foundation (CNCF). The platform enables an ecosystem for
microservices deployment and cloud-native applications and services, a system design
methodology that embraces rapid change, large scale, and resilience. Containers,
microservices, immutable infrastructure, and declarative APIs exemplify this
methodology.
Kubernetes provides a way to deploy and manage containerized applications at scale
across a cluster of servers. It does this by providing a set of APIs and tools that enable
users to define the desired state of their applications (for example, how many replicas
of a container should be running) and then automatically deploy and manage the
containers to meet that desired state.
Kubernetes is designed to be highly available and to provide self-healing capabilities,
meaning that it can automatically recover from failures and ensure that applications are
running as intended. It also provides load balancing, auto-scaling, and rolling updates,
enabling users to easily manage and scale their applications and services.
Kubernetes is commonly used in cloud-native environments and is supported by
a wide range of cloud providers. The major players are Amazon Web Services (AWS),
Microsoft Azure, and Google Cloud Platform (GCP). Kubernetes is also widely used in
on-premises environments and hybrid cloud deployments, as shown in Figure 4-1.

2014 2015: 2016: 2017: 2018:

Google CNCF Mainstream Enterprise KubeCon

Figure 4-1. Timeline of the Kubernetes project

87
© Luca Berton 2023
L. Berton, Ansible for Kubernetes by Example, https://doi.org/10.1007/978-1-4842-9285-3_4
Chapter 4 Ansible for K8s Tasks

Kubernetes was started by Google in 2014. The predecessor was called Borg, a
distributed cluster manager created by Google to manage its fleet automatically. With
the advent of cloud computing, managing the workload manually became nearly
impossible, and applying normal paradigms was also nearly impossible. Docker
Corporation came up with a solution that allowed it to scale up and down very easily
with immediate benefits. The main advantage of Docker is to create a bundle that can be
executed locally, in a remote machine, or in the cloud in the same manner.
Docker has successfully bridged the gap between developers and operators. Docker
solved the problem of packaging the software and making it accessible on different
platforms. Containers can support any programming language: Ruby, Python, Java, C,
C++, and so on.
A lot more was required to manage the workloads, and that was why Kubernetes
was created. It was a disruptive approach at the time that Google innovated the whole
industry. The official project was announced publicly on June 10th, 2014, and it had a
great following and was released on the same day as DockerCon, the biggest event about
containers worldwide. The 1.0 version was for web app-style stateless low-scale systems.
The initial Google development team was pretty small: Tim Hockin, Kelsey Hightower,
Craig McLuckie, Joe Breda, Brian Grant, Xiaohui Chen, and the original manager, Chen
Goldberg.
Mesos had a better scheduling feature, so, at the time, companies like Netflix,
Airbnb, and Apple were looking for a more mature solution with scaling capabilities up
to 10,000 instances. A lot changed from that release and the project grew up to where it
is today. In retrospect, the correlation between Docker and Kubernetes looks inevitable,
but it wasn’t so clear at the time.
The best thing was that Google created the Kubernetes container orchestration
platform and published it as open-source. More companies contributed to the project
than ever before. They created the open-source community behind and established a de
facto standard. The core part is written in the GO programming language.
Kubernetes was built in a way called Promise Theory that ensures that you can
run the container and don’t need to worry about the underlying complexity. Still, it is
completely abstract by the hardware and any possible failure that might happen during
the lifecycle. This means that when a Kubernetes cluster is alive, it can be broken at
any time, but Kubernetes’ job is to make sure that the application is always running
and serving your traffic. Another important technology is the declarative layer on top
of Docker.

88
Chapter 4 Ansible for K8s Tasks

OpenShift was started in 2010 as a platform-as-a-service (PaaS) to run containers.

Google and Red Hat worked together on the open-source project. Clayton Coleman
was the first contributor at the time from Red Hat. The key advantage of Kubernetes
compared to other orchestration technologies (Dockeswarm and Mesos) was that it
came out with a new project built from scratch, with the vision of a platform built for
enterprise users and with plenty of extension possibilities. This vision quickly excited the
engineers, who immediately recognized their workflows’ value.
The most appreciated feature of Kubernetes is a clean API that manages container
workload and works across multiple computers. Once you introduce the network, partial
failure modes and different scheduling constraints are available across the capacity.
The killer application for Kubernetes was Pokémon Go, the popular augmented
reality game developed by Niantic and running on the Google Kubernetes Engine (GKE).
Once they launched in July 2016, they experienced 50 times the amount of load that they
expected. Kubernetes was able to meet the demand even if it was challenging at the time,
pushing the limits of the infrastructure.
At DockerCon 2017, Docker announced the official support of Kubernetes, which
was the beginning of a successful partnership between the two products. After this
announcement, many competitors embraced Kubernetes: Mesos, Docker, and Pivotal.
By the end of 2017, Amazon Web Server launched a Kubernetes product.
That was the official signal that Kubernetes was here to stay in the market. The
major force was probably the vibrant community supporting Kubernetes. The developer
experience of using containers and Docker being very easy attracted new people. It was a
product that went above and beyond the expectations of the user cases.
Increasing velocity, efficiency, portability, scaling, and reducing latency are some
of the great success stories of adopting Kubernetes in your organization. Benefits
include a decline in an increase in production load, increase in traffic without issues,
SLA achievements, MTTR improvements, cost savings, reduction of load times, better
hardware utilization, integration of local data centers and hybrid cloud environments,
huge active customers, a high number of transactions per second, smooth production
releases per day, more COVID-19 tests scheduled, redesign and standardize CI/CD
pipeline, and decreased software development time.

89
Chapter 4 Ansible for K8s Tasks

Kubernetes Objects
The atomic unit to deploy any Kubernetes object is the pod. The pod contains one or
more containers. It is easy to scale pods to scale your application. The pod is executed
in a worker node of Kubernetes. The complete number of worker nodes is called a data
plane because they run your applications.
A Service is a logical way to expose your application and connect a pod without
worrying about the internal IP addresses of pods.
To deploy a pod, you use PodSpec in Kubernetes, a simple YAML text document
(also called YAML manifest) that is a blueprint for Kubernetes to deploy your pod. The
most important instruction is the image specified in the spec area, which defines the
image to use. DockerHub is often the default Container Registry, so you can just specify
ubuntu:22:10. To specify another Container Registry, you need to specify the full URI
and the image name.
The pod is ephemeral, meaning you lose data when the pod is terminated or
destroyed.
Volumes persistently store the data of the pod. Volumes are available to be
connected for any other pod. You specify the volume in the spec area of the PodSpec.
You can mount all the many network storage protocols, for example, NFS, SMB,
CephFS, Lustre, GlusterFS, JuiceFS, and OpenEBS. This is supported by the Container
Storage Interface (CSI) driver for Kubernetes. CSI drivers are available for popular
storage devices (NetApp, Synology, HP, Cisco HyperFlex, Datatom-Infinity, Datera, Dell
EMC (PowerMax, PowerScale, PowerStore, Unity, and VxFlexOS), HPE, HPE Ezmeral,
Huawei Storage, IBM Block Storage, IBM Spectrum Scale, IBM Cloud Block Storage VPC,
Intel PMEM-CSI, HP TrueNAS, and many cloud storage services (AWS Elastic Block
Storage, AWS Elastic File System, AWS FSx for Lustre, Azure Blob, Azure Disk, Azure
File, cloudscale.ch, GCE Persistent Disk, Google Cloud Filestore, Google Cloud Storage,
Hetzner Cloud Volumes, HyperV, oVirt, Tencent Cloud Block Storage, Tencent Cloud File
Storage, Tencent Cloud Object Storage, vSphere, and so on).
The ConfigMap file keeps the configuration in a key-value database.
A Secret is a way to store sensitive information such as usernames, passwords, SSH
keys, and keypairs. Note that Secrets are base64 encoded but not encrypted.
Namespaces within a cluster allow for logical groups of the common resources of an
application and set user role permissions for access resources and set limits. If they’re
not specified, your resources are created under the default namespace.

90
Chapter 4 Ansible for K8s Tasks

Deployment is commonly used to deploy pods without dealing with pods. Inside a
Deployment, you can add a ReplicaSets to specify the expected pod count and ensure
that the amount of pods matches the desired count for deploying your application
or service. You can also set a label for pods to make them easily recognizable. If a
pod terminates for whatever reason, the Deployment spins up another pod replica to
maintain the desired pod count of the ReplicaSets.
ResourceQuota sets the hard requests and limits for CPU or memory resources per
namespace. Each pod resource request and limit is validated against these hard limits.
You can set some special settings using taints, which are settings of the worker node.
Tolerations are on the pod and should match the settings on the worker node.
Affinities specify properties for node scheduling and execution to the control plane.
For example, instance types or availability zones for Amazon Web Services.
The DaemonSet is very similar to Deployment without a ReplicaSet. Log collector is a
popular use case of the DaemonSet.

Control Plane vs Data Plane

Kubernetes nodes is divided into two types: a control plane and a data plane (see
Figure 4-2). The difference is in the nodes used to manage your Kubernetes cluster
(control plane) and the ones used to execute the user requests (data plane).

Worker Node 1
Pod 1 Pod 2 Pod 3
Control Plane
Container Engine
API server
User kubelet kube-proxy
Interface Scheduler
Worker Node 2
CLI Controller-Manager
Pod 1 Pod 2 Pod 3
etcd
Container Engine

kubelet kube-proxy

Figure 4-2. Control plane vs data plane

91
Chapter 4 Ansible for K8s Tasks

The Kubernetes Control Plane has at least two API server nodes and three etcd
nodes. Kubernetes automatically detects and repairs unhealthy control plane nodes in
order to guarantee the continuity of the service. The data plane is a fleet of worker nodes
where you can allocate your Kubernetes resources. The worker nodes are often called
only nodes.
The Controller Manager (running inside the control plane) runs all the time in the
Kubernetes cluster because it maintains the cluster in the desired status. Every time
you send a command to Kubernetes, you express an intent or a wish to the Controller
Manager and check if you are allowed and if there are enough resources to perform the
action. The Cloud Controller translates your wish into a request for each cloud provider.
The scheduler specifies where to allocate resources for your Kubernetes resources. The
commands are performed via restful API.
The cluster can execute your container using a large variety of container runtime
engines. The most commonly used are Docker and Podman, but your setup might vary
according to your use cases.
The kubelet is a vital process that connects directly to the API server of the control
plane nodes in the worker node.
Kubernetes operators enable you to specify custom resource manifests for your
cluster.

kubectl
The kubectl command-line utility interacts with the Kubernetes cluster and Kubernetes
objects. By default, it interacts with the default namespace.
In the following sections, you are going to create a simple nginx deployment and
service to deploy an Nginx web server in your Kubernetes cluster.
The kubeconfig configuration file is the default authentication method to read the
parameters to connect to the Kubernetes cluster. The file is called config and is stored
in the .kube directory under your home directory. The file is a YAML document with
the Kubernetes hostname, endpoint of the Kubernetes API server, certificate authority
data, username, and authentication method. Your Kubernetes administrator team
usually provides this configuration file. More details are in the “Ansible Troubleshooting”
section.

92
Chapter 4 Ansible for K8s Tasks

Note You can also investigate inside your pod running the shell (usually bash for
Linux) using this command:
$ kubectl exec -it nginx-pod -n ansible-examples -- /bin/bash
Where nginx-pod is the pod name and ansible-examples is the namespace
name. After running the command, you obtain a root shell inside the contained
confirmed by the # of the command-line output, similar to this: bash-5.0#

GitOps Continuous Deployment

Developers worldwide recognize the advantage of testing the codebase early as a
DevOps principle. It’s a common practice to launch a test suite to test your code each
time it is sent to the server. Behind the scenes, a pipeline of code is created to launch the
testing at every submission of code (commit) in the SCN repository, usually Git.
A more modern approach uses Pipeline as Code to define deployment pipelines
through source code.
The Pipelines as Code model, like all the “as code” models, has a lot of benefits:

• Version tracking: Changes are tracked and teams can use the current
version or roll back to previous configurations versions

• Audit trails: Transparency is the key advantage for all the

stakeholders, external auditors, and developers; it shows every
change made to the code and by whom.

• Ease of collaboration: The code is the source of truth, all the

stakeholders can suggest changes, improvements, updates, or
deletes.

• Knowledge sharing: Stakeholders share best practices, using

templates and popular design patterns, and share code snippets so
teams can learn from each other.
Continuous Integration (CI) enables the testing of the application in every change
(every commit in Git). It’s a common best practice enabled by using automatic
pipelines triggered by hooks. Errors are detected and corrected as soon as possible, and
increments are small and gradual.

93
Chapter 4 Ansible for K8s Tasks

Organizations can speed up the release process when any new code update or
change is made through the rigorous automated test process.
After a successful Continuous Integration (CI) build, a test suite with a bunch of
integration tests is executed. The next stage of software automation is Continuous
Delivery (CD).
Continuous Delivery (CD) takes care of the release and deployment process. A fully
automated CI/CD workflow enables a more frequent and reliable release process in your
organization by taking advantage of the DevOps methodologies.
The conjunction with the Git repository, Kubernetes enables the GitOps use case. You
can manage Kubernetes manifest files stored inside it. The main use case of the GitOps
approach is to allow systems convergence when updating or installing new applications.

Jenkins
Jenkins Pipeline is the most commonly used open-source option to integrate CI/CD
and GitOps.
Jenkins Pipeline is a suite of plugins that allows a Jenkins user to define a Jenkins job
in a Jenkinsfile. A Jenkinsfile is a text file that defines a Jenkins Pipeline and is checked
into source control. In a Jenkins Pipeline, you can define a series of tasks, called stages,
that are executed in a specific order. For example, you might have a stage that checks
out code from a version control system, another stage that builds the code, and another
stage that runs tests. Each stage consists of one or more steps, which are individual tasks
that are executed in the context of the stage. One of the main advantages of a Jenkins
Pipeline is that it is built for the Jenkins software in the Groovy programming language
with declarative syntax in a Jenkinsfile. The declarative syntax is easier to read and write,
but the scripted syntax provides more flexibility.
You can define the entire build process in a single Jenkinsfile, which can be version-
controlled and shared among team members. This makes it easy to manage and track
changes to the build process over time, and it allows you to automate the entire build
and deployment process.
Developers and system administrators worldwide create Jenkinsfiles to automate their
tasks every day, store them in a source control management system (SCM), and execute
them via a Jenkins server. The Jenkinsfile is the single source of truth (SSOT) about how a
deployment executes. The advantage of using a Jenkins Pipeline is the predictability of the
workflow and the opportunity to share the result with all the relevant stakeholders.

94
Chapter 4 Ansible for K8s Tasks

VMWare Tanzu Application Platform

The VMWare Tanzu Application Platform is a platform-as-a-service (PaaS) solution
that enables organizations to develop, deploy, and manage applications on top of
Kubernetes. It provides a set of tools and services that allow developers to focus on
building their applications rather than worrying about the underlying infrastructure.
Here are some key features of the VMWare Tanzu Application Platform:

• It is created on the Kubernetes open-source platform, which makes

it easy to manage and scale applications in a consistent way across
environments.

• It provides a range of pre-built services and tools, such as databases,

message queues, and monitoring, which can be easily integrated into
applications.

• It includes a range of tools for developing, testing, and deploying

applications, including an integrated development environment
(IDE), a continuous integration/continuous delivery (CI/CD)
platform, and a registry for storing and sharing Docker images.

• It provides a range of security and compliance features, encryption of

data at rest and in transit, and the ability to enforce policies for access
to resources.

Overall, the VMWare Tanzu Application Platform is designed for organizations to

easily build, manage, and deploy applications in a cloud-native environment. Some
organizations use the VMWare Tanzu Application Platform instead of Jenkins for their
deployment.

Set Up Your Laboratory

Creating a laboratory to experiment with the Kubernetes cluster is handy for testing
code, creating a new configuration, or experimenting with an upgrade of the Kubernetes
platform. The control plane coordinates all the activities of your cluster. A minimum of
three nodes is generally recommended for a production Kubernetes cluster. If one goes
down, the redundancy is compromised because the etcd and the control plane instance
are lost. Developers usually add more control plane nodes to mitigate this risk.

95
Chapter 4 Ansible for K8s Tasks

Creating a Kubernetes cluster from scratch, even if possible, is a daunting task, and
there are better ways to save time than reinventing the wheel. Minikube and Kubespray
are convenient software ways to deploy Kubernetes clusters. The Minikube software is
popular for non-production clusters for the laboratory, whereas Kubespray is focused
more on deploying Kubernetes on cloud providers.
You can deploy your Kubernetes cluster on physical or virtual machines. If none of
these is available, cloud providers have plenty of resources available at their fingertips,
often for a free testing trial.

Virtual Machines
In a home laboratory, you can use an old laptop of bare metal or some virtual machines.
For managing virtual machines, there are many alternatives for the hypervisor software
(also known as a virtual machine monitor, VMM, or virtualizer). This platform assigns
physical resources to virtual machines (see Figure 4-3).

Guest operating
system
Guest operating
system Hypervisor

Hypervisor Operating system

Hardware Hardware

Type 1 native Type 2 hosted

Figure 4-3. Type 1 vs type 2 hypervisors

A long list of hypervisor software is available in modern operating systems. The

following names are likely familiar: Microsoft Hyper-V, Oracle Virtualbox, VMWare
ESX(i), KVM, Xen, QEMU, LXC, Proxmox, and Parallels Desktop for Mac.
The procedure varies from software to software but involves the decision of how
many physical resources are assigned to the virtual machine, creating a virtual machine
(“guest” system), and running like a virtual computer. Once the operating system is
installed, you can install all the Kubernetes components. This might be a long process
because you need to manually download and install all the different open-source
components from scratch and match the different versions. This task might be fun for

96
Chapter 4 Ansible for K8s Tasks

some, but difficult for most. The following sections discuss some solutions that help you
spin up a Kubernetes cluster in your laboratory or production without having to spend
months building everything from scratch.

Raspberry Pis
Raspberry Pi is a popular small single-board computer (SBCs) series. Raspberry Pi 4
Model B (see Figure 4-4) is the latest at the time of writing this book.

Figure 4-4. Raspberry Pi 4 Model B Attribution: Michael H. (Laserlicht) /

Wikimedia Commons / CC BY-SA 4.01

It was developed by the Raspberry Pi Foundation in the United Kingdom in

association with Broadcom. The original aim of the project was to teach computer
science in schools and developing countries. The low cost (from $35) boosted the
popularity of the original model. The key advantages are a modular, open design
combined with the low energy consumption of the ARM processor. Some cutting-
edge technologies are available as well: a fast ARM processor, memory from 1 GiB to
4 GiB, fast data transfer by USB 3.0 and Gigabit Ethernet, onboard wireless network
connectivity, and Bluetooth 5.0, with support for two 4K HDMI displays. It is used for
applications such as smart home hubs, game consoles, robot brains, media servers, VPN
servers, home computers, and more.
It’s possible and relatively affordable to buy a bunch of Raspberry Pi minicomputers
and use them to create a home laboratory for experimentation. Nowadays, they are
available in many variants (RAM, eMMC, wireless) that are shipped worldwide. It is
very interesting to use the latest Raspberry Pi Compute Module 4 cluster board for the
clustering experimentation, a stripped-down edition without any connectors available
from $25. Recent years saw a rise in the popularity of some interesting motherboards

1
https://commons.wikimedia.org/wiki/File:Raspberry_Pi_4_Model_B_-_Side.jpg

97
Chapter 4 Ansible for K8s Tasks

that plug in and use the Raspberry Pi Compute Module as a cluster or a server farm.
Jeff Geerling has become popular on YouTube for testing and crowdsourcing cluster
board projects. Installing Kubernetes software on the Raspberry Pi is usually via the SSH
terminal interface. Remember that the Raspberry Pi processor uses the ARM processor
architecture, which might differ from the one use in your main computer. There are
many alternative ways to install Kubernetes on a single Raspberry Pi or a cluster of
Raspberry Pis. It depends on your experience with Kubernetes; among all, the easiest
is MicroK8s. It simply deploys using the microk8s snap with classic confinement and
interacts using the microk8s command-line tool. You can start and stop the cluster,
save batteries, and use the microk8s start and microk8s stop command-line. You
can verify the status using the microk8s status command: running or not running.
You can enable some Kubernetes features, for example, the Kubernetes Dashboard (see
Figure 4-5), using the microk8s enable dashboard command. More features are also
available: rbac, dns, registry, istio, helm, ingress, prometheus, hostpath-storage,
and so on. Once enabled, the Kubernetes Dashboard is available using the microk8s
dashboard-proxy command. Just remember to add the following line to the /boot/
firmware/cmdline.txt file to enable the cgroup on kernel bootstrap, because it is
disabled by default on the ARM platform:

cgroup_enable=memory cgroup_memory=1

Figure 4-5. microk8s Kubernetes Dashboard

98
Chapter 4 Ansible for K8s Tasks

Kubespray
The Kubespray project focuses on deploying the Kubernetes platform on cloud
providers: Amazon Web Services (AWS), Microsoft Azure, Google Cloud Platform
(GCP), OpenStack, VMWare vSphere, Equinix Metal, Oracle Cloud Infrastructure
(experimental), and bare metal. Kubespray is the preferred choice when you want to
deploy a cluster and don’t want to manually compile the entire Kubernetes ecosystem.
Kubespray enables the deployment of a Kubernetes cluster using simple configuration
management powered by Ansible (Playbooks and Inventory), provisioning tools, and
platform knowledge.
The result is the following:

• Cluster high availability

• Modularity (choices between network plugins)

• Support for most popular Linux distributions, including Flatcar

Container Linux by Kinvolk, Debian, Ubuntu LTS (16.04, 18.04, 20.04,
22.04), RHEL/CentOS/Oracle Linux (7, 8, 9), Fedora (CoreOS, 35, 36),
openSUSE (Leap 15/Tumbleweed), Alma Linux/Rocky Linux (8, 9),
Kylin Linux Advanced Server V10, and Amazon Linux 2

• Tested by CI/CD

• Kubernetes version 1.22+, Ansible v2.11+, Jinja 2.11+ and

python-netaddr

The first step of the installation process is to set up the cluster IPs inside the
inventory/mycluster file, as shown in Listing 4-1.

Listing 4-1. Kubespray inventory/mycluster File

node1 ip=10.3.0.1
node2 ip=10.3.0.2
node3 ip=10.3.0.3
[kube_control_plane]
node1
node2
[etcd]

99
Chapter 4 Ansible for K8s Tasks

node1
node2
[kube_node]
node2
node3
[k8s_cluster:children]
kube_node
kube_control_plane

You can customize the variables for the Ansible groups all and k8s_cluster using
these files:

• inventory/mycluster/group_vars/all/all.yml

• inventory/mycluster/group_vars/k8s_cluster/k8s-cluster.yml

Once everything is configured, you can execute your playbook to set up your nodes:

$ ansible-playbook -i inventory/mycluster --become --become-user=root

cluster.yml

The output is long and the execution time depends on the performance of your
hardware. A successful execution ends in a Kubernetes cluster that’s ready to use.

OpenShift Local
OpenShift Local (formerly CodeReady Containers) is an easy and fast way to deploy an
OpenShift cluster for the laboratory in a virtual machine in minutes.
The purpose of the deployment is only local and for a laboratory, for testing
purposes, because it’s meant to be ephemeral. A control plane and the worker node are
embedded in a single-node virtual machine. It supports many operating systems—Red
Hat Enterprise Linux/CentOS x86_64 (7, 8, and 9) and two latest stable Fedora releases,
Windows 10 x86_64 Fall Creators Update (version 1709+), and macOS (BigSug 11+ Intel
Chip x86_64 or Apple Silicon arm64). The minimum resources required are four vCPU,
8 GiB memory, and 35 GiB of storage. A special edition base on Podman container
runtime is designed for Apple Silicon arm64 SoC processors (M1, M1 Pro, M1 Max, M1
Ultra, M2, and so on) and requires two physical CPU cores, 2 GiB of available physical
memory, and 35 GiB of storage space.

100
Chapter 4 Ansible for K8s Tasks

At the moment of writing this book, the latest release is version 2.12, based on using
Red Hat OpenShift Container Platform 4. You should use the full OpenShift installer for
other OpenShift Container Platform use cases, such as cloud, data center, headless, or
multi-developer setups.
It is a stripped-down release of OpenShift with the following current limitations:

• Single node: Control plane and worker node

• Runs in a virtual machine: Different from external networking

• Upgrade path: It doesn’t support an upgrade path other than

downloading an updated, newer virtual machine

• Cluster monitoring operator: It is non-functional and disabled

You can start downloading the crc command-line tool via the official Red Hat
website (https://console.redhat.com/openshift/create/local). On Windows and
macOS, a guided installer is provided with an easy and straightforward process; just
following the workflow, as shown in Figure 4-6.

Figure 4-6. OpenShift Local guided install on macOS

101
Chapter 4 Ansible for K8s Tasks

A successful installation allows you to run the crc command-line utility from your
terminal and move forward with the OpenShift Local setup. You can download the latest
version of the OpenShift Local virtual machine and store it under the current user’s .crc
directory. Then bootstrap with the following command:

$ crc setup

At the moment, the downloaded virtual machine is 3.09 GiB and requires 31
GiB storage once decompressed. After the successful download, you can start your
environment. It creates the local-only *.crc.testing domain to provide an environment
to test your containers using the local IPv4 address range starting with 172.x.y.z.

$ crc start

The first start requires more time because all the initialization utilities need to run
as well as the Kubernetes Operators. You can verify the running status of the OpenShift
Local cluster anytime using the status parameter:

$ crc status

You can access your OpenShift Local cluster via the command line using the oc
command-line utility (similar to kubectl) and using the API endpoint https://api.
crc.testing:6443 or via the web console at https://console-openshift-console.
apps-crc.testing.
Using the oc-env parameter of the oc command with a running cluster, you can
configure the needed variable to interact with the OpenShift Local cluster and the path
of the oc command-line utility. It’s common practice to combine this with the eval
terminal expression:

$ eval $(crc oc-env)

Two users were automatically created during the setup process with different
capabilities:

–– Administrator: username: kubeadmin with an auto-generated

password. For example, zc3Qx-ejEii-KALgK-5pbYU
–– Standard user: username: developer with simple password:
developer

102
Chapter 4 Ansible for K8s Tasks

To log in to your OpenShift Local cluster, just provide the desired username
(-u parameter) and password (-p parameter):

$ oc login -u developer -p developer https://api.crc.testing:6443

Otherwise, you can use the token retrieved from the OpenShift console:

$ oc login --token=sha256~xxxxxx --server=https://api.crc.testing:6443

You can check which user you’re logged in as using this command:

$ oc whoami

The OpenShift Local cluster has also a container image registry inside. With the same
users, you can access the OpenShift Local Dashboard via the web address (shown in
Figure 4-7). It’s at https://console-openshift-console.apps-crc.testing.

Figure 4-7. OpenShift Local console

hetzner-ocp4
When your target cloud provider is Hetzner, you can use the hetzner-ocp4 project
by RedHat-EMEA-SSA-Team (https://github.com/RedHat-EMEA-SSA-Team/
hetzner-ocp4). It deploys and consumes the Red Hat OpenShift Container Platform 4
environment as a sandpit in the CentOS Root Server bare metal servers from Hetzner.
Currently, it supports CentOS Stream 8 and Red Hat Enterprise Linux 8 and 9.

103
Chapter 4 Ansible for K8s Tasks

The tool relies on Ansible to provision, install, start, stop, and delete the Red Hat
OpenShift Container Platform 4 cluster, renew the SSL certificates using the “Let’s
Encrypt” project, and set up the public DNS records. Current tools allow the use of three
out-of-the-box DNS providers (AWS Route53, Cloudflare, DigitalOcean, and GCP DNS)
or setting a few environment variables (Azure, CloudFlare, Gandi, and TransIP).

Create a Cluster with Minikube

The easiest and smallest Kubernetes cluster comprises only one node. Obviously, this
setup is good enough for a workstation laboratory, but not for a production environment,
which requires more nodes to guarantee high availability.
Minikube is a Go implementation of Kubernetes lightweight that creates one virtual
machine on your local machine and deploys a one-node simple cluster. Minikube is
available for the most recent operating systems—Linux, macOS, and Windows.
First of all, you need a fresh install of a modern operating system, for example,
Ubuntu 22.04 LTS, for the official repository. During the setup, the configuration enables
the additional packages openssh and docker. Make sure to enable the Allow Password
Authentication over SSH command in the SSH Setup options. This simplifies the initial
configuration, but feel free to move to the PKI authentication anytime.
You need to install the minikube command-line tool. Once it’s successfully installed,
you can start the minikube by typing the following:

$ minikube start

Underneath the tool, download the latest Kubernetes images and set up a virtual
machine; for example, using VirtualBox with 2 CPUs, 2GiB of memory, and 20 GiB of
storage. Inside the virtual machine, the tool downloads the Minikube ISO, sets up a
virtual IP address, configures the Docker container runtime, downloads the kubelet and
kubeadm, launches the Kubernetes cluster, configures the cluster permission, and verifies
the component health.
You can also start the minikube with more or fewer resources by specifying the
memory parameter with the desired amount of RAM in MiB and the amount of CPUs:

$ minikube start --memory=8192 --cpus 2

104
Chapter 4 Ansible for K8s Tasks

This will yield the output shown in Figure 4-8.

Figure 4-8. minikube start command-line output

Assigning more resources usually results in a faster environment and more space to
experiment as a data plane. Still, you should compromise within the available physical
resources. At least 1 GiB of RAM is recommended to handle operating system/virtual
machine overhead. The minimum amount of memory required by Kubernetes is at least
2048 MiB of memory. Recommended setup requires four CPUs, 10 GiB of memory, and
as much storage as needed.
Alternatively, you can set the memory and CPUs settings as a configuration as
follows:

$ minikube stop
$ minikube config set memory 8192
$ minikube config set cpus 4
$ minikube start

Once started, you can start sending commands via the kubectl command to execute
your Kubernetes Pods, Services, Deployment, and so on.
As you can see, all control planes and workers are running on the same machine.
You can verify using the kubectl command and the get nodes parameter. The output
shows only one node, called minikube, has a Ready status with control-plane,master
roles and the Kubernetes running version.
Moreover, the minikube dashboard command enables the Kubernetes Dashboard
and proxy and returns the URL for the connections.

105
Chapter 4 Ansible for K8s Tasks

The result is a dashboard opened in the current browser, as shown in Figure 4-9.

Figure 4-9. minikube Kubernetes Dashboard

Kubeadm
Kubeadm is a tool to build a Kubernetes cluster following the best-practice “fast paths”
for creating domain knowledge of the clusters’ lifecycle management, self-hosted
layouts, dynamic discovery services, and so on.
It provides a simple kubeadm command-line utility with the following options:
• init, sets up the initial Kubernetes control-plane node.

• join, sets up a Kubernetes additional control plane node and joins it

to the cluster or a worker node.

• upgrade, upgrades the Kubernetes cluster to a newer version.

• reset, reverts any changes by init or join made to the current host.

106
Chapter 4 Ansible for K8s Tasks

K3s Lightweight Kubernetes

K3s is a CNCF sandbox project created in 2020 that was originally developed by Rancher
Labs and packaged as a single binary with a very small footprint (<100 MB). It creates a
Kubernetes cluster that’s simple to install, run, and auto-update from Raspberry Pi to
AWS a1.4x large 32 GB server. It is a good fit for Internet of Things (IoT), edge computing,
and ARM processors. The lightweight storage backend is based on sqlite3 and extensible
with etcd3, MySQL, and PostgreSQL. K3s can run in a container or as a native Linux
service. With Rancher and K3s, you can manage more than 1,000 edge clusters using a
GitOps Continuous Delivery pipeline.
When the cluster is up and running, you can interact with it using simple commands.

$ export KUBECONFIG=/etc/rancher/k3s/k3s.yaml

As usual, you can interact with the cluster using the kubectl command-line utility.
For example, the following command displays all the pods in all the namespaces of the
cluster:

$ kubectl get pods --all-namespaces

Kubernetes Upgrade
Kubernetes has a vibrant community with four releases per year, so expect a new version
every three months. At the moment of writing this book, the latest Kubernetes version is
1.26.1. In Kubernetes jargon, the values are Major 1, Minor 26 Patch 1. Major are versions
with huge, with impacts that break compatibility with past releases. Minor versions
introduce new features and deprecate some APIs. Patch versions are usually just bug-
fixing releases and are automatically applied by the control plane every month. The
Kubernetes lifecycle of version 1.26 was generally available and released on December 9,
2022, entered maintenance mode on December 28, 2023, and will enter its official end-
of-life phase on February 24, 2024. See Figure 4-10.

development generally maintenance end-of-life

available

Figure 4-10. Kubernetes releases lifecycle

107
Chapter 4 Ansible for K8s Tasks

These constant upgrades guarantee bug fixing and the latest codebase and features
on your cluster. You must proceed one minor version at a time. You can’t jump by two
versions. Note that you can’t revert the upgrade process either.
Always plan for an upgrade. Read the documentation, the changelog, and the release
notes carefully to understand if some API items are going to be deprecated. It’s a best
practice to create a checklist and identify which team performs the upgrade. Usually,
security, operation, and infrastructure are involved in this operation. Define, test, and
document upgrade procedures. Have a rollback/fallback plan in place and know how to
process it. You need to assess the impact on your environment and test in Dev and UAT
before promoting to production. Possible minor service interruptions might happen, so
it’s a good idea to notify your organization or your business continuity officer.
If an upgrade is unsatisfactory for your organization and you need the previous
release, you need to start a new version and restore your data. Define what to do if there
are some conflicts between versions of Kubernetes.
You should always begin the upgrade with the control plane, and only when it works
fine, upgrade the data plane (when using managed workgroup). Nodes and add-ons can
be upgraded after a successful upgrade of the control plane. A robust upgrade procedure
identifies which team is responsible and which is tested in the test environment. Have a
Plan B if something is not working as expected. A step-by-step documentation procedure
is a good upgrade strategy to share with colleagues in your organization. A test of the
main API used in your organization is always good for strategies. Identify downstream
dependencies of the operators, services, and addons used in your Kubernetes cluster,
determine the correct order to upgrade components (control plane first, data place
second), and plan for API changes.
The following upgrade process is for a cluster installed using the kubeadm command-
line tool. For Kubernetes clusters installed with other tools, the procedure is similar.
The plan begins with upgrading the control plane nodes. You need to set the cordon
and drain node offline settings one by one, thus preventing scheduling and evicting any
workloads, upgrade the kubelet and kubectl packages, restart the kubelet service, and
bring the uncordon node back online.
After upgrading the control plane nodes, you can upgrade worker nodes. Begin by
upgrading the kubeadm tool, executing the kubeadm upgrade node command, setting
the node in drain mode, upgrading kubelet and kubectl, and restarting the kubectl
process, then you uncordon the node.

108
Chapter 4 Ansible for K8s Tasks

Create a Cluster with kOps

Kubernetes Operations (kOps) deploys a production-grade Kubernetes cluster. It
provides installation, upgrades, and management, and enables you to be specific with
your favorite cloud provider. kOps is tightly integrated with the unique features that your
cloud provider supports, so it can be a better choice if you know that you will only be
using one cloud provider platform in the foreseeable future.
kOps officially supports AWS (Amazon Web Services) and GCE (Google Cloud
Platform); beta support is available for DigitalOcean, Hetzner, and OpenStack, and alpha
support is available for Azure. The initial kOps command-line utility supports Linux,
macOS, and Windows operating systems.
Its significant advantages are as follows:

• Supports dry-runs and automatic idempotency

• YAML manifest-based configuration

• Multi-architecture with ARM64 support

The installation procedure changes based on the target cloud provider platform,
but involves the use of the kops command-line tool with some parameters. The create
cluster parameter actually creates the Kubernetes cluster when the relevant variables are
set for your environment. Some cloud providers require that you set permission to a service
account beforehand. You can customize and update the cluster configuration as well as
validate the configuration of the cluster to ensure your cluster is working as expected.

Configure Ansible for Kubernetes

The good news is that an Ansible collection is designed to interact with any Kubernetes
cluster. Under the hood, all these resources interact with the Kubernetes API and create
Ansible resources through the Python programming language.
You can interact with the Kubernetes cluster using the kubernetes.core Ansible
collection. At the moment of writing this book, these resources are available:

• helm: Add, update, and delete Kubernetes packages using the

package manager Helm

• helm_info: Obtain information about a Helm package deployed into

the Kubernetes cluster

109
Chapter 4 Ansible for K8s Tasks

• helm_plugin: Add, update, and delete Helm plugins

• helm_plugin_info: Obtain information about your Helm plugins

• helm_repository: Add, update, and delete Helm repositories

• helm_template: Render Helm chart templates

• k8s: Interact with the Kubernetes (K8s) objects

• k8s_cluster_info: Obtain Kubernetes clusters, any APIs available,

and their respective versions

• k8s_cp: Copy files and directories to and from a pod

• k8s_drain: Set drain, cordon, or uncordon for a node in the

Kubernetes cluster

• k8s_exec: Execute a command in a pod

• k8s_info: Obtain information about Kubernetes objects

• k8s_json_patch: Apply JSON patch operations to existing

Kubernetes objects

• k8s_log: Fetch logs from a Kubernetes resource

• K8s_rollback: Roll back Kubernetes object deployments and

DaemonSets

• k8s_scale: Set the scale parameter for a deployment, ReplicaSet,

replication controller, or job

• k8s_service: Add, update, and delete services on Kubernetes

• k8s_taint: Set the taint attribute for a node in a Kubernetes cluster

• kubectl: Execute commands in Pods on a Kubernetes cluster

(connection plugin)

• k8s_config_resource_namer: Generate resource names for the given

resource of type ConfigMap, Secret (filter plugin)
• k8s: Fetch containers and services for one or more Kubernetes
clusters and group by cluster name, namespace, namespace_
services, namespace_pods, and labels (inventory plugin)

110
Chapter 4 Ansible for K8s Tasks

• k8s: Provide access to the full range of Kubernetes APIs

(lookup plugin)

• kustomize: Use the kustomization.yaml file to build a set of

Kubernetes resources (lookup plugin)

Let me highlight the kubernetes.core.k8s module, as it is probably the most

important of the Ansible collection.
The kubernetes.core collection is not included in the Ansible Core, which includes
only the ansible.builtin collection. When you try to use a module that’s not present in
your system, you might end up with the following Ansible error:

ERROR! couldn't resolve module/action 'kubernetes.core.k8s'.

You can verify the presence of the kubernetes.core Ansible collection using the
following ansible-galaxy command:

$ ansible-galaxy collection list

This command shows all the Ansible collections in your system with the relative
version. The latest version of the kubernetes.core Ansible collection is 2.3.2 at the
moment of writing this book.
The collection requires Python 3 installed in your system and the kubernetes Python
library. If they are not present, you can install them manually using the PIP Python
package manager. The pip utility is available for the most recent version of Linux, by
installing the package python3-pip. The command for installing the PyYAML jsonpatch
kubernetes Python libraries dependencies is as follows:

$ pip3 install PyYAML jsonpatch kubernetes

When you want to interact with the Helm package manager, you also need the
helm Python library. However, it uses the legacy setup.py install method, which is
discouraged by pip 23.1 and later. You can read more at https://github.com/pypa/pip/
issues/8559; the workaround at the moment for the Helm package is to use the extra
--use-pep517 parameter.

$ pip3 install --use-pep517 helm

When the Python dependencies are satisfied, you can install the Ansible collection
manually or automatically using the ansible-galaxy command-line utility included in
every Ansible installation.

111
Chapter 4 Ansible for K8s Tasks

–– manual: Just type the action (collection install) followed by the

collection name. The full command looks like this:

$ ansible-galaxy collection install kubernetes.core

–– automatic: Require creating a requirements.yml file, a particular

YAML file with the list of all the Ansible collections to install in the
target system. See Listing 4-2.

Listing 4-2. The requirements.yml File

---
collections:
- name: cloud.common
- name: kubernetes.core

You can execute the requirements.yml file by specifying the -r parameter using the
ansible-galaxy command-line tool:

$ ansible-galaxy install -r collections/requirements.yml

After the manual or automatic Ansible collection installation process, your system is
configured to interact with the kubernetes.core Ansible collection.

Tip The kubernetes.core Ansible collection supports the Ansible Turbo

mode. By default, this feature is disabled. It requires an additional cloud.common
Ansible collection installed.
The AnsibleTurboModule class is inherited from the standard AnsibleModule
class, which spawns a little Python daemon, and the module logic runs inside
this Python daemon. The result is a drastic improvement in performance
because Python modules are loaded one time, and Ansible can reuse an existing
authenticated session. The daemon run in a single process exposes a local
UNIX socket for communication and kills itself after 15 seconds. You can set the
ENABLE_TURBO_MODE variable to true or 1 using the environment statement in
the Play section of the Ansible Playbooks.

112
Chapter 4 Ansible for K8s Tasks

Once you have successfully installed the Python and Ansible dependencies, you
need to execute the authentication to your cluster in order to execute API requests via
the command-line tool.
For the Kubernetes cluster, you need the kubectl command-line utility. Many
options are available as binary or by using your favorite package managers. For example,
in macOS, you can download it as a binary, use Macports, or use Homebrew (https://
brew.sh/):

$ brew install kubernetes-cli

The kubectl Linux package is available in the DEB and RPM packages format or as a
binary. In Windows, you can download and install the kubectl binary from the website
or use the Chocolatey, Scoop, or “winget” package manager.
For the OpenShift cluster, the oc command-line utility is needed instead. In macOS
with Homebrew, you can find in the package:

$ brew install openshift-cli

You don’t need any modifications in the Ansible configuration file ansible.cfg to
use Ansible with Kubernetes. This file allows you to fine-tune some settings and enables
custom Ansible plugins. You can also customize directory paths in your system.

Ansible Troubleshooting
Ansible troubleshooting can be difficult because more often than not, the fatal error root
cause is related to a Kubernetes or Ansible configuration. The following errors might
appear when you try to configure your environment to interact with a Kubernetes or
OpenShift cluster.

401 unauthorized
Kubernetes authenticates the API requests using bearer tokens, client certificates, or an
authentication proxy through authentication plugins.
You might obtain the Kubernetes 401 unauthorized fatal error message when you try
to access your cluster when trying to execute some code without an authorization token
or using an invalid token.

113
Chapter 4 Ansible for K8s Tasks

The root cause is related to Kubernetes authentication and not to Ansible Playbook
or Ansible configuration. It could be because you are not authorized to access the
namespace or resource or simply because you are not authenticated. Bear in mind
that the authentication token expires after a specified amount of time, so you need to
authenticate again in your Kubernetes cluster.
The full fatal error message looks similar to the following:

fatal: [localhost]: FAILED! => {"changed": false, "error": 401, "msg":

"Namespace example: Failed to retrieve requested object: b'{\"kind\":\"St
atus\",\"apiVersion\":\"v1\",\"metadata\":{},\"status\":\"Failure\",\"me
ssage\":\"Unauthorized\",\"reason\":\"Unauthorized\",\"code\":401}\\n'",
"reason": "Unauthorized", "status": 401}

Kubernetes
In a Kubernetes cluster, obtaining a valid authentication token requires the setup of a
context for your cluster with the kubectrl command-line utility.
You can obtain a valid Kubernetes authentication token by first entering the
username and password and saving them as credentials using the following command:

$ kubectrl config set-credentials developer/foo.example.com

–-username=developer --password=developer

With:

• kubectrl: Specifies the command-line tool to connect to the

Kubernetes cluster

• config: Specifies the action to configure the cluster

• set-credentials developer/foo.example.com Creates a new

credential named developer/foo.example.com (I prefer to use the
same cluster name, but you can specify any name)

• –-username=developer: Specifies the username for the login

• --password=developer: Specifies the password for the login

114
Chapter 4 Ansible for K8s Tasks

Then you need to configure the cluster connection server:

$ kubectl config set-cluster foo.example.com –-insecure-skip-tls-

verify=true –-server=https://foo.example.com

With:

• kubectrl: Specifies the command-line tool to connect to the

Kubernetes cluster

• config: Specifies the action to configure the cluster

• set-cluster foo.example.com: Creates a new cluster named foo.

example.com (you can specify any name)

• –-insecure-skip-tls-verify=true: Skips TLS certificates validation

• –-server=https://foo.example.com: Specifies the cluster URI

Then you can join in a context the cluster and credentials information:

$ kubectl config set-context default/foo.example.com/developer

–-user=developer/foo.example.com –-namespace=default –-cluster=foo.example.com

With:

• kubectrl: Specifies the command-line tool to connect to the

Kubernetes cluster

• config: Specifies the action to configure the cluster

• set-context default/foo.example.com/developer: Creates a new

context named default/foo.example.com/developer (you can
specify any name)

• –-user=developer/foo.example.com: Skips TLS certificates

validation

• ––-namespace=default: Specifies the default namespace (in this

case, default)
• –-cluster=foo.example.com: Specifies the cluster configuration name
Than you can finally use this context and obtain the authentication token:

$ kubectl config use-context default/foo.example.com/developer

115
Chapter 4 Ansible for K8s Tasks

From now on, all the commands are sent from within this context to the Kubernetes
cluster that you configured.

OpenShift
You can obtain the OpenShift cluster using the oc command and specifying the
username and password:

$ oc login -u developer -p developer https://api.openshift.example.com:6443

With:

• oc: Specifies the command-line tool to interact with OpenShift cluster

• login: Specifies the action to authenticate in the cluster

• -u developer: Specifies the username for the login

• -p developer: Specifies the password for the login

• https://api.openshift.example.com:6443: Specifies the cluster

URI and the port

As an alternative to a username and password, you can authenticate using a token

obtained via the OpenShift console. After successfully logging in to the web user
interface, just choose Copy Login Command under your username in the top-right
menu area, as shown in Figure 4-11.

Figure 4-11. OpenShift console token generation

The token is a SHA256 series of alphanumeric characters displayed in a new window

on the screen, as shown in Figure 4-12.

116
Chapter 4 Ansible for K8s Tasks

Figure 4-12. OpenShift console token example

Once generated, the token is usable via the command-line oc command like the
following:

$ oc login --token=sha256~xxxxxx --server=https://api.openshift.example.

com:6443

With:

• oc: Specifies the command-line tool to interact with OpenShift cluster

• login: Specifies the action to authenticate in the cluster

• --token=sha256~xxxxxx: Specifies the usage of the access token

• https://api.openshift.example.com:6443: Specifies the cluster

URI and the port
After successfully authenticating, an authentication token is stored in the .kube
directory under the config file on the local machine. You can customize the filename
and path using the KUBECONFIG environment variable in the terminal.

$ export KUBECONFIG=/home/devops/.kube/config-developer

In the same way, you should instruct Ansible to read the environment variable by
adding the following environment lines in the Play area of your Ansible Playbook:

[...]
environment:
KUBECONFIG: "/home/devops/.kube/config-developer"
[...]

117
Chapter 4 Ansible for K8s Tasks

It is not possible to use the ~ character to refer to the UNIX user’s home directory as
is treated as a plain character and looking for a directory named ~ instead.

x509 error
Another common error is when the certification authority is not shared between the
client and the server. You will receive the following onscreen error when executing the
following command:

$ oc login -u developer -p developer --server=https://api.openshift.

example.com:6443

The parameters were discussed in the previous section. The fatal error on the screen
is as follows:

error: x509: "kube-apiserver-service-network-signer" certificate is

not trusted

You can solve adding the additional certification authority file with the
--certificate-authority parameter followed by the path on disk of the certificate.
You can acquire more information by increasing the verbosity level to six, the highest
available, with the --loglevel 6 parameter.
A successful login procedure can be verified using this OpenShift command:

$ oc whoami
developer

kubeconfig
The kubectl command-line utility and the kubernetes.core Ansible collection both
require a valid kubeconfig file with the authentication information to connect to the
Kubernetes cluster. This book assumes that you have a valid kubeconfig file to connect
to your Kubernetes cluster. The file contains the following key information:

• cluster-ca-cert: The Kubernetes cluster certification authority

• endpoint: Kubernetes cluster endpoint (IP or DNS of the
master node)

• cluster-name: K8s cluster name

118
Chapter 4 Ansible for K8s Tasks

• cluster-username: Username or service account

• secret-token: A secret token of the username of the service account

This is a schematic of a typical kubeconfig file:

---
apiVersion: v1
kind: Config
current-context:  <cluster-name>
clusters:
- name: <cluster-name>
  cluster:
    certificate-authority-data: <cluster-ca-cert>
    server: <endpoint>
contexts:
- name:  <cluster-name>
  context:
    cluster:  <cluster-name>
    user:  <cluster-username>
preferences: {}
users:
- name:  <cluster-username>
  user:
    token: <secret-token>

The default path of this file is in the .kube user directory and is named config.
You can customize the path of the kubectl command-line utility using the
KUBECONFIG environment variable as follows:

$ export KUBECONFIG=$HOME/.kube/config

For Ansible code, you can specify a custom path by specifying the kubeconfig
module parameter or via the K8S_AUTH_KUBECONFIG environment variable. The code in
this book assumes that your kubeconfig file is stored under the ~/.kube/config path

119
Chapter 4 Ansible for K8s Tasks

(the default path), where the ~ (tilde symbol) indicates the home of your user. You can
customize this path using the kubeconfig parameter for each Ansible module. The most
common usage with the k8s Ansible module is as follows:

kubernetes.core.k8s:
kubeconfig: '~/.kube/config'

Configure a Python Virtual Environment

A Python Virtual Environment is a tool to keep your application environment separate
from the system-wide environment. Ansible, like all Python applications, requires some
Python libraries as dependencies. Maintaining up-to-date all the Python dependencies
of Ansible and of the Ansible collection without interfering with your Linux system can
be challenging without a Python Virtual Environment.
The first step is to create the virtual environment using the Python venv module,
preceded by the -m parameter. In the following example, the virtual environment is
named venv; feel free to customize it according to your needs:

$ python3 -m venv venv

After creating the venv virtual environment, you need to activate it using the script
for your operating system activate for Linux/macOS e activate.bat for Windows. You
can find the script in the directory named like your virtual environment (venv). You can
return every time to this state using the activate script.

$ source venv/bin/activate

Once inside the venv virtual environment, prepend the prompt to the name (venv).
It’s a best practice to upgrade the package manager pip and setuptools to the latest
versions:

(venv) $ pip3 install --upgrade pip setuptools

On some systems, the pip/pip3 command tool is not available out of the box.
It might be necessary to install the python3-pip package or use following a manual
installation process.

$ python -m ensurepip --upgrade

120
Chapter 4 Ansible for K8s Tasks

After upgrading the pip installed, you can install the Python PyYAML, jsonpatch and
kubernetes dependency:

(venv) $ pip3 install PyYAML jsonpatch kubernetes

For Ansible modules that interact with the Helm package manager, the extra helm
Python library is required with the --use-pep517 parameter workaround:

(venv) $ pip3 install --use-pep517 helm

You can save all the Python libraries and their dependencies using the pip3 freeze
command and save the result to requirements.txt. Sharing this file with someone else
or another system will result in the same configuration of this virtual environment.

(venv) $ pip3 freeze > requirements.txt

Listing 4-3 shows the contents of the requirements.txt at the time of writing this
book; it includes the ansible-builder and ansible-runner utilities.

Listing 4-3. The requirements.txt File

ansible-builder==1.2.0
ansible-runner==2.3.1
bindep==2.11.0
cachetools==5.2.0
certifi==2022.12.7
charset-normalizer==2.1.1
distro==1.8.0
docutils==0.19
google-auth==2.15.0
helm==0.4
idna==3.4
kubernetes==25.3.0
lockfile==0.12.2
oauthlib==3.2.2
packaging==22.0
Parsley==1.3
pbr==5.11.0
pexpect==4.8.0

121
Chapter 4 Ansible for K8s Tasks

ptyprocess==0.7.0
pyasn1==0.4.8
pyasn1-modules==0.2.8
python-daemon==2.3.2
python-dateutil==2.8.2
PyYAML==6.0
requests==2.28.1
requests-oauthlib==1.3.1
requirements-parser==0.5.0
rsa==4.9
six==1.16.0
types-setuptools==65.6.0.2
urllib3==1.26.13
websocket-client==1.4.2

If you want to use the Ansible Turbo feature, you must install the cloud.common
Ansible collection using the ansible-galaxy command-line utility:

(venv) $ ansible-galaxy collection install cloud.common

The kubernetes.core Ansible collection is a must-have in your system to execute

Kubernetes automation:

(venv) $ ansible-galaxy collection install kubernetes.core

You can deactivate your virtual environment at any time and return to your system
console. You should activate the virtual environment every time you want to use it. The
virtual environment preserves all installed libraries and collections.

$ deactivate

When you want to use your virtual environment, just execute the activate
script again.
When the complexity of the Python Virtual Environment grows too much, it can
become challenging to manage the Python library dependencies and Ansible module
dependencies. You can use the Ansible Execution Environment instead.

122
Chapter 4 Ansible for K8s Tasks

Configure an Ansible Execution Environment

The Ansible Execution Environment is a way to containerize the execution of the Ansible
invented by Red Hat. It maintains archives of the separation of the operating system
dependencies, Python dependencies, and Ansible collections without interfering with
your Linux system. It’s the evolution of Python Virtual Environment and can be executed
natively in a Kubernetes cluster. It supersedes manual Python Virtual Environments,
Ansible module dependencies, and Ansible Tower bubblewrap.
Ansible Execution Environment relies on the Ansible Builder, the ansible-builder
command, to create an Ansible Execution Environment. Ansible Builder produces a
directory with the build context for the container image. It contains the Containerfile,
along with any other files that need to be added to the image.
The Ansible Runner, the ansible-runner command utility, executes the Ansible
Playbook in an Ansible Execution Environment. The Ansible Runner enables you to run
the Ansible Execution Environment as a container in the current machine.
You can install the ansible-builder, and ansible-runner command-line utilities
via the DNF package manager (Fedora, CentOS, and Red Hat Enterprise Linux) or via the
pip package manager:

$ pip3 install ansible-builder

$ pip3 install ansible-runner

Listing 4-4 creates a custom Ansible Execution Environment for Kubernetes called
my_kube. The execution-environment.yml file contains the blueprint of the newest
Execution Environment, specifying that you want to customize the operating system
packages (system: bindep.txt), the Python libraries (python: requirements.txt),
and the Ansible collections (galaxy: requirements.yml). See Listing 4-4. You’re going
to see these files one by one. Before the build step, you install and upgrade the pip and
setuptools, and afterwards, you list the file in the root directory. The building process
will not work in disconnected or “air-gapped” environments. Ensure that your systems
are connected to the Internet because the process must download some resources in
order to build the Ansible Execution Environment.

123
Chapter 4 Ansible for K8s Tasks

Listing 4-4. The execution-environment.yml File

---
version: 1
dependencies:
  galaxy: requirements.yml
  python: requirements.txt
  system: bindep.txt
additional_build_steps:
  prepend: |
                RUN pip3 install --upgrade pip setuptools
  append:
    - RUN ls -al /

The append part is an optional additional step that demonstrates that you can run
commands after building your Ansible Execution Environment.
As demonstrated in Listing 4-5, the requirements.yml file specifies the list of Ansible
collections to download. In this case, cloud.common for Ansible Turbo and kubernetes.
core for the Ansible resources for Kubernetes.

Listing 4-5. The requirements.yml File

---
collections:
- name: cloud.common
- name: kubernetes.core

The requirements.txt file specifies the Python library dependencies, in that case,
the JSONpatch, kubernetes, and PyYAML Python libraries, as shown in Listing 4-6.

Listing 4-6. The requirements.txt File

kubernetes>=12.0.0
PyYAML>=3.11
jsonpatch

You can also specify some operating system package files, in this case, the git Source
Code Management package for using the rpm and dpkg package installation methods, as
shown in Listing 4-7.

124
Chapter 4 Ansible for K8s Tasks

Listing 4-7. The bindep.txt File

git [platform:rpm]
git [platform:dpkg]

You can start building the my_kube Ansible Execution Environment using the
ansible-builder command and specifying the build parameter with the name of the
newest Ansible Execution Environment after the -t parameter. Optionally, you can
obtain the output in verbose mode using the -v 3 parameter. The full command is like
the following:

$ ansible-builder build -t my_kube -v 3

The command requires connecting to the Container Registry to download the base
image and the Ansible Builder image. When you use it in an enterprise environment,
type the following command to authenticate to the Red Hat Container Registry with your
Red Hat Portal username and password:

$ podman login registry.redhat.io

Once the Ansible Builder has completed the execution successfully, a context/
Containerfile file is created. You can build your image using this command:

$ podman build -f context/Containerfile -t my_kube context

The Ansible Runner enables you to run the Ansible Execution Environment as
a container in the current machine. The syntax resembles the ansible-playbook
command-line utility, specifying the name of the Ansible Execution Environment:

$ ansible-runner run -p ping.yml --inventory inventory --container-

image=my_kube .

This uses the simplest Ansible INI inventory file against the localhost and using
the local connection (instead of the SSH connection):

localhost ansible_connection=local

The Ansible Playbook shown in Listing 4-8 uses the popular ansible.builtin.ping
Ansible module to test the successful connection to the target node. The ping module
executes a connection to the target node and executes some sample Python code. In this
way, you can test the connection and the ability of the Python interpreter to listen and
execute code on the target node.
125
Chapter 4 Ansible for K8s Tasks

Listing 4-8. The ping.yml Ansible Playbook File

---
- name: test
  hosts: all
  tasks:
    - name: test connection
      ansible.builtin.ping:

In this section, you learned how to create a custom my_kube Ansible Execution
Environment with the Ansible Builder and how to execute an Ansible Paybook inside of
it using the Ansible Runner command-line tools. Since it’s a container, you can deploy
your Ansible Environment in your Container Registry. You can execute your Ansible
Playbook scripts in the local machine or in a Kubernetes cluster.

Tip The Ansible Automation Platform easily manages the Ansible Execution

Environments and the execution against any Kubernetes and OpenShift cluster.
Under the Ansible Controller, you can create a container group specifying a
namespace, a service account with roles that allow to launch and manage
pods in the namespace, the OpenShift or Kubernetes bearer token, and the SSL
certification authority certificate associated with the cluster.

Create a Namespace
You can automate the creation of the ansible-examples namespace in Kubernetes
or project in OpenShift. You just need a few parameters for the kubernetes.core.k8s
Ansible module. The k8s Ansible module is part of the kubernetes.core collection; it’s
very powerful and allows you to interact with the Kubernetes API like the kubectl or
oc command-line utilities. The name parameter specifies the namespace that you want
to create or verify if present. The api_version parameter specifies the Kubernetes API
version; the default is v1 for version 1. The kind parameter specifies an object model,
which is namespace for this use case. As with other modules, the state parameter
determines if an object should be created (the - present option), updated (the - patched
option), or deleted (the - absent option). You can customize the myproject variable with
your namespace in the code or via an Ansible extra variable.
126
Chapter 4 Ansible for K8s Tasks

The full Ansible Playbook is shown in Listing 4-9.

Listing 4-9. The ns_create.yml Ansible Playbook File

---
- name: k8s ns
  hosts: all
  vars:
    myproject: "ansible-examples"
  tasks:
    - name: namespace present
      kubernetes.core.k8s:
        api_version: v1
        kind: Namespace
        name: "{{ myproject }}"
        state: present

You are going to execute the playbook in the current Ansible Controller so you can
use the localhost in the inventory with a local connection, as shown in Listing 4-10.

Listing 4-10. The Inventory File

localhost ansible_connection=local

The code executes the ansible-playbook command. The full command is as

follows:

$ ansible-playbook -i inventory ns_create.yml

A successful execution output includes the following:

• Managed host: localhost

• Play recap status: ok=2 changed=1

• Task status:

TASK [namespace ansible-examples present]

changed: [localhost]

127
Chapter 4 Ansible for K8s Tasks

You receive the changed status when the namespace is successfully created in the
Kubernetes cluster, whereas you receive the ok status when the namespace was already
present in the cluster (Ansible idempotency property).
You can verify that the namespace list uses the kubectl command-line utility:

$ kubectl get namespace | grep ansible-examples

ansible-examples Active 10m

Instead, OpenShift calls namespace as a project but the result is similar to the
previous:

$ oc projects | grep ansible-examples

ansible-examples

The previous Ansible Playbook supposes that you perform the authentication in the
Ansible Controller outside the Ansible Playbook using the kubectl for Kubernetes or
oc for OpenShift command. It’s also possible to perform the authentication step in the
Ansible Playbook using the k8s_auth Ansible module, which is part of the kubernetes.
core Ansible collection. If you want to execute the authentication directly in the Ansible
Playbook, you need to add a task before executing any command to the cluster. The
get access token task connects to the host address with the specified username and
password credentials and saves the result in the k8s_auth_results runtime variable.
The variable contains the token under the k8s_auth.api_key variable. Every time you
use the k8s module, the extra parameter api_key is specified with the value of the token
registered from the previous task. Listing 4-11 shows the full Ansible Playbook.

Listing 4-11. The ns_create_auth.yml Ansible Playbook File

---
- name: k8s ns
  hosts: all
  vars:
    myproject: "ansible-examples"
    k8s_username: "kubeadmin"
    k8s_password: "password"
    k8s_host: "https://api.k8s:6443"
    k8s_validate: true

128
Chapter 4 Ansible for K8s Tasks

  tasks:
    - name: get access token
      kubernetes.core.k8s_auth:
        username: "{{ k8s_username }}"
        password: "{{ k8s_password }}"
        host: "{{ k8s_host }}"
        validate_certs: "{{ k8s_validate }}"
      register: k8s_auth_results

- name: namespace present

      kubernetes.core.k8s:
        api_key: "{{ k8s_auth_results.k8s_auth.api_key }}"
        api_version: v1
        kind: Namespace
        name: "{{ myproject }}"
        state: present

In the real world, it’s better to store any variables with sensitive data (beginning
with k8s_ of the previous Ansible Playbook) in an Ansible vault that’s encrypted and
protected. For OpenShift authentication, use the community.okd.openshift_auth
Ansible module in the first task.

Report Namespaces
Kubernetes cluster administrators often need to list all the namespaces (projects
in OpenShift jargon). Reporting is a simple, repetitive, and boring task that can be
automated with Ansible. This can be easily achieved using Ansible Playbook with
two tasks. The first task uses the Ansible query lookup plugin and invokes the Ansible
kubernetes.core.k8s module to interact with the Kubernetes API and return a list of
all the namespaces in the cluster. The result of the Ansible lookup plugin is saved in the
projects runtime variable using the ansible.builtin.set_fact Ansible module. You
can use this variable to perform further tasks, such as print the result onscreen using
the ansible.builtin.debug Ansible module. Listing 4-12 shows the full ns_list.yml
Ansible Playbook file.

129
Chapter 4 Ansible for K8s Tasks

Listing 4-12. The ns_list.yml Ansible Playbook File

---
- name: k8s ns
  hosts: all
  tasks:
    - name: list all namespaces
      ansible.builtin.set_fact:
        projects: "{{ query('kubernetes.core.k8s', api_version='v1',
kind='Namespace') }}"

- name: display the result

ansible.builtin.debug:
msg: "{{ projects }}"

The easier execution is locally in the Ansible Controller, specifying the localhost in
the inventory, as shown in Listing 4-13.

Listing 4-13. The Inventory File

localhost ansible_connection=local

The Ansible code can be executed using the ansible-playbook command included
in every Ansible installation. A successful execution output includes the following:

• Managed host: localhost

• Play recap status: ok=3 changed=0

• Task status:

TASK [print]
ok: [localhost] => {
    "msg": [
        {
            "apiVersion": "v1",
            "kind": "Namespace",
[...]

130
Chapter 4 Ansible for K8s Tasks

You receive the ok statuses for the two tasks and a long JSON file containing the list of
all the namespaces in the Kubernetes cluster. To filter the output in a more synthetic and
useful way, see the next section about using the community.general.json_query filter
plugin and JMESPath.

Report Deployments in Namespace

Reporting deployments in namespaces is another repetitive and useful task for a
Kubernetes user or administrator. First of all, you need to create a deployment in order to
list them in the Kubernetes/OpenShift cluster.
The Kubernetes Deployment is a central object that allows you to specify the number
of replicas of your pods. The pod definition is inside the Deployment YAML manifest
document in the following example. The deployment.yaml document is a simple
Deployment YAML file that deploys the latest tag of the container named nginx with two
replicas listening on port TCP 3000 under the ansible-examples namespace, as shown
in Listing 4-14.

Listing 4-14. The deployment.yaml Kubernetes File

---
apiVersion: apps/v1
kind: Deployment
metadata:
  name: nginx
  namespace: ansible-examples
spec:
  replicas: 2
  selector:
    matchLabels:
      app: nginx
  template:
    metadata:
      labels:
        app: nginx
    spec:
      containers:

131
Chapter 4 Ansible for K8s Tasks

- name: nginx

          image: nginx:1.22
          imagePullPolicy: Always
          ports:
            - containerPort: 3000

You can apply your deployment to the Kubernetes cluster manually using the
kubectl command-line utility or via the kubernetes.core.k8s Ansible module:

$ kubectl apply -f deployment.yaml

A successful execution of the kubectl command returns the following message:

deployment.apps/nginx created

You can check the status of the nginx deployment using the following command:

$ kubectl get deployment nginx

The output confirms that your nginx deployment has been running for three
minutes (under the AGE column):

NAME READY UP-TO-DATE AVAILABLE AGE
nginx 2/2 1 1 3m

You can also check your Kubernetes resources such as the pods allocated by your
deployment using this command:

$ kubectl get pods

If you don’t specify the kubectl namespace, try to list the resources of the default
namespace. You should always specify the namespace (the -n ansible-examples
parameter) in order to obtain the full list of pods. If you see the following message on the
screen, it means you forgot to specify your namespace:

No resources found in the default namespace.

$ kubectl get pods -n ansible-examples

The output confirms your nginx pod, plus a random identifier inside the ansible-
examples namespace:

NAME READY STATUS RESTARTS AGE
nginx-98fabf9-pvw89 2/2 Running 0 5m
132
Chapter 4 Ansible for K8s Tasks

As you can see, the pod had 2 replicas and the status is Running for 5 minutes.

Note When Kubernetes cannot download the Container Image from the Container
Registry; it reports the statuses ImagePullBackOff and CrashLoopBackOff.
These two statuses are connected, meaning that the Kubernetes cluster tried to
download the image from the registry but without success. The full workflow is
from ErrImagePull, then ImagePullBackOff, and then keep trying. The root
cause could be a misspelled image name or tag, a registry issue, a rate limit in
the registry (see Free vs. Pro accounts in Docker Hub), a private registry without
specifying the Secret for the Pod, or simply a bit of bad luck about connectivity.

You can automate the application of your Kubernetes deployment.yaml file using the
following Ansible Playbook and the k8s module, as demonstrated in Listing 4-15.

Listing 4-15. The deployment.yml Ansible Playbook File

---
- name: k8s deployment
  hosts: all
  vars:
    myproject: "ansible-examples"
  tasks:
    - name: namespace present
      kubernetes.core.k8s:
        api_version: v1
        kind: Namespace
        name: "{{ myproject }}"
        state: present

- name: deployment present

      kubernetes.core.k8s:
        src: deployment.yaml
        namespace: "{{ myproject }}"
        state: present

133
Chapter 4 Ansible for K8s Tasks

Now execute your code in the Ansible Controller machine using the inventory, as
shown in Listing 4-16.

Listing 4-16. The Inventory File

localhost ansible_connection=local

A successful execution using the ansible-playbook command output includes the

following:

• Managed host: localhost

• Play recap status: ok=3 changed=2

• Task status:

TASK [namespace ansible-examples present]

changed: [localhost]
TASK [deployment present]
changed: [localhost]

You receive the two changed statuses for the two tasks, which means the deployment
request was successfully sent to the Kubernetes cluster API and executed. The two
changed statuses mean that the namespace and the deployment were created. If one of
the two is already present, you receive an ok status.

TASK [deployment present]

ok: [localhost]

Note You’ll receive a Kubernetes return code 404 error when the namespace is
not present and a 403 error when the namespace is terminated.

The following code lists all the deployments in the specified namespace. As
you explored in the previous section, the Ansible query lookup plugin invokes the
kubernetes.core.k8s module to interact with the Kubernetes API and return a list
of all the deployments in the cluster. The query lookup plugin (sometimes shortened
as q) invokes the kubernetes.core.k8s module, specifying the deployment as the
kind parameter and ansible-examples as a namespace parameter. You can list other
134
Chapter 4 Ansible for K8s Tasks

Kubernetes objects by changing the kind parameter value or other namespaces by

changing the namespace parameter value. The lookup plug returns the full JSON received
from the Kubernetes cluster.
You need to filter the output to extract a single element: the name of the deployment
in the namespace. The most useful Ansible filter plugin to extract a single element is the
Ansible json_query filter plugin. It’s specifically designed to select a single element or
a data structure from a JSON output. Its full name is community.general.json_query,
which means that you need the additional community.general Ansible collection
installed in your system. The community.general.json_query.
You can install the collection using the ansible-galaxy command-line tool:

$ ansible-galaxy collection install community.general

See Chapter 2 when you want to automate this process using a requirements.
yml file. The Ansible module requires the Python jmespath library to be installed in
the system:

$ pip3 install jmespath

Follow the guidance of the previous section about how to install these dependency
resources manually or automatically. With the Ansible json_query filter plugin, you can
use the JMESPath popular query language for JSON. Considering the Kubernetes output,
you apply the [*].metadata.name filter pattern which means extracting for each element
of the list (the star between brackets [*]) the key metadata, and displaying the subkey
name specified using the dot notation. The result of the Ansible lookup plugin and the
Ansible filter is saved in the deployments runtime variable using the ansible.builtin.
set_fact Ansible module. It’s displayed onscreen using the Ansible debug module.
Listing 4-17 shows the full Ansible Playbook code.

Listing 4-17. The deployment_list.yml File

---
- name: k8s deployments
  hosts: all
  tasks:
    - name: list all deployments
      ansible.builtin.set_fact:

135
Chapter 4 Ansible for K8s Tasks

deployments: "{{ query('kubernetes.core.k8s', kind='Deployment',

namespace='ansible-examples') | community.general.json_query('[*].
metadata.name') }}"

- name: display the result

ansible.builtin.debug:
msg: "{{ deployments }}"

You can execute the ansible-playbook command on the Ansible Controller

machine using the inventory, as shown in Listing 4-18.

Listing 4-18. The Inventory File

localhost ansible_connection=local

A successful execution output includes the following:

• Managed host: localhost

• Play recap status: ok=3

• Task status:

TASK [display the result]

ok: [localhost] => {
    "msg": [
        "nginx"
    ]
}

Create a Pod
A pod is an atomic unit that deploys one or more containers in a Kubernetes cluster. You
can automate the creation of the nginx pod in the ansible-example namespace with
Ansible module k8s in your Kubernetes/OpenShift cluster. The k8s Ansible module is
part of the kubernetes.core collection and manages Kubernetes objects by applying
a YAML object manifest to the src parameter. You also specify the namespace in the
namespace parameter; it refers to the Ansible variable myproject defined in the Play area

136
Chapter 4 Ansible for K8s Tasks

of the Playbook. The present state indicates the final result of the execution to create
the object if not present or to verify its existence if it’s already present in the Kubernetes
cluster.
The Ansible Playbook shown in Listing 4-19 executes the Kubernetes pod.yaml
manifest file to create the pod in the desired namespace.

Listing 4-19. The pod.yml Ansible Playbook File

---
- name: k8s pod
  hosts: all
  vars:
    myproject: "ansible-examples"
  tasks:
    - name: namespace present
      kubernetes.core.k8s:
        api_version: v1
        kind: Namespace
        name: "{{ myproject }}"
        state: present

- name: pod present

      kubernetes.core.k8s:
        src: pod.yaml
        namespace: "{{ myproject }}"
        state: present

The full Kubernetes YAML manifest file is shown in Listing 4-20.

Listing 4-20. The pod.yaml Kubernetes File

---
apiVersion: v1
kind: Pod
metadata:
name: nginx
spec:

137
Chapter 4 Ansible for K8s Tasks

  containers:
    - name: nginx
      image: nginx:latest
      ports:
        - containerPort: 80

As usual, execute the playbook code in your Ansible Controller with the localhost in
the inventory, as demonstrated in Listing 4-21.

Listing 4-21. The Inventory

localhost ansible_connection=local

A successful execution using the ansible-playbook command output includes the

following:

• Managed host: localhost

• Play recap status: ok=2 changed=1

• Task status:

TASK [pod present]

changed: [localhost]

You can check the nginx pod running as a result of the operation using the
Kubernetes kubectl command-line utility for a Kubernetes cluster:

$ kubectl get pods -n ansible-examples

In a similar way, you can list the pods in a namespace in an OpenShift cluster with
the oc command line:

$ oc get pods -n ansible-examples

The output displays the pod name, the number of pods (also called the replica count
1/1), the current status (Running), and the number of restarts:

NAME READY STATUS RESTARTS AGE
nginx 1/1 Running 0 2m

138
Chapter 4 Ansible for K8s Tasks

Create a Secret
Secrets are not encrypted; they are encoded in base64 and stored in etcd on an
encrypted volume. Note that when you decide to encrypt your data, you can’t return to
the non-encrypted status.
Listing 4-22 shows you how to automate the creation of the mysecret Kubernetes
secret in the ansible-examples namespace in the Kubernetes/OpenShift cluster using
the Ansible module k8s.

Listing 4-22. The secret.yml Ansible Playbook File

---
- name: k8s secret
  hosts: all
  tasks:
    - name: namespace present
      kubernetes.core.k8s:
        api_version: v1
        kind: Namespace
        name: "{{ myproject }}"
        state: present

- name: secret present

      kubernetes.core.k8s:
        src: secret.yaml
        state: present

You can create two data fields: username and password, with the value. The
value must be encoded as base64. Many encoders and decoders exist for the base64
encodings. In Listing 4-23, the following key and values are used:

• username = admin

• password = mysupersecretpassword

139
Chapter 4 Ansible for K8s Tasks

Listing 4-23. The secret.yaml Kubernetes File

---
apiVersion: v1
kind: Secret
metadata:
  name: mysecret
  namespace: ansible-examples
type: Opaque
data:
  username: YWRtaW4=
  password: bXlzdXBlcnNlY3JldHBhc3N3b3Jk

You can execute this code locally in your Ansible Controller machine using the
inventory, as shown in Listing 4-24.

Listing 4-24. The Inventory File

localhost ansible_connection=local

A successful execution using the ansible-playbook command output includes the

following:

• Managed host: localhost

• Play recap status: ok=2 changed=1

• Task status:

TASK [secret present]

changed: [localhost]

After execution, you can list the secrets in the ansible-examples namespace using
the kubectl command-line utility for a Kubernetes cluster:

$ kubectl get secrets -n ansible-examples

You get the same command and output for the OpenShift cluster using the oc
command-line utility:

$ oc get secrets -n ansible-examples

140
Chapter 4 Ansible for K8s Tasks

The output displays the secret name, the secret type (Opaque), the amount of fields
(2), and the age of creation (2m54s):

NAME TYPE DATA AGE
mysecret Opaque 2 2m54s

Use a Service to Expose Your App

The Kubernetes Service exposes your application to stakeholders without worrying
about the internal IP addresses of pods. Your customers need to be able to connect
to your Kubernetes pods in order to access a service or application. Now that you
understand that pods can be ephemeral, it makes no sense to rely on the pod IP address
for your services.
Instead, the best way is to use a Kubernetes Service that is designed to take care of
this use case. The Kubernetes Service is continuously updated by the pod statuses with
of healthy pods. The Kubernetes Service provides a constant IP address and port that act
as an entry point to a group of pods. This information doesn’t change for as long as the
service exists. Internal and external clients can reach your application running in a group
of pods by connecting to the service IP and ports. These connections are routed to one of
the pods behind the Kubernetes Service.
The service.yaml document is a simple Deployment YAML file that deploys your
service, exposing the internal port TCP 3000 of your application named nginx-example
to the external TCP 80 (HTTP protocol), as shown in Listing 4-25.

Listing 4-25. The service.yaml Kubernetes File

---
apiVersion: v1
kind: Service
metadata:
  name: nginx-example
spec:
  selector:
    app: nginx-example
  ports:

141
Chapter 4 Ansible for K8s Tasks

- protocol: TCP

port: 80
targetPort: 3000

You can create your nginx-example service by applying the service.yaml Service
YAML file to the Kubernetes cluster with the kubectl command or via the kubernetes.
core.k8s Ansible module:

$ kubectl apply -f service.yaml

Successful execution of the kubectl command returns the following message:

service/nginx-example created

You can also list the service nginx-example in your cluster with the following
kubectl command:

$ kubectl get service nginx-example -n ansible-examples

NAME TYPE CLUSTER-IP EXTERNAL-IP PORT(S) AGE
nginx-service ClusterIP 172.30.112.45 <none> 80/TCP 4m

The output confirms that your nginx-example service has been running for four
minutes; it listens on port 80/TCP at the 172.30.112.45 Cluster IP and has no external
IP associated with it.
For reference, the following command operates in the OpenShift cluster:

$ oc get services -n ansible-examples

The Ansible Playbook shown in Listing 4-26 automates the execution of the service.
yaml Kubernetes YAML manifest file using the k8s module.

Listing 4-26. The service.yml Ansible Playbook File

---
- name: k8s service
  hosts: all
  vars:
    myproject: "ansible-examples"
  tasks:
    - name: k8s service
      kubernetes.core.k8s:

142
Chapter 4 Ansible for K8s Tasks

        src: service.yaml
        namespace: "{{ myproject }}"
        state: present

Execute your code using the ansible-playbook command on your Ansible

Controller machine using the inventory file, as shown in Listing 4-27.

Listing 4-27. The Inventory File

localhost ansible_connection=local

Successful execution using the ansible-playbook command output includes the

following:

• Managed host: localhost

• Play recap status: ok=2 changed=1

• Task status:

TASK [k8s service]

changed: [localhost]

You can verify the result using the kubectl command-line utility for a Kubernetes
cluster and the oc command-line utility for the OpenShift cluster.

Kubernetes Networking
Kubernetes uses five techniques to track pod status and direct traffic to the
appropriate pods:

–– ClusterIP: Accessible only internally within the cluster.

–– NodePort: Exposes a static port (the NodePort) on each node’s IP

that can be accessed from outside the cluster by the address
NodeIP:NodePort. The NodePort service connects to a ClusterIP
service that is created automatically for your NodePort.

–– LoadBalancer: Exposes a load balancer externally. The

LoadBalancer service connects to the NodePort and ClusterIP,
automatically created.
143
Chapter 4 Ansible for K8s Tasks

–– Ingress: Reduces the amount of load balancer for HTTP and HTTPS
defining traffic routes.

–– ExternalName: Maps to a DNS name. It returns a CNAME record with

its value.

The simplest network configuration is the ClusterIP service type. This is accessible
from within the cluster only. A ClusterIP service IP address is assigned from the cluster
range. The ClusterIP service routes the traffic to the nodes. Behind the scene, each
node runs a kube-proxy container for this task. The kube-proxy container creates the
appropriate IPtables firewall rules to redirect the ClusterIP traffic to the appropriate Pod
IP address. This type of service is used by frontend pods to redirect traffic to the backend
pods or to act like load balancers.
In the Kubernetes YAML for ClusterIP service, the most important field is the
selector. It determines which pods serve as endpoints for this service (see Listing 4-28).

Listing 4-28. The ClusterIP.yaml Kubernetes File

---
apiVersion: v1
kind: Service
metadata:
  name: "nginx-example"
  namespace: "mynamespace"
spec:
  type: ClusterIP
  selector:
    app: "nginx"
  ports:
    - name: http
      protocol: TCP
      port: 80
      targetPort: 9376

The NodePortservice type is similar to a ClusterIP service, but it also opens a port
on each node. Opening a port allows access to the service from inside the cluster (using
the ClusterIP). External users can connect directly to the node on the NodePort.

144
Chapter 4 Ansible for K8s Tasks

A random port is opened on the local node for each service unless the nodeport
property specifies a specific port (see Listing 4-29). The kube-proxy container forwards
traffic from the port to the service’s cluster IP and then to the pod that’s updating the
IPtable rules.

Listing 4-29. The NodePort.yaml Kubernetes File

---
apiVersion: v1
kind: Service
metadata:
  name: "nginx-example"
  namespace: "mynamespace"
spec:
  type: NodePort
  selector:
    app: "nginx"
  ports:
    - name: http
      protocol: TCP
      port: 80
      targetPort: 9376
      nodeport: 25000

The LoadBalancer service type extends the NodePort service type by adding a load
balancer in front of all nodes. This means Kubernetes requests a load balancer and
registers all the nodes. The load balancer doesn’t detect where the pods for a specific
service are running. Consequently, the load balancer adds all worker nodes as backend
instances. The Load Balancer service type uses the classic load balancers by default. This
means there is a classic load balancer in front of all instances listening to your requests.
The load balancer routes the requests to the nodes through an exposed port. You can
also use the network load balancer instead of the classic load balancer.
The classic load balancer processes the requests one by one when they arrive from
the Internet. It then forwards the request to one of the Kubernetes instances on a specific
port. When there is a service listening to a specific port, it acts like a second-layer load
balancer for the backend pods that handle the requests. Listing 4-30 illustrates this.

145
Chapter 4 Ansible for K8s Tasks

Listing 4-30. The LoadBalancer.yaml Kubernetes File

---
apiVersion: v1
kind: Service
metadata:
  name: "nginx-example"
  namespace: "mynamespace"
spec:
  type: LoadBalancer
  selector:
    app: "nginx"
  ports:
    - name: http
      port: 80
      targetPort: 9376

You can reduce the number of load balancers using the Kubernetes ingress object.
An ingress object exposes HTTP and HTTPS routes outside the cluster and routes traffic
to your services according to defined traffic rules. Ingress objects use ingress controllers,
which fulfill the ingress rules and requests (usually using a load balancer). Using the
ingress objects and controllers, you can transition from one load balancing per service
to one load balancer per ingress and route to multiple services. Traffic can be routed to
the proper service using path-based routing. When using Kubernetes, you have many
ingress controller options, for example, Envoy Controllers and NIGNX controllers.
The load balancer picks the proper target group based on the path. The load
balancer then forwards the request to one of the Kubernetes instances on the application
ports. The service listens at a specific port and balances the requests to one of the pods
of the application or service, as demonstrated in Listing 4-31.

Listing 4-31. The Ingress.yaml Kubernetes File

---
apiVersion: networking.k8s.io/v1
kind: Ingress
metadata:
name: "nginx-example"

146
Chapter 4 Ansible for K8s Tasks

  namespace: "mynamespace"
  annotations:
    nginx.ingress.kubernetes.io/rewrite-target: /
spec:
  ingressClassName: nginx-example
  rules:
    - http:
        paths:
          - path: /webapp
            backend:
              service:
                serviceName: webapp
                servicePort: 9376

The ExternalName service type is used to connect to a resource outside of the cluster.
For example, this could be a database or file system resource that is in the network
outside the cluster. The pods that require access to the external resource service require
access to the resource to connect to the resource service, which returns the external
resource endpoints. This type of service is helpful when you decide to migrate the
external resource to the Kubernetes cluster.
When the resource is deployed in the cluster, you can update the service type to
ClusterIP. The application can continue to use the same resource endpoint.
The most important property in the YAML manifest is the externalName field, which
specifies where the service maps to (see Listing 4-32).

Listing 4-32. The ExternalName.yaml Kubernetes File

---
apiVersion: v1
kind: Service
metadata:
  name: "nginx-example"
  namespace: "mynamespace"
spec:
  type: ExternalName
  externalName: mydb.example.com

147
Chapter 4 Ansible for K8s Tasks

You can apply all your network YAML files in this section to the Kubernetes
cluster using the kubectl command-line utility or via the kubernetes.core.k8s
Ansible module:

$ kubectl apply -f ClusterIP.yaml

Substitute the name of the YAML file with the relevant code in this section:
ClusterIP.yaml, NodePort.yaml, LoadBalancer.yaml, Ingress.yaml, or
ExternalName.yaml.

Scale Your App

Resilience is an important property that you want to achieve from your services or
applications. Containers enable you to achieve this property by running multiple
replicas of the same pod. Cost optimization involves running the minimum available
pods (a minimum of two) when there is low traffic and being able to scale up or down
following the traffic needs.
Some business-critical services must be designed with high availability and fault
tolerance. Your application must be created to be cloud-ready or, even better, cloud-
native in order to unlock the full potential of cloud computing.
Imagine a service scenario created by two pods. If there is a failure of one container
or the infrastructure under the container, the service can survive one container failure.
In high availability (HA), the application survives the failure, but the traffic is
impacted by poor performance. Even if the service survived failure, it withstands failure.
To be fault tolerant (FT), your application needs two instances in order to operate
at its full potential. The service needs to be designed in order to recover at least the
minimum amount of pods.
Since Kubernetes version 1.21, PodDisruptionBudget (PDB) is a way to limit the
number of concurrent disruptions that your application experiences, thus allowing for
high availability. Meanwhile, it allows the cluster administrator to manage the nodes of
the cluster by manually draining a node or preventing an upgrade from taking too many
copies of the application offline. PDB allows you to specify and overprovision in order
to guarantee the high availability of your service. PDB is different from auto-scaling
because it overprovisions the application on purpose. For full disclosure, if there is a
failure in the production infrastructure and you need to have a Disaster Recovery (DR)
infrastructure.

148
Chapter 4 Ansible for K8s Tasks

Scaling is the best feature of the Kubernetes platform and allows you to match
customer demands with resources. You can scale up (increase) your deployment using
the scale parameter of the kubectl command or using the kubernetes.core.k8s_scale
Ansible module in order to change the number of replicas from the current number to
three of your nginx deployment:

$ kubectl scale deployment nginx --replicas=3

The output confirms your nginx deployment was successfully scaled:

deployment.apps/nginx scaled

To check the number of replicas, you need to use the get parameter of the kubectl
command:

$ kubectl get deployments

As you can see in the output, the nginx is now running on three replicas confirmed
by the 3/3 in the READY column:

NAME READY UP-TO-DATE AVAILABLE AGE
nginx 3/3 3 3 125m

In the same way, you can scale down (reduce) your deployment using the scale
parameter of the kubectl command. You do this in order to change the number or
replicas of your deployment from the current number to one of your deployment:

$ kubectl scale deployment nginx --replicas=1

The output confirms your nginx deployment was successfully scaled:

deployment.apps/nginx scaled

To check the number of replicas, you need to use the get parameter of the kubectl
command:

$ kubectl get deployments

149
Chapter 4 Ansible for K8s Tasks

As you can see in the output, the nginx is now running on one replica confirmed by
the 1/1 in the READY column:

NAME READY UP-TO-DATE AVAILABLE AGE
nginx 1/1 1 1 31m

With Ansible, you can scale a deployment, ReplicaSet, replication controller, and job
using the kubernetes.core.k8s_scale Ansible module. The k8s_scale Ansible module
is part of the kubernetes.core collection and it interacts with the Kubernetes API like
the kubectl or oc command-line utilities.
For example, let’s suppose you want to increase the pod count for your nginx
deployment inside the ansible-examples namespace. Figure 4-13 gives a clear overview
of the result of the scale-up operation in your Kubernetes cluster dashboard. Note the
pod count from two to ten. Consequently, all the other resources are increased as well.

Figure 4-13. Kubernetes pod resource utilization

The Ansible Playbook shown in Listing 4-33 changes the replica count to ten of the
deployment nginx in the namespace ansible-examples. The Ansible Playbook waits for
a timeout maximum of 120 seconds (two minutes) for the operation to complete.

150
Chapter 4 Ansible for K8s Tasks

Listing 4-33. The scale_up.yml Ansible Playbook File

---
- name: k8s scale
  hosts: all
  vars:
    myproject: "ansible-examples"
    mydeployment: "nginx"
    myreplica: 10
    mytimeout: 120
  tasks:
    - name: scale Deployment
      kubernetes.core.k8s_scale:
        api_version: v1
        kind: Deployment
        name: "{{ mydeployment }}"
        namespace: "{{ myproject }}"
        replicas: "{{ myreplica }}"
        wait_timeout: "{{ mytimeout }}"

Execute your code in the Ansible Controller machine using the inventory, as shown
in Listing 4-34.

Listing 4-34. The Inventory File

localhost ansible_connection=local

Successful execution using the ansible-playbook command output includes the

following:

• Managed host: localhost

• Play recap status: ok=2 changed=1

• Task status:

TASK [scale up]

changed: [localhost]

151
Chapter 4 Ansible for K8s Tasks

The changed status means the deployment changed the replica count from the
Kubernetes cluster API. If the replica count already matches the current pod count, you
would obtain the ok status.
When the pod count is changed by a big number and the time is not sufficient,
you might receive a failed status with the following message on the screen: Resource
scaling timed out. This simply means that the Kubernetes cluster is taking longer than
expected to execute the operation; it’s not a failure message. A more deep analysis of the
fatal error message, for example, reveals the operation in progress:

"message": "ReplicaSet \"nginx-689b466988\" is progressing."

Ansible allows you to scale your replica count when you reach a specific number of
pods. The Ansible Playbook shown in Listing 4-35 reduces the pod count from 10 to 5 of
the deployment nginx in the namespace ansible-examples.

Listing 4-35. The scale_down.yml Ansible Playbook File

---
- name: k8s scale
  hosts: all
  vars:
    myproject: "ansible-examples"
    mydeployment: "nginx"
    curr_replica: 10
    myreplica: 5
  tasks:
    - name: scale down
      kubernetes.core.k8s_scale:
        api_version: v1
        kind: Deployment
        name: "{{ mydeployment }}"
        namespace: "{{ myproject }}"
        current_replicas: "{{ curr_replica }}"
        replicas: "{{ myreplica }}"

152
Chapter 4 Ansible for K8s Tasks

The outcome is similar to the previous Ansible Playbook, except that Kubernetes
is going to reduce the replica count from 10 to 5 pods when the current_replicas
parameter is reached. This playbook is helpful for performing the scale operation when
the threshold is reached.

Auto-scaling
The next level is to use metrics to automatically scale your cluster.

• Horizontal scaling adds/remove pods of the same size capacity.

• Vertical scaling keeps the same number of pods, but you change the
size of capacity.

Kubernetes Cluster Autoscaler (CA) allocates the right amount of resources to

guarantee that your application meets the demands of the traffic. The change is
performed for your pods in the auto-scale group in your cluster in order to enable
Horizontal Pod Autoscaler (HPA) or Vertical Pod Autoscaler (VPA).
Kubernetes Cluster Autoscaler can be deployed, but it requires permission to
interact with your cluster. The Kubernetes Cluster Autoscaler needs a service account to
interact with the auto-scaling group.
Table 4-1 lists the necessary configuration parameters in your Kubernetes cluster.

Table 4-1. Auto-Scaling API Endpoints

Key Value

k8s.io/cluster-autoscaler/my-cluster Owned
k8s.io/cluster-autoscaler/enabled True

You can download the Cluster Autoscaler via the kubernetes/autoscaler GitHub
repository. The installation process might vary, but it’s relatively simple and usually
involves a YAML manifest file, according to the Kubernetes release version or cloud
provider. For the Amazon Web Services (AWS) provider, you just need to download and
install the cluster-autoscaler-autodiscover.yaml YAML manifest file.
After a successful download, you can apply it to your Kubernetes cluster:

$ kubectl apply -f cluster-autoscaler-autodiscover.yaml

153
Chapter 4 Ansible for K8s Tasks

You can also tune some parameters in the kube-system namespace:

$ kubectl -n kube-system edit deployment.apps/cluster-autoscaler

Set the appropriate Kubernetes version to the Cluster Autoscaler image tag with the
following command. Replace 1.21 with your Kubernetes version value.

$ kubectl set image deployment cluster-autoscaler \

-n kube-system \
cluster-autoscaler=k8s.gcr.io/autoscaling/cluster-autoscaler:v1.21

Pods status pile up in a Pending status and they transition to Ready when they can
handle the traffic.
The Vertical Pod Autoscaler has three working modes:

• Off: Recommended only for simulations

• Initial: Initialization only

• Auto: Performing the action on your cluster

It’s always a good practice to start with the Off mode and increase the more you have
data about your traffic and usage. The wrong combination of horizontal and vertical
Autoscaler can be difficult to troubleshoot and not obtain the desired result.

Tip The k8s.gcr.io registry is frozen and considered an “old” registry since

April, 2023; no further images for Kubernetes and related subprojects are pushed
in it. The registry.k8s.io is the community-owned image onward.

Update Your App

This is the most controversial topic because users expect to access the application or
service all the time. Instead, the development team wants to deploy new versions as fast
as possible. Sometimes the release process happens several times a day, with continuous
delivery. The rolling update is the solution in the Kubernetes world, as shown in
Figure 4-14.

154
Chapter 4 Ansible for K8s Tasks

Deployment Deployment

Image v1 Image v2

Image v1 Rolling update Image v2

Image v1 Image v2

Figure 4-14. Kubernetes rolling update

The Kubernetes rolling update enables you to update an application while the
users are connected to it. Under the hood, Kubernetes updates pod instances with
a new one performing a zero downtime deployment update on nodes with enough
available resources. The mechanism behind it is similar to the scale mechanism. Like in
application scaling, when a deployment is publicly exposed, the service load balances
the traffic on the available pods during the update. Having more than one copy of the
pods ensures the high availability of the application while the update is in progress. The
rolling update mechanism also offers a rollback option if something goes wrong during
the update process; you can simply return to the previous version of the pod anytime.

Note Using the latest tag is discouraged because it fails when rolling an
update from image:latest to a new image:latest, even when the image tag
has been updated. The latest tag is identical from the Kubernetes point of view
to the current value.

Suppose you are currently running a Nginx web server using the nginx:1.22 (image
nginx tag 1.22) in the nginx deployment. The following command changes the image
of the nginx deployment to the nginx:1.23 (image nginx tag 1.23). Manually, you can
specify the set image parameter for deployment to set a new image or a new image tag
to deploy.

$ kubectl set image deployments/nginx=nginx:1.23

The command tells the Kubernetes nginx deployment to use a different image and
initialize a rolling update process to deploy the image nginx with tag 1.23.

155
Chapter 4 Ansible for K8s Tasks

You can verify the status by running the following command for the nginx deployment:

$ kubectl rollout status deployment nginx

You can perform a rollback in the nginx deployment using this command:

$ kubectl rollout undo deployments nginx

Logs are recorded in the Kubernetes cluster and you can review them using this
command:

$ kubectl rollout history deployment nginx

You can automate the rolling update of the deployment using the Ansible k8s
module and specifying a new deployment YAML manifest file.
The Kubernetes nginx deployment manifest file uses the following image definition
image: nginx:1.22. The updated nginx deployment manifest file reports the new image
and tag image: nginx:1.23.
The full Kubernetes updated YAML deployment manifest file is shown in
Listing 4-36.

Listing 4-36. The deployment2.yml Kubernetes Manifest File

- name: nginx

          image: nginx:1.23
          imagePullPolicy: Always
          ports:
            - containerPort: 3000

You can automate the application of your Kubernetes deployment.yaml file using the
deployment.yml—the same Ansible Playbook that you used in the “Report Deployments
in Namespace” section. You just need to change the value of the src parameters for the
module, specifying the newest deployment manifest filename. Line 16 changes from
src: deployment.yaml to src: deployment2.yaml.
Successful execution using the ansible-playbook command output includes the
following:

• Managed host: localhost

• Play recap status: ok=3 changed=1

• Task status:

TASK [namespace ansible-examples present]

ok: [localhost]
TASK [deployment present]
changed: [localhost]

You received the changed statuses for the two tasks, which means the deployment
request was sent to the Kubernetes cluster API and the rolling update process is ongoing.

Assign Resources to Kubernetes K8s Pods

Out of the box, a Kubernetes cluster is configured to allow the usage of as many
resources as available in the nodes. This behavior is understandable and potentially
allows a Kubernetes service to serve many customers using all the cluster resources.
On the other end, excessive resource consumption from a resource leak or an unstable
or rogue pod could consume all the resources and not allow other pods to use them.
Imagine how dangerous it could be if a program gets out of control and consumes all the
available CPU. The root cause may a bad developer, bad code, or simply bad luck—the

157
Chapter 4 Ansible for K8s Tasks

most important thing is being in control of your Kubernetes clusters. You can apply CPU
and memory resource limitations to your Kubernetes cluster to prevent this possible
harmful scenario.
Observability is the pillar of every successful service. Predicting trends and obtaining
meaningful information from the data is more important than ever.

Metrics
In order to guarantee a healthy Kubernetes cluster, the Cluster Administrator Team and
the Service Reliability Team must acquire some data and perform some predictions.
Managing metrics and logs in your Kubernetes cluster is the way to guarantee the
business continuity of your application. The OpenTelemetry (https://opentelemetry.
io/) project is trying to standardize metrics all over the world. The project delivers tools,
a Software Development Kit, and an API to be instrumental in the generation, collection,
and export of telemetry data (logs, metrics, and traces). They help you analyze the
performance and behavior of your software. Prometheus metrics store the data inside
Prometheus. PromQL is the query language to search across the data.
Control planes and data planes create metrics in the Prometheus format. You can see
them in the RAW format (JSON) via the command-line command:

$ kubectrl get –raw /metrics

You can represent them in a graphical way from a Grafana dashboard. Performance
monitoring, setting alarms, and acquiring insights from the logs and metrics are all of
great value for making decisions in your organization.

CPU Resources
Kubernetes CPU resources are defined in millicores. You can specify any integer value by
adding the letter m to the end. Each virtual CPU core has a value of 1000m. If you want to
specify 50 percent of a CPU core, you use the value 500m, whereas for two full CPU cores,
you use the value 2000m.
Set a limitation on the usage of the CPU for your containers. A container is
guaranteed to be allocated as much CPU as it requests when the cluster has enough
capacity available. To specify a soft limitation, use the CPU request for a container using
the resources:requests field in the container resource manifest. To specify a hard
limitation, use the CPU limit, also including the resources:limits.
158
Chapter 4 Ansible for K8s Tasks

The vish/stress container image is a famous CPU stress test that allows you to test
your cluster, set some limits and requests, and then see the results. You can find it in
many Container Registries, such as Docker Hub (https://hub.docker.com/r/vish/
stress).
Listing 4-37’s Playbook executes the CPU stress test cpu-demo-pod in the cpu-
example namespace, setting some CPU limits and requests.

Listing 4-37. The cpu.yml Ansible Playbook File

---
- name: k8s cpu
  hosts: all
  vars:
    myproject: "cpu-example"
  tasks:
    - name: namespace present
      kubernetes.core.k8s:
        kind: Namespace
        name: "{{ myproject }}"
        state: present
        api_version: v1
    - name: cpu pod
      kubernetes.core.k8s:
        state: present
        definition:
          apiVersion: v1
          kind: Pod
          metadata:
            name: cpu-demo
            namespace: "{{ myproject }}"
          spec:
            containers:
              - name: cpu-demo-pod
                image: vish/stress
                resources:

159
Chapter 4 Ansible for K8s Tasks

                  limits:
                    cpu: "1"
                  requests:
                    cpu: "0.5"
                args:
                  - -cpus
                  - "2"

As usual, you execute the code in your Ansible Controller with the localhost in the
inventory, as shown in Listing 4-38.

Listing 4-38. The Inventory File

localhost ansible_connection=local

You can execute the cpu.yml playbook using the ansible-playbook command-
line tool:

$ ansible-playbook -i inventory cpu.yml

Note that the Kubernetes cluster in this stress test respects the 0.5 CPU request (soft
limitation) with a maximum of 1.0 CPU request (hard limitation), despite the process
asking for more resources and 2.0 CPUs. You can verify this behavior using the kubect
command line and specifying the cpu-example namespace and the cpu-demo pod name:

$ kubect get pod cpu-demo --namespace=cpu-example --output=yaml

[...]
    resources:
      limits:
        cpu: "1"
      requests:
        cpu: 500m
[...]

You can use the cpu.yaml playbook, changing the pod name to set CPU limits and
requests to any pod in your Kubernetes cluster.

160
Chapter 4 Ansible for K8s Tasks

Memory Resources
Kubernetes memory resources are defined in bytes and multiples. A mebibyte value for
memory is commonly used, following the two powers 1024 KiB instead of the thousands
1000 kB.
You can limit the memory usage for pods to prevent dangerous memory leaks. The
Kubernetes cluster honors the limits and requests specified by the user. To specify a
memory request for a container, include the resources:requests field in the container
resource manifest. To specify a memory limit, specify resources:limits.
The polinux/stress container image is a famous memory stress test that allows you
to test your cluster and set some limits and requests about RAM usage. You can find the
image on the most popular Container Registry, for example, Docker Hub (https://hub.
docker.com/r/polinux/stress).
The playbook in Listing 4-39 executes the memory demo application memory-demo-
pod in the mem-example namespace, thus setting some memory limits and requests.

Listing 4-39. The ram.yml Ansible Playbook File

---
- name: k8s memory
  hosts: all
  vars:
    myproject: "mem-example"
  tasks:
    - name: namespace present
      kubernetes.core.k8s:
        kind: Namespace
        name: "{{ myproject }}"
        state: present
        api_version: v1
    - name: memory pod
      kubernetes.core.k8s:
        state: present
        definition:
          apiVersion: v1

161
Chapter 4 Ansible for K8s Tasks

          kind: Pod
          metadata:
            name: memory-demo
            namespace: "{{ myproject }}"
          spec:
            containers:
              - name: memory-demo-pod
                image: polinux/stress
                resources:
                  requests:
                    memory: "100Mi"
                  limits:
                    memory: "200Mi"
                command: ["stress"]
                args: ["--vm", "1", "--vm-bytes", "500M", "--vm-hang", "1"]

As you can see, the pod has set a memory request of 100 MiB and a memory limit of
200 MiB. During execution, the pod attempts to allocate 500 MiB of memory, which is
not allowed.
You can verify the successful application of a pod limitation using the following
command:

$ kubect get pod memory-demo --namespace=mem-example --output=yaml

[...]
      resources:
        limits:
          memory: 200Mi
        requests:
          memory: 100Mi
[...]

Namespace Resources
Assigning resources per pod is very granular but might be daunting. A more durable
approach is to define resource needs at a namespace level.

162
Chapter 4 Ansible for K8s Tasks

After creating a namespace, you can add a ResourceQuotas Kubernetes object that
defines the CPU and memory limits and requests for each namespace. A more powerful
Kubernetes object is LimitRange. Unlike the ResourceQuotas Kubernetes object, which
looks at the namespace as a whole, a LimitRange Kubernetes object applies to an
individual pod.

GPU Resources
At the moment, Kubernetes doesn’t provide a native tool to manage the Graphics
Processing Units (GPU) resources of the cluster. The GPUs are a critical component in
order to build and train deep learning (DL) and machine learning (ML) algorithms and
bring artificial intelligence (AI) initiatives to production. At the moment, it’s possible to
allocate the GPU resources in a static way and only on the worker node of your pod. The
Run:AI startup from Tel Aviv, Israel created the Atlas AI Orchestration Platform to pool
the available GPU resources, efficiently optimize the scheduling, distribute the workload,
and set a prioritization schedule in your cluster. More details are available on the official
website at https://www.run.ai.

Configure a Pod to Use a Volume for Storage

A pod is ephemeral; the file system of a pod lives only as long as the pod does. Therefore,
when a pod terminates or restarts, any file system changes are lost. Volumes allow you to
assign more consistent storage, independent from the pod’s lifecycle. This is especially
important for stateful applications, data lakes (structured, semi-structured, and
unstructured data), and databases.
For example, let’s suppose you want to execute a Redis server. Redis is a popular BSD
open-source license, in-memory data structure store used as a NoSQL database, cache,
and message broker. The Ansible Playbook shown in Listing 4-40 creates the volume-
example namespace with a pod from the redis container image. It also has a redis-
storage storage that stores the data created in your application.

163
Chapter 4 Ansible for K8s Tasks

Listing 4-40. The storage.yml Ansible Playbook File

---
- name: k8s volume
  hosts: localhost
  gather_facts: false
  connection: local
  vars:
    myproject: "volume-example"
  tasks:
    - name: create namespace
      kubernetes.core.k8s:
        kind: Namespace
        name: "{{ myproject }}"
        state: present
        api_version: v1

- name: pod with storage

      kubernetes.core.k8s:
        state: present
        namespace: "{{ myproject }}"
        definition:
          apiVersion: v1
          kind: Pod
          metadata:
            name: redis
          spec:
            containers:
              - name: redis
                image: redis
                volumeMounts:
                  - name: redis-storage
                    mountPath: /data/redis
            volumes:
              - name: redis-storage
                emptyDir: {}

164
Chapter 4 Ansible for K8s Tasks

The inventory file in Listing 4-41 executes the Ansible Playbook in the current
Ansible Controller with a local connection.

Listing 4-41. The Inventory File

localhost ansible_connection=local

Successful execution using the ansible-playbook command output includes the

following:
• Managed host: localhost
• Play recap status: ok=2 changed=2
• Task status:

TASK [create volume-example namespace]

changed: [localhost]
TASK [pod with storage]
changed: [localhost]

You receive the changed status when the namespace was successfully created in the
Kubernetes cluster and the pod was successfully created. Otherwise, you’ll receive the
ok status when the namespace or the pod was already present in the cluster (Ansible
idempotency property). You can verify the successful operation via the command line or
via the dashboard, like the one shown in Figure 4-15.

Figure 4-15. Storage pod in the OpenShift console

165
Chapter 4 Ansible for K8s Tasks

Apply Multiple YAML Files at Once on Kubernetes

Deploying complex applications or services with Kubernetes requires many objects
(namespaces, pods, services, deployments, and so on) in a precise order. Especially with
the fast deploying applications, this task might need to be repeated often. If you perform
manually or in a deployment pipeline, the following code is beneficial.
Ansible module k8s and the Ansible lookup plugin fileglob (both included in all
the installations) enable you to list files to match a pattern. The fileglob lookup plugin
is distributed as ansible.builtin.fileglob. You can specify the following parameter:

- "./defs/*.yaml"

The *.yaml command list all the files with the yaml extension under the defs
directory on the target node. Obviously, the *.yaml YAML text files are assumed to be
Kubernetes object manifest files. Ansible will process them in alphabetical order. Some
services require a precise order, and a list of files becomes handy in this use case. The
following Ansible list deploys a namespace, a pod, and a service, in this order:

        - "./defs/namespace.yaml"
        - "./defs/pod.yaml"
        - "./defs/service.yaml"

Ansible processes each manifest file using the k8s module (full name kubernetes.
core.k8s), which you read about in previous sections. You are going to use the template
lookup plugin to read the contents of each manifest file. The plugin is very powerful
because it enables retrieval of the contents of the file and template with the Jinja2
programming language. Jinja is a web template engine for the Python programming
language.
The full Ansible Playbook is shown in Listing 4-42.

Listing 4-42. The multiple.yml Ansible Playbook File

---
- name: k8s multiple
  hosts: all
  tasks:
    - name: k8s resources
      kubernetes.core.k8s:

166
Chapter 4 Ansible for K8s Tasks

definition: "{{ lookup('template', '{{ item }}') | from_yaml }}"

      with_fileglob:
        - "./defs/namespace.yaml"
        - "./defs/pod.yaml"
        - "./defs/service.yaml"

The following inventory file executes the Ansible Playbook in the current Ansible
Controller with a local connection. See Listing 4-43.

Listing 4-43. The Inventory File

localhost ansible_connection=local

The YAML manifest files are already included in the book in the previous sections.
Make sure that the YAML manifest files exist in the target node before launching
the code. If a YAML manifest file is not found, Ansible proceeds to the next file without
performing the Kubernetes requests. Successful execution using the ansible-playbook
command output includes the following:
• Managed host: localhost
• Play recap status: ok=2 changed=1
• Task status:

TASK [k8s resources]

changed: [localhost] => (item=./defs/namespace.yaml)
changed: [localhost] => (item=./defs/pod.yaml)
changed: [localhost] => (item=./defs/service.yaml)

When the objects are already present in the Kubernetes cluster, you obtain an ok:
[localhost] status.

Key Takeaways
In this chapter, you learned how to install a Kubernetes cluster and how to create the
objects inside of it. Kubernetes is an amazing technology that enables and simplifies the
execution of world-scale applications and services with many features. Under the hood,
it uses a lot of objects for services, deployment, and networking. This chapter explored

167
Chapter 4 Ansible for K8s Tasks

the main components and explained how to deploy your Hello application and service.
You can take advantage of the combination of Kubernetes for orchestration with Ansible
automation to simplify the software lifecycle of your cluster and the application running
inside of it. From installation to the deployment of applications, Ansible is the Swiss
Army Knife kit. In the next chapters, you are going to dip your toes into the use cases
of software deployment and cluster management, and you’ll explore the most popular
Kubernetes cloud providers.

168
CHAPTER 5

Ansible for K8s Data

Plane
Ansible is a valuable tool for all IT professionals, DevOps personnel, application
developers, site reliability engineers (SREs), and IT operations teams worldwide.
This chapter explores some exciting use cases for the modern development team.
Applying Agile methodologies and continuous integration, as well as quickly deploying
applications to customers, is critical in today’s fast-paced world. Software development
teams in many organizations have embraced the software development lifecycle (SDLC)
methodology (see Figure 5-1).

Requirements analysis

Deployment Planning

Testing Architectural design

Software development

Figure 5-1. The SDLC model

The following stages apply to the SDLC methodology:

1. Requirements analysis: Identify the problems and use cases.

2. Planning: Define the costs and resources needed.

3. Architecture design: Determine the design specification

requirements.

4. Software development: Build the software.

169
© Luca Berton 2023
L. Berton, Ansible for Kubernetes by Example, https://doi.org/10.1007/978-1-4842-9285-3_5
Chapter 5 Ansible for K8s Data Plane

5. Testing: Test for defects and deficiencies and verify the meeting of
the specifications.

6. Deployment: Launch the product to the customer and get feedback.

As you can notice, the SDLC methodology is a cyclical process that uses the Agile
methodology. For example, in a sprint timeframe, you must release a minimum viable
product (MVP) and enrich its features. Other methodologies that can be used are the
waterfall model, the spiral model, extreme programming (XP), and so on.
The twelve-factor application1 is an important methodology for building software-
as-a-service (SaaS) applications that can be applied to any programming language with
any combination of backing services (database, queue, memory cache, and so on). The
twelve factors are based on the experience of development, operation, and scaling of
hundreds of thousands of applications on the Heroku platform:

1. Codebase: Create one codebase tracked in revision control, many

deploys.

2. Dependencies: Explicitly declare and isolate dependencies.

3. Config: Store the config in the environment.

4. Backing services: Treat backing services as attached resources.

5. Build, release, run: Strictly separate build and run stages.

6. Processes: Execute the app as one or more stateless processes.

7. Port binding: Export services via port binding.

8. Concurrency: Scale out via the process model.

9. Disposability: Maximize robustness with fast startup and graceful

shutdown processes.

10. Dev/prod parity: Keep development, staging, and production as

similar as possible.

11. Logs: Treat logs as event streams.

12. Admin processes: Run admin/management tasks as one-off

processes.

1
https://12factor.net/

170
Chapter 5 Ansible for K8s Data Plane

The twelve-factor application principles lay the groundwork for developing scalable,
resilient, and maintainable applications that can operate on any cloud platform. A
good understanding of the twelve-factor application methodology is an indispensable
requirement for constructing cloud-native applications that capitalize on the advanced
functionalities of Kubernetes.
The Kubernetes data plane is populated by the cloud-native applications that
your development team creates to satisfy the needs of your target customers and
stakeholders. You can take advantage of Ansible in order to increase the velocity of
deploying your software.

Configuring a Java Microservice

Configuring a Java microservice consists of several steps. First, a cloud-native application
follows the architecture design pattern and the structure of a set of loosely coupled,
collaborating microservices. This will enable continuous integration and continuous
delivery, as well as make the application easier to scale.
Next, you need to create three classes: configuration, application, and main. The
configuration class defines the beans and their dependencies, while the application
class sets up the application context, and the main class is used to run the application.
You should also create a build file to specify how the application should be built and
configured (often a Jenkinsfile).
Finally, you need to configure the external services, such as databases and messaging
systems, that the microservice might use to store or process data. This involves setting up
the appropriate connection parameters and security protocols. Once all of this is done,
you can deploy the microservice and start using it.
To configure a Java microservice with Kubernetes, you first need to set up the
Kubernetes cluster and create the necessary resources, such as deployments and
services. After the cluster is set up, you create the configuration file that defines the
application components and their dependencies, such as databases and messaging
systems. This configuration file will be used by Kubernetes to create the application pods
and services. After this is done, you need to deploy the application to the cluster using a
deployment tool such as Kubernetes or Docker. Once the application runs, you need to
configure the Kubernetes resources to ensure that the application is secure and performs
optimally. Finally, you need to monitor the application to ensure it runs correctly and is
not wasting your resources (memory leaks, CPU throttling, and so on).

171
Chapter 5 Ansible for K8s Data Plane

As an example, evaluate the Quarkus project, a Java framework distributed under

the open-source Apache License 2.0 by Red Hat, as a runtime environment. Its popular
Kubernetes Native Java stack is perfectly tailored for OpenJDK HotSpot and GraalVM,
with a small footprint. It’s compatible with the most famous Java libraries and standards.
The major benefits are a millisecond startup, low memory utilization, and a smaller
container image compared to the traditional JVM execution stack.
A popular framework for creating Java microservices applications is Spring Boot. At
the moment of writing this book, Spring Boot is the world’s most popular Java framework
and is supported by a great community of developers worldwide. The Spring Initializr
enables you to start your projects and quickly package then as JARs.
There are multiple options for containerizing a Spring Boot application. Since you
are already building a Spring Boot JAR file, you only need to call the plugin directly. The
most common build automation tools are Maven and Gradle command-line utilities:

• Apache Maven: $ mvn spring-boot:build-image

• Gradle: $ gradle bootBuildImage

For the Windows platform, use the gladlew and mvnw commands instead.

The Demo Java Web Application

The following example shows how you can build web and RESTful applications using
Spring Web integrated on Spring MVC and distribute them as an embedded containers
powered by Apache Tomcat. This is a pure Java application that creates a simple RESTful
API interface.
The first step is to start with Spring Initializr from the website at https://start.
spring.io/. Then select either Gradle or Maven and the Java language that you want
to use. Then specify in the Dependencies Spring Web. When you’re ready, choose the
Generate button to download the sample ZIP file archive with the web application
already configured. The final selection with Gradle build tool, Java language, and Spring
Boot with Spring Web is shown in Figure 5-2.

172
Chapter 5 Ansible for K8s Data Plane

Figure 5-2. Spring Initializr for the Spring Web project

The web-generated ZIP file archive contains the boilerplate for your web application
using the Spring framework. Once it’s downloaded and extracted from the ZIP file
archive, remove the sample file called Application.java from the src/main/java/com/
example/demo directory and create the HelloController.java file in the same directory
with the code shown in Listing 5-1.

Listing 5-1. The HelloController.java File

package com.example.demo;

import org.springframework.web.bind.annotation.GetMapping;
import org.springframework.web.bind.annotation.RestController;

@RestController
public class HelloController {
@GetMapping("/")

173
Chapter 5 Ansible for K8s Data Plane

public String index() {

return "Ansible For Kubernetes By Examples";
}
}

The code uses the Spring MVC. The HelloController class is flagged as a

@RestController to handle web requests. @GetMapping maps / to the index() method.
When invoked from a browser or by using curl on the command line, the method
returns the text message specified. That is because @RestController combines
@Controller and @ResponseBody, two annotations that result in web requests returning
data rather than a view.
You need to customize the Spring Initializr simple application class file
Application.java in the src/main/java/com/example/demo directory, as shown in
Listing 5-2.

Listing 5-2. The Application.java File

package com.example.demo;
import java.util.Arrays;
import org.springframework.boot.CommandLineRunner;
import org.springframework.boot.SpringApplication;
import org.springframework.boot.autoconfigure.SpringBootApplication;
import org.springframework.context.ApplicationContext;
import org.springframework.context.annotation.Bean;

@SpringBootApplication
public class Application {
        public static void main(String[] args) {
                SpringApplication.run(Application.class, args);
        }
        @Bean
        public CommandLineRunner commandLineRunner(ApplicationCont
ext ctx) {
                return args -> {
                        String[] beanNames = ctx.getBeanDefinitionNames();
                        Arrays.sort(beanNames);
                        for (String beanName : beanNames) {

174
Chapter 5 Ansible for K8s Data Plane

                                System.out.println(beanName);
                        }
                };
        }
}

Note that the main() method, the first method invoked by the Java interpreter,
launches the application using Spring Boot. You can run your application using the
following command in your favorite terminal:

$ gradle bootRun

If you use Maven, run the following command in a terminal window (in the
complete) directory:

$ mvn spring-boot:run

The results are shown in Figure 5-3.

Figure 5-3. Executing the Gradle build tool

You can easily test your application running locally in your workstation and
accessing your API interface connecting via the localhost:8080 interface, as shown in
Figure 5-4.

175
Chapter 5 Ansible for K8s Data Plane

Figure 5-4. Output of the Hello application

When your application is packaged as a JAR file, you can use the Containerfile shown
in Listing 5-3 to deploy as a container image copying all the content of the target/*.jar
directory in the image.

Listing 5-3. The Containerfile

FROM openjdk:21
VOLUME /tmp
ARG JAR_FILE
COPY target/*.jar app.jar
ENTRYPOINT ["java","-jar","/app.jar"]

You can now compile your application as a container using the Maven or Gradle
build tool:

• Maven: $ mvn spring-boot:build-image

• Gradle: $ gradle bootBuildImage

The output of the container build process looks similar to the output shown in
Figure 5-5 provided by the Gradle build tool.

176
Chapter 5 Ansible for K8s Data Plane

Figure 5-5. Output of the Hello application

At the end of the process, the demo image with 0.0.1-SNAPSHOT tag is available in
your workstation. You can analyze the resulting image with Docker Desktop, as shown in
Figure 5-6.

Figure 5-6. Hello container with your application

177
Chapter 5 Ansible for K8s Data Plane

Once the container image is built, you can execute locally using the Docker
command (similar to the command for Podman):

$ docker run --name=demo demo:0.0.1-SNAPSHOT

The execution starts the container with the JVM and the Tomcat web server inside
and produces the output shown in Figure 5-7.

Figure 5-7. Running the containerized Java application

Now you are ready to upload your image to a Container Registry. When your
Container Registry requires authentication, you need to execute the following command:
$ docker login

Now you can rename the image to match the format of your Container Registry
(Docker suggests the default Docker Hub) and perform an upload (push operation) of
your image:

$ docker tag demo:0.0.1-SNAPSHOT lucab85/demo

$ docker push lucab85/demo

A successful operation looks like Figure 5-8. It reports the image digest information
on the screen:

latest: digest: sha256:2b07da21327fc17f9a93f789d4dd1e0b859bab126c91b5b044

af97356772e816 size: 5325
178
Chapter 5 Ansible for K8s Data Plane

Figure 5-8. Upload the container image to Container Registry

After successfully uploading to the Container Registry, you can execute a Kubernetes
deployment of this application. The Kubernetes deployment shown in Listing 5-4
is needed.

Listing 5-4. The hello-deployment.yaml Manifest

apiVersion: apps/v1
kind: Deployment
metadata:
  name: ansiblebyexamples
spec:
  replicas: 2
  selector:
    matchLabels:
      app: ansiblebyexamples
  template:
    metadata:
      labels:
        app: ansiblebyexamples
    spec:
      containers:

179
Chapter 5 Ansible for K8s Data Plane

- name: ansiblebyexamples

          image: demo:0.0.1-SNAPSHOT
          ports:
            - containerPort: 8080
          imagePullPolicy: Always

Listing 5-5. The hello-service.yaml Manifest

apiVersion: v1
kind: Service
metadata:
  name: ansiblebyexamples
spec:
  selector:
    app: ansiblebyexamples
  ports:
    - port: 80
      targetPort: 8080
  type: LoadBalancer

You can execute the Kubernetes YAML manifest file manually via the Ansible
Playbook.

$ kubectl apply -f hello-deployment.yaml

tateless: Deploying PHP Guestbook Application

S
with Redis
Every stateless application is perfectly scalable in Kubernetes. Stateless means that
an instance can be stopped, restarted, or duplicated at any time without any data loss
or inconsistent behavior. You can use the Redis database to store your data structure.
Redis is a well-known key-value database, with in-memory cache and a message broker.
The data durability is optional. Figure 5-9 illustrates the architecture of the Guestbook
application you’ll be deploying using Redis.

180
Chapter 5 Ansible for K8s Data Plane

Guestbook

Guestbook Redis

Guestbook

Figure 5-9. Architecture of the Guestbook application

To deploy a PHP Guestbook application with Redis, you first need to set up a Redis
database server to store the application data. After this is done, you can deploy three
replicas of the frontend Guestbook application in the Kubernetes cluster. Once the
application has been deployed, it can be tested and used by your users.
First of all, you need a Redis leader deployment to store your data using the latest
version (7.0.8) of the official redis image from the Docker Hub Container Registry. This
should run with one replica and listen on the network port 6379.

Listing 5-6. The redis-deployment.yaml Manifest File

apiVersion: apps/v1
kind: Deployment
metadata:
  name: redis-db
  labels:
    app: redis
    role: db
    tier: backend
spec:
  replicas: 1
  selector:
    matchLabels:
      app: redis
  template:
    metadata:
      labels:

181
Chapter 5 Ansible for K8s Data Plane

        app: redis
        role: db
        tier: backend
    spec:
      containers:
        - name: leader
          image: "docker.io/redis:7.0.8"
          resources:
            requests:
              cpu: 100m
              memory: 100Mi
          ports:
            - containerPort: 6379

The Redis Service YAML file is shown in Listing 5-7.

Listing 5-7. The redis-service.yaml Manifest File

apiVersion: v1
kind: Service
metadata:
  name: redis-db
  labels:
    app: redis
    role: db
    tier: backend
spec:
  ports:
    - port: 6379
      targetPort: 6379
  selector:
    app: redis
    role: db
    tier: backend

182
Chapter 5 Ansible for K8s Data Plane

You also need to create a configuration file that defines the application components
and their dependencies, such as databases and messaging systems. Kubernetes will
use this configuration file to create the application pods and services. Once the cluster
is set up and the configuration file is created, you can deploy the application using a
deployment tool such as Kubernetes or Docker. See Listing 5-8.

Listing 5-8. The frontend-deployment.yaml Manifest File

apiVersion: apps/v1
kind: Deployment
metadata:
  name: frontend
spec:
  replicas: 3
  selector:
    matchLabels:
      app: guestbook
      tier: frontend
  template:
    metadata:
      labels:
        app: guestbook
        tier: frontend
    spec:
      containers:
        - name: php-redis
          image: gcr.io/google_samples/gb-frontend:v5
          env:
            - name: GET_HOSTS_FROM
              value: "dns"
          resources:
            requests:
              cpu: 100m
              memory: 100Mi
          ports:
            - containerPort: 80

183
Chapter 5 Ansible for K8s Data Plane

The Guestbook application service looks like Listing 5-9.

Listing 5-9. The frontend-service.yaml Manifest File

apiVersion: v1
kind: Service
metadata:
  name: frontend
  labels:
    app: guestbook
    tier: frontend
spec:
  type: LoadBalancer
  ports:
    - port: 80
  selector:
    app: guestbook
    tier: frontend

Once the application has been deployed, you can test and use it. Finally, you need to
configure the Kubernetes resources to ensure that the application is secure and performs
optimally. One improvement to the architecture is to use some Redis “followers” nodes
between the Redis database and the frontend. With the “followers” improvement, you
can implement the high availability for the single replica Redis database (also called the
“leader”).
You can easily deploy your stateless Guestbook application using the “Apply Multiple
YAML Files at Once on Kubernetes” Ansible Playbook; for more details, see Chapter 4.

Kustomize: Do More with Less

Kustomize is a fantastic tool for generating Kubernetes YAML resources. It saves time
creating usable resource artifacts for your use cases. It creates Kubernetes objects from
a special “kustomization” file. It provides a way to define multiple objects in a single
manifest file, as well as options to customize and override settings on existing objects.
Kustomize also allows you to leverage the power of templating and variables, making it
easier to customize and deploy objects to Kubernetes clusters. Kustomize can be used

184
Chapter 5 Ansible for K8s Data Plane

to manage multiple environments, create complex deployments, and manage resources

across multiple namespaces. You are going to use Kustomize in the following section.
The Kustomize tool has been supported since Kubernetes version 1.14 (kustomize build
flow version 2.0.3).
You can view the resources found in a directory containing a kustomization file by
running the following command:

$ kubectl kustomize <directory>

To apply the resources in your Kubernetes cluster, use the --kustomize or -k

parameter of the kubectl command:

$ kubectl apply –kustomize <directory>

Ansible supports the kubernetes.core.kustomize lookup plugin to execute the

Kustomization directory. The Ansible Playbook shown in Listing 5-10 executes the
customization in the kustomization directory, in the namespace specified in the
myproject variable.

Listing 5-10. The kustomization.yml File

---
- name: k8s ns
  hosts: all
  vars:
    myproject: "ansible-examples"
    kustomize_dir: "kustomization"
  tasks:
    - name: namespace {{ myproject }} present
      kubernetes.core.k8s:
        api_version: v1
        kind: Namespace
        name: "{{ myproject }}"
        state: present
      - name: create kubernetes resource using lookup plugin
        kubernetes.core.k8s:
          namespace: "{{ myproject }}"
          definition: "{{ lookup('kubernetes.core.kustomize',
dir=kustomize_dir) }}"
185
Chapter 5 Ansible for K8s Data Plane

tateful: Deploying WordPress and MySQL

S
with Persistent Volumes
A stateful application requires the persistence of data, contrary to stateless, which
doesn’t. A simple example is the common application based on the web framework
WordPress deployed on top of a MySQL database. To deploy WordPress and MySQL with
Kubernetes, you need persistent volumes to store the database and the application data.
First, you have to create a Kubernetes cluster and configure the necessary resources.
This will enable you to deploy WordPress and MySQL as two separate services. You
need a PersistentVolume to store data with PersistentVolumeClaims declared at the
deployment level. Figure 5-10 illustrates this architecture.

WordPress MySQL

PersistentVolume PersistentVolume

Figure 5-10. Architecture of WordPress with the MySQL application

Next, you create the configuration file for each service. They define the components
of the application and their dependencies, such as databases and messaging systems.
Once the configuration files are created, you can deploy the application to Kubernetes
using a customization file (see Listing 5-11) that links to the Deployment, Service, and
PersistentVolume Kubernetes resources needed by your application (see Listings 5-12
and 5-13).

Listing 5-11. The kustomization.yaml File

secretGenerator:
  - name: mysql-pass
    literals:
      - password=YOUR_PASSWORD
resources:
  - mysql.yaml
  - wordpress.yaml

186
Chapter 5 Ansible for K8s Data Plane

Listing 5-12. The mysql.yaml File

apiVersion: v1
kind: Service
metadata:
  name: wordpress-mysql
  labels:
    app: wordpress
spec:
  ports:
    - port: 3306
  selector:
    app: wordpress
    tier: mysql
  clusterIP: None
---
apiVersion: v1
kind: PersistentVolumeClaim
metadata:
  name: mysql-pv-claim
  labels:
    app: wordpress
spec:
  accessModes:
    - ReadWriteOnce
  resources:
    requests:
      storage: 20Gi
---
apiVersion: apps/v1
kind: Deployment
metadata:
  name: wordpress-mysql
  labels:
    app: wordpress

187
Chapter 5 Ansible for K8s Data Plane

spec:
  selector:
    matchLabels:
      app: wordpress
      tier: mysql
  strategy:
    type: Recreate
  template:
    metadata:
      labels:
        app: wordpress
        tier: mysql
    spec:
      containers:
        - image: mysql:5.7.41
          name: mysql
          env:
            - name: MYSQL_ROOT_PASSWORD
              valueFrom:
                secretKeyRef:
                  name: mysql-pass
                  key: password
          ports:
            - containerPort: 3306
              name: mysql
          volumeMounts:
            - name: mysql-persistent-storage
              mountPath: /var/lib/mysql
      volumes:
        - name: mysql-persistent-storage
          persistentVolumeClaim:
            claimName: mysql-pv-claim

188
Chapter 5 Ansible for K8s Data Plane

Listing 5-13. The wordpress.yaml File

apiVersion: v1
kind: Service
metadata:
  name: wordpress
  labels:
    app: wordpress
spec:
  ports:
    - port: 80
  selector:
    app: wordpress
    tier: frontend
  type: LoadBalancer
---
apiVersion: v1
kind: PersistentVolumeClaim
metadata:
  name: wp-pv-claim
  labels:
    app: wordpress
spec:
  accessModes:
    - ReadWriteOnce
  resources:
    requests:
      storage: 20Gi
---
apiVersion: apps/v1
kind: Deployment
metadata:
  name: wordpress
  labels:
    app: wordpress

189
Chapter 5 Ansible for K8s Data Plane

spec:
  selector:
    matchLabels:
      app: wordpress
      tier: frontend
  strategy:
    type: Recreate
  template:
    metadata:
      labels:
        app: wordpress
        tier: frontend
    spec:
      containers:
        - image: wordpress:6.1.1-apache
          name: wordpress
          env:
            - name: WORDPRESS_DB_HOST
              value: wordpress-mysql
            - name: WORDPRESS_DB_PASSWORD
              valueFrom:
                secretKeyRef:
                  name: mysql-pass
                  key: password
          ports:
            - containerPort: 80
              name: wordpress
          volumeMounts:
            - name: wordpress-persistent-storage
              mountPath: /var/www/html
      volumes:
        - name: wordpress-persistent-storage
          persistentVolumeClaim:
            claimName: wp-pv-claim

190
Chapter 5 Ansible for K8s Data Plane

You can apply the resources in your Kubernetes cluster by manually executing this
command:

$ kubectl apply -k ./

Or you can automate with Ansible using the kubernetes.core.kustomize lookup

plugin to execute the Kustomization directory described in the previous section.
Kubernetes operators often manage complex stateful applications on Kubernetes.
Writing an operator is often a full-fledged software project, using low-level APIs and
writing boilerplate code. The Operator SDK uses the Kubernetes controller-runtime
library to make writing operators easier by providing a high-level API, tools for
scaffolding and code generation, and extensions for use cases. The Operator SDK can
generate the boilerplate for an Ansible Operator but also for Golang and Helm.

Security Namespace (Pod Security Admission)

Enforcing proper Kubernetes security can save you time and protect your organization
from unexpected behavior. The best practice of pod security is useful to apply pod
security standards (pod security policy) at the namespace level. This can be a simple
warning message, or you can’t deny the deployment when the containers don’t
meet the acceptance criteria. You implement the policy at the namespace level. This
policy defines the security rules for the namespace, including who can access what
resources and the types of traffic that are allowed. Previously, the Pod Security Policy
was configured to ensure that the security rules are enforced for the pods within the
namespace. PodSecurityPolicy was previously deprecated in Kubernetes version 1.21
and removed from Kubernetes in version 1.25. Kubernetes implements role-based
access control (RBAC). You can configure RBAC to ensure that only authorized users can
access the resources in the namespace. Once this is done, the Pod Security Standards
will be applied at the Namespace Level. Since Kubernetes version 1.23, Pod Security
Admission (PSA) has been enabled in Kubernetes per cluster and namespace lever by
default. It enables you to use these built-in Pod Security Standards modes:

• enforce (baseline)
• audit (restricted)

• warn (restricted)

191
Chapter 5 Ansible for K8s Data Plane

Using Kubernetes labels of the built-in Pod Security Admission, it is possible to

enable pod security standards at the namespace level. You can use the following label to
set the pod security standard policy:

pod-security.kubernetes.io/<MODE>: <LEVEL>

• MODE sets the pod security standard modes: enforce,

audit, warn

• LEVEL sets the pod security standard levels: privileged,

baseline, or restricted

The following command enables the “warn” pod security standard on the baseline
each time you try to retrieve the “latest” tag:

$ kubectl label --overwrite ns example pod-security.kubernetes.io/

warn=baseline pod-security.kubernetes.io/warn-version=latest

From now on, whenever you try to retrieve a “latest” tag for a pod, you will receive
a warning on output that warns you about the policy violation:

violate PodSecurity "restricted:latest"

The warning message reminds you about the best practices of Kubernetes to specify
a specific tag instead of using the “latest” tag.

Security Pod Resources (AppArmor)

AppArmor is a Linux security tool that works by granting access first rather than applying
restrictions. Sometimes AppArmor is confused with SELinux; the main differences are
listed in Table 5-1.

Table 5-1. AppArmor vs SELinux

Feature AppArmor SELinux

Access control Security profiles based on paths Security policies based on file labels
Linux distributions Mainly used on SUSE and Ubuntu Primarily used on RHEL/Fedora
systems
Level of control Medium High

192
Chapter 5 Ansible for K8s Data Plane

Kubernetes has supported AppArmor since version 1.4. At the moment of writing
this book, Kubernetes’ annotation is the way to use the feature. When promoted to
General Availability (GA), each annotation will be a Kubernetes Object field. AppArmor
is used to restrict a container’s access to resources; you need to create an AppArmor
profile for the container. This profile will define the security rules for the containers, such
as which files and directories it can access, which system calls it can make, and which
other containers or processes it can interact with. Additionally, you need to configure the
container runtime to ensure that the AppArmor profile is applied when the container is
started. Kubernetes-supported container runtimes that support AppArmor technology
are Docker, CRI-O, and containerd. Once this is done, the container will be restricted to
the access rules defined in the AppArmor profile.
AppArmor profiles are specified per pod by adding the following annotation:

container.apparmor.security.beta.kubernetes.io/<container_name>:
<profile_ref>

You can deploy an application protected with AppArmor using the kubectl or
Ansible kubernetes.core.k8s module. There is no native Kubernetes way to load
AppArmor profiles onto nodes. However, you can use Ansible as an initialization script
to enable it.

Security Pod Syscalls (seccomp)

A system call, or syscall, is a direct request from your application to the operating system
where it is executed. The typical use case is accessing hardware resources, launching
threads, communicating with other processes, and interacting with internal kernel
services.
You can use the Linux secure computing mode, called seccomp, to enable a "secure"
state one-way transition, where it is possible to use only exit, sigreturn, read, and
write on an already-open file descriptor system call. Of course, you can restrict the
syscalls of the container creating a seccomp profile for the container. A profile defines
a sandbox of the privileges of a process, restricting the call from the userspace to the
kernel. This profile will define the syscalls that the container can make, as well as the
parameters and arguments that can be used when making the syscalls.

193
Chapter 5 Ansible for K8s Data Plane

Additionally, you need to configure the container runtime at bootstrap to ensure that
the seccomp profile is applied. Once you do this, the container will be restricted to the
syscalls defined in the seccomp profile. Seccomp requires at least Linux kernel 2.6 and
Kubernetes version 1.19. The SCMP_ACT_LOG seccomp profile logs all system calls.
The snippet of a Kubernetes spec YAML manifest shown in Listing 5-14 enables the
seccomp profile audit.json file defined in the localhost node.

Listing 5-14. The seccomp.yaml Manifest File

[...]
  securityContext:
    seccompProfile:
      type: localhost
      localhostProfile: profiles/audit.json
[...]

Listing 5-15. The audit.json seccomp Profile File

{
"defaultAction": "SCMP_ACT_LOG"
}

Ansible Dynamic Inventory

Ansible Dynamic Inventory enables you to generate Ansible Inventories automatically.
The information is read from external sources such as cloud providers, CMDBs,
and inventory management systems. This allows users to deploy infrastructure and
applications dynamically and automate them quickly. The inventory is stored in a YAML
file, which is then used by Ansible to determine which hosts to target for tasks. Ansible
Dynamic Inventory can be used for several different use cases, such as dynamic scaling,
provisioning, and application deployment.

Tip Ansible can also use multiple inventory sources at the same time. You can
mix and match dynamic and statically managed inventory sources in the same
Ansible run.

194
Chapter 5 Ansible for K8s Data Plane

Additionally, Ansible Dynamic Inventory can be used to manage multiple environments

and ensure that the right resources are allocated for each environment. See Listing 5-16.

Listing 5-16. The ansible.cfg File

[inventory]
enable_plugins = kubernetes.core.k8s

The Ansible plugin kubernetes.core.k8s searches for all files with the .k8s.yml or
.k8s.yaml suffixes in the current working directory, as demonstrated in Listing 5-17.

Listing 5-17. The inventory.k8s.yml File

plugin: kubernetes.core.k8s
connections:
- namespaces:
- ansible-examples

You can list all the containers in the namespace ansible-examples using the
inventory.k8s.yml file and the ansible-inventory command:

$ ansible-inventory -i inventory.k8s.yml --list

If the location is given to the -i parameter in Ansible, it’s a directory (or as so

configured in the ansible.cfg file).
The output of the ansible-inventory command displays a JSON file with all the
pods in the specified namespace (or projects when run in OpenShift).

• Namespace: ansible-examples

• list view

            "nginx-689b466988-4z9qq_nginx": {
[...]
                "ansible_kubectl_namespace": "ansible-examples",
                "ansible_kubectl_pod": "nginx-689b466988-4z9qq",
                "container_image": "nginx",
                "container_name": "nginx",
                "container_ready": false,
                "container_state": "Waiting",

195
Chapter 5 Ansible for K8s Data Plane

                "labels": {
                    "app": "nginx",
                    "pod-template-hash": "689b466988"
                },
                "object_type": "pod",
                "pod_host_ip": "192.168.50.15",
                "pod_ip": "10.131.0.3",
                "pod_name": null,
                "pod_node_name": "compute-2",
                "pod_phase": "Running",
                "pod_resource_version": "15542143",
                "pod_uid": "2994bacc-28b0-47fa-b44e-4b28f9ea7d5b"
[...]

The graph view is helpful for displaying in a graph style the pod in the selected
namespaces (or projects when in OpenShift):

$ ansible-inventory -i inventory.k8s.yml --graph

The output of the execution of the ansible-inventory command with the graph
parameter displays a tree style output with all the pods in the specified namespace (or
projects when run in OpenShift).

• Namespace: ansible-examples

• graph view

196
Chapter 5 Ansible for K8s Data Plane

You can execute an Ansible Playbook, for example for remediation in the inventory,
that’s dynamically generated by Ansible Dynamic Inventory.
The easiest Ansible Playbook is ping.yml (see Listing 5-18), which tests the
connection between the Ansible controller and the target nodes.

Listing 5-18. The ping.yml File

---
- name: test
  hosts: all
  gather_facts: false
  tasks:
    - name: test connection
      ansible.builtin.ping:

You can execute the code using the ansible-playbook command combined with the
Ansible Dynamic Inventory configuration file for the ansible-examples namespace:

$ ansible-playbook -i inventory.k8s.yml ping.yml

Successful execution using the ansible-playbook command output includes the

following:

• Play recap status: ok=1 changed=0

• Task status:

TASK [test connection]

ok: [nginx-689b466988-4z9qq_nginx]
ok: [nginx-689b466988-55jnd_nginx]

You might incur the “Failed to create temporary directory” error message
when you try to connect to the remote pods. It is a permission error: “In some cases,
you may have been able to authenticate and did not have permissions on the
target directory”. The message is trying to point out a possible solution: Consider
changing the remote tmp path in ansible.cfg to a path rooted in "/tmp".
The root cause could be in the authentication phase or could be a lack of permission to
create temporary files by the connection user. See Figure 5-11.

197
Chapter 5 Ansible for K8s Data Plane

Ansible target host

Ansible controller
1. authentication

2. create temporary directory

3. copy task generated files

4. execute task generated files

Figure 5-11. Architecture of the Ansible connection

The default behavior for Ansible to connect to any target hosts is to execute the
authentication phase and then copy a file generated by every single task of your
Ansible Playbook. Ansible connects to remote servers and executes code with the same
username. You can specify the connection username with the ansible_user variable
on the inventory. You can change the running user by specifying remote_user in the
playbook or globally in the ansible.cfg file. After a successful connection, Ansible
tries to create a temporary directory in the home of the user of the connection, where
it copies the task files if they doesn’t already exist. If Ansible is unable to create the
temporary directory, if that user does not have a home directory, or if their home
directory permissions do not allow them to write access, you can customize the path of
the temporary directory via the ansible.cfg file. For example, you can use the following
path in the /tmp directory:

remote_tmp = /tmp/.ansible-${USER}/tmp

Ansible pipelining executes the Ansible modules on the target directly without the
prior file transfer, consequently reducing the network operations. Another pleasant
side-effect is the increase in performance when enabled. By default, Ansible pipelining
is disabled. You can enable it using the ANSIBLE_PIPELINING=True environment variable
or setting the pipelining=true key in the [connection] and [defaults] sections of the
ansible.cfg file.

198
Chapter 5 Ansible for K8s Data Plane

Key Takeaways
You can develop the application of tomorrow by applying Agile methodologies. Similarly,
you can deploy modern cloud-native applications in the Kubernetes cluster to obtain a
faster workflow and more reliable applications or services for your customers. You can
also mix and match stateless and stateful applications and apply security policies to the
pod as desired. Some applications require a simple microservice architecture, whereas
others require a more complex design. Using Kubernetes, you can focus on the high
availability, scale your workflow, and increase the velocity with Ansible.

199
CHAPTER 6

Ansible for K8s
Management
Ansible for Kubernetes helps organizations reduce operational costs by automating
many manual processes associated with deploying and managing large-scale distributed
systems like those used in modern production environments. This includes tasks such
as provisioning nodes, configuring network settings, and setting up storage volumes for
persistent data stores across multiple nodes in a cluster environment.
All this can be accomplished quickly using Ansible’s easy-to-use playbooks written
in the YAML format. Overall, Ansible provides great advantages when implementing
Kubernetes in an organization’s IT architecture, due to its ease of use combined
with advanced capabilities. It allows developers and system administrators alike to
create complex configurations quickly while ensuring consistency across different
infrastructures regardless of whether they are located locally or remotely hosted in
public clouds providers such Amazon Web Services (AWS) EC2 instances or other cloud
providers running Linux operating system distributions like Ubuntu LTS Server Edition.
Modern cloud-native applications embrace the workflow shown in Figure 6-1.

DEV SIT UAT PROD

Figure 6-1. The DEV, SIT, UAT, and PROD workflow

• DEV: Development (software development team)

• SIT: System Integration Test (software developer and QA engineer)

• UAT: User Acceptance Test (selected clients)

• PROD: Production (public users)

201
© Luca Berton 2023
L. Berton, Ansible for Kubernetes by Example, https://doi.org/10.1007/978-1-4842-9285-3_6
Chapter 6 Ansible for K8s Management

Organizations embrace the DEV, SIT, UAT, and PROD workflow in order to take
advantage of automating tasks using the Ansible technology and to have a consistent
environment at every stage of the execution and save time.
Managing multiple Kubernetes clusters to create the four environments needed for
DEV, SIT, UAT, and PROD workflow can be a daunting process that burns IT resources
in your IT department. Automating the process with Ansible enables you to have a
consistent environment between the four stages and focus on the distribution of cloud-
native applications to your target audience and stakeholders.

The Helm Package Manager

The Helm Package Manager, originally created by Microsoft and written in the GO
language, is available as an open-source project and has become the most convenient
way to package and distribute applications for Kubernetes. A package includes all the
necessary objects for a Kubernetes application to run. Helm is a graduate project of the
Cloud Native Computing Foundation.
Helm is constantly used for installing, upgrading, and publishing packages in the
Kubernetes cluster. The major advantages of the Helm software are as follows:

• Create standard and reusable templates

• Eliminate deploy errors

You can easily search for all the available Helm packages using the directory called
Artifact Hub https://artifactshub.io/ (there are 9161 results at the moment of
writing this book). Helm retrieves software from packages called “charts” in Helm slang.
The packages are distributed in repositories.
To use Helm, you must install the helm command-line tool in your workstation. You
can find that tool on the official website https://helm.sh/. Helm is available for most
modern operating systems: Linux (i386, amd64, arm, arm64,ppc64le, and s390x), macOS
(Intel and Apple Silicon), and Windows (amd64). It’s also possible to install Helm using
the most common package managers: Homebrew, Chocolatey, Scoop, GoFish, and Snap.
To report the Helm packages deployed inside the cluster, you can use the manual
helm list command. This command displays all the Helm packages installed in the
current namespace, as well as their versions and other information.

202
Chapter 6 Ansible for K8s Management

You can also use the helm status command to get detailed information about a
specific package, such as its deployed version, release history, and a list of resources that
were created when the package was deployed.
Additionally, you can use the helm history command to view a list of the releases
that were installed or deleted in the current namespace. Additionally, the helm search
command can be used to search for packages in the repositories that are currently
configured.
The kubernetes.core Ansible collection enables you to automate the execution
of tasks in Helm and obtain the same information in the Ansible way. The following
modules are specific to Helm-related tasks:

• helm: Add, update, and delete Kubernetes packages using the

package manager Helm

• helm_info: Obtain information about a Helm package deployed into

the Kubernetes cluster

• helm_plugin: Add, update, and delete Helm plugins

• helm_plugin_info: Obtain information about your Helm plugins

• helm_repository: Add, update and delete Helm repositories

• helm_template: Render Helm chart templates

In the following section, you are going to apply the Ansible modules to some
practical examples.
Ansible Helm modules rely on the helm command-line utility that should be
installed in your system. When Ansible is unable to use the helm utility, the execution of
the code terminates with the following fatal runtime error:

"msg": "Failure when executing Helm command. Exited 1

In most cases, it’s simply necessary to install the helm utility. When some
dependencies are missing, the message on the screen shows the exact library missing:

/bin/helm\", line 24, in <module>\n import glib\nModuleNotFoundError: No

module named 'glib'\n"

The message points out that the glib Python library is missing on your system. You
could install the following:

$ pip install glib

203
Chapter 6 Ansible for K8s Management

Helm Repositories
A Helm repository enables you to simply install packages (“charts” in Helm slang) just
as a package manager does in an operating system. The repository is a web server that
houses an index.yaml file and some packaged charts. To manage Helm repositories,
you can use the manual helm repo command or the Ansible kubernetes.core.helm_
repository module. This command and the Ansible module allow you to add, remove,
and list Helm repositories.
The most famous and widely used repository is the Bitnami repository. Bitnami
is famous for the library of installers and software packages for web applications and
software stacks, as well as virtual appliances, acquired by VMWare in May, 2019. The
Bitnami repository hosts packages like Redis, PostgreSQL, MySQL, ExternalDNS,
RabbitMQ, Apache Kafka, Keycloak, MongoDB, MiniO, Metrics Server, WordPress,
Thanos, MariaDB, NGINX, ElasticSearch, MetailLB, InfluxDB, Apache ZooKeeper,
Fluentd, Prometheus, Cassandra, Grafana, Harbor, Apache Spark, MariaDB Galera,
Kibana, Odoo, Drupal, Contour, Memcached, Apache Tomcat, Apache AirFlow, Jenkins,
Redmine, JupyterHub, Apache, Argo CD, HashiCorp Consul, MediaWiki, Moodle,
Logstash, PyTorch, DokuWiki, Magento, Matomo, Prestashop, Apache Solr, ASP .NET
Core, Kong, Apache MXNet, Parse, SonarQube, Joomla, EJBCA, HAProxy, OpenCart,
phpBB, TensorFlow, Appsmith, Jaeger, Gitea, Pinniped, and more.

Add Helm Repository

You can add a new Helm repository manually by using the helm repo command:

$ helm repo add <repo-name> <repo-url>

You can automate the Helm repository management using the kubernetes.core.
helm_repository Ansible module.
The following helm_repo_present.yml Ansible Playbook adds the Bitnami
repository to your cluster or verifies that it is already present. When you are dealing
with a private repository, you can also specify the path of a Certification Authority
certificate (ca_cert parameter) and the API key (api_key parameter), enable the
domain credentials (pass_credentials parameter), and specify basic authentication
(repo_username and repo_password parameters) or a specific path for the Helm binary

204
Chapter 6 Ansible for K8s Management

(binary_path parameter). The full Ansible Playbook looks like the following for the
bitnami public Helm repository. Simply customize the helm_chart_name and helm_
chart_url variables for another Helm repository, as demonstrated in Listing 6-1.

Listing 6-1. The helm_repo_present.yml File

---
- name: k8s helm repo
  hosts: all
  vars:
    helm_chart_name: "bitnami"
    helm_chart_url: "https://charts.bitnami.com/bitnami"
  tasks:
    - name: helm repo present
      kubernetes.core.helm_repository:
        name: "{{ helm_chart_name }}"
        repo_url: "{{ helm_chart_url }}"
        repo_state: present

You can execute the code using the ansible-playbook command and specifying the
inventory file and the playbook filename:

$ ansible-playbook -i inventory helm_repo_present.yml

When you want to execute the Playbook code in your Ansible Controller, you specify
the localhost in the inventory, as shown in Listing 6-2.

Listing 6-2. Inventory

localhost ansible_connection=local

A successful execution using the ansible-playbook command output includes the

following:

• Play recap status: ok=2 changed=1

• Task status:

TASK [helm repo present]

changed: [localhost]

205
Chapter 6 Ansible for K8s Management

After executing the Ansible Playbook, your Kubernetes cluster has the bitnami
repository in place. You can confirm this using the manual helm command:

$ helm repo list

The report shows you the name and URL of the installed Helm repository in the
Kubernetes cluster.

Remove Helm Repository

To remove a repository, you can use the manual helm repo command:

$ helm repo remove <repo-name>

Otherwise, you can automate this process using the Ansible Playbook in your
Kubernetes cluster. The helm_repo_absent.yml Ansible Playbook shown in Listing 6-3
removes the bitnami Helm repository from the Kubernetes cluster.

Listing 6-3. The helm_repo_absent.yml File

---
- name: k8s helm repo
  hosts: all
  vars:
    helm_chart_name: "bitnami"
  tasks:
    - name: helm repo removed
      kubernetes.core.helm_repository:
        name: "{{ helm_chart_name }}"
        repo_state: absent

You can execute this in your Ansible Controller, specifying the localhost in the
inventory (see Listing 6-4).

Listing 6-4. Inventory

localhost ansible_connection=local

You can execute the code using the ansible-playbook command and specifying
the -i parameter for the inventory file and the playbook filename:

$ ansible-playbook -i inventory helm_repo_absent.yml

206
Chapter 6 Ansible for K8s Management

A successful execution using the ansible-playbook command output includes the

following:

• Play recap status: ok=2 changed=1

• Task status:

TASK [helm repo removed]

changed: [localhost]

After successful execution of the Ansible Playbook, your Kubernetes cluster has
removed the bitnami repository. You can confirm this using the manual helm command:

$ helm repo list

The output of the command shows the name and URL of the installed Helm
repositories in your Kubernetes cluster.

Helm Packages
When a Helm repository is configured in your Kubernetes cluster, you can install
any Helm packages you need. Many Helm charts are available worldwide. The Helm
packages are also called “charts”. You can use the manual helm command or the Ansible
kubernetes.core.helm module.

Install Helm Package

The following Ansible Playbook installs the popular NGINX web server using the nginx
charts from the Bitnami repository. You can enable this repository using the code in the
previous section. Feel free to customize the variable with the one in your Kubernetes
cluster:

• chart_name: The name of the chart to install (nginx-server)

• chart_ref: The name of the repository to retrieve the chart

(bitnami/nginx)
• myproject: The name of the target Kubernetes namespace/
OpenShift Project (ansible-examples)

207
Chapter 6 Ansible for K8s Management

The full Ansible Playbook is shown in Listing 6-5.

Listing 6-5. The helm_present.yml File

---
- name: k8s helm
  hosts: all
  vars:
    chart_name: "nginx-server"
    chart_ref: "bitnami/nginx"
    myproject: "ansible-examples"
  tasks:
    - name: helm chart present
      kubernetes.core.helm:
        name: "{{ chart_name }}"
        namespace: "{{ myproject }}"
        chart_ref: "{{ chart_ref }}"
        release_state: present

You can execute the code in your Ansible Controller using the localhost inventory,
as demonstrated in Listing 6-6.

Listing 6-6. Inventory

localhost ansible_connection=local

You can execute your helm_present.yml Ansible Playbook using the ansible-
playbook command and specifying the -i parameter for the inventory file and the
playbook filename:

$ ansible-playbook -i inventory helm_present.yml

A successful execution using the ansible-playbook command output includes the

following:

• Play recap status: ok=2 changed=1

• Task status:

208
Chapter 6 Ansible for K8s Management

TASK [helm chart present]

changed: [localhost]

Taking advantage of the Ansible idempotency property when you execute your
code with the desired state already present, the code returns only an ok status. You can
verify the installed status of the Helm “chart” package in your Kubernetes cluster in the
following section (see the Report Helm Package section).

Remove Helm Package

In the same way as you install a Helm “chart” package, you can automate its removal
using the Ansible Playbook in your Kubernetes cluster. Feel free to customize the
variable with the one in your Kubernetes cluster:

• chart_name: The name of the chart to install (nginx-server)

• myproject: The name of the target Kubernetes namespace/

OpenShift Project (ansible-examples)

The full Ansible Playbook is shown in Listing 6-7.

Listing 6-7. The helm_absent.yml File

---
- name: k8s helm
  hosts: all
  vars:
    chart_name: "nginx-server"
    myproject: "ansible-examples"
  tasks:
    - name: helm chart absent
      kubernetes.core.helm:
        name: "{{ chart_name }}"
        namespace: "{{ myproject }}"
        release_state: absent

When you execute the code in your Ansible Controller, you use the localhost
inventory (see Listing 6-8).

209
Chapter 6 Ansible for K8s Management

Listing 6-8. Inventory

localhost ansible_connection=local

The ansible-playbook command executes your code and specifies the -i parameter
for the inventory file and the helm_absent.yml filename:

$ ansible-playbook -i inventory helm_absent.yml

A successful execution using the ansible-playbook command output includes the

following:

• Play recap status: ok=2 changed=1

• Task status:

TASK [helm chart absent]

changed: [localhost]

You can verify the removed status of the Helm “chart” package in your Kubernetes
cluster in the following section (see Report Helm Package).

Report Helm Package

You can verify the status of the installed Helm “chart” packages using the manual helm
command or the Ansible kubernetes.core.helm_info module. It might be useful to the
release_state parameter (default to deployed and failed):
• all: Show all releases

• deployed: Show only deployed releases

• failed: Show only failed releases

• pending: Show only pending releases

• superseded: Show only superseded releases

• uninstalled: Show only uninstalled releases

• uninstalling: Show only releases that are currently being

uninstalled

210
Chapter 6 Ansible for K8s Management

Feel free to customize the variable of the following code:

• chart_name: The name of the chart to install (nginx-server)

• myproject: The name of the target Kubernetes namespace/

OpenShift Project (ansible-examples)

The full code of the Ansible Playbook shown in Listing 6-9 displays the Helm charts.

Listing 6-9. The helm_info.yml File

---
- name: k8s helm
  hosts: all
  vars:
    chart_name: "nginx-server"
    myproject: "ansible-examples"
  tasks:
    - name: gather chart information
      kubernetes.core.helm_info:
        name: "{{ chart_name }}"
        release_namespace: "{{ myproject }}"

The localhost inventory shown in Listing 6-10 executes the code in the Ansible
Controller.

Listing 6-10. Inventory

localhost ansible_connection=local

You can execute the code using the ansible-playbook command and specifying the
inventory file and the playbook filename:

$ ansible-playbook -i inventory helm_info.yml

211
Chapter 6 Ansible for K8s Management

A successful execution using the ansible-playbook command output includes the

following:

• Play recap status: ok=2 changed=1

• Task status:

TASK [gather chart information]

changed: [localhost]

The output of the helm_info.yml playbook shows information about the installed
“chart” packages: application version, chart version, revision, status, and updated
date time.

Helm Plugins
The Helm plugins are software add-ons that expand Helm’s functionalities.
To report the Helm plugins deployed inside the cluster manually, you can use the
helm plugin list command. This command will list all the plugins installed in the
current namespace, as well as their versions and other information.
Additionally, you can use the helm plugin install command to install a plugin and
the helm plugin update command to update a plugin. The helm plugin uninstall
command can be used to uninstall a plugin. Finally, the helm plugin status command
can be used to get detailed information about a specific plugin. You can automate Helm
plugin management using the kubernetes.core.helm_plugin Ansible module.

Install Helm Plugin

Many Helm plugins expand Helm functionality with new pieces of software. For
example, the Helm plugin called helm env allows users to view the environment
variables. The following Ansible Playbook downloads the latest release of the helm env
plugin and installs it in your Kubernetes cluster. Feel free to customize the plugin path
and the Kubernetes namespace/OpenShift project variables:
• myplugin_name: The name of the plugin to install (env)

• myplugin_path: The name of the repository to retrieve the plugin

(https://github.com/adamreese/helm-env)

212
Chapter 6 Ansible for K8s Management

The full Ansible Playbook is shown in Listing 6-11.

Listing 6-11. The helm_plugin_present.yml file

---
- name: k8s helm plugin
  hosts: all
  vars:
    myplugin_name: "env"
    myplugin_path: "https://github.com/adamreese/helm-env"

  tasks:
    - name: helm plugin present
      kubernetes.core.helm_plugin:
        plugin_name: "{{ myplugin_name }}"
        plugin_path: "{{ myplugin_path }}"

state: present

The localhost inventory executes your code in the Ansible Controller

(see Listing 6-12).

Listing 6-12. Inventory

localhost ansible_connection=local

You can execute the code with the ansible-playbook command:

$ ansible-playbook -i inventory helm_present.yml

A successful execution using the ansible-playbook command output includes the

following:

• Play recap status: ok=2 changed=1

• Task status:

TASK [helm plugin present]

changed: [localhost]

213
Chapter 6 Ansible for K8s Management

You can verify the installed Helm plugin in the following section (see Report Helm
Plugin).
A possible improvement to the code is to use a list of namespaces and loop them
with the Ansible Playbook.

Remove Helm Plugin

When a plugin is not needed anymore in your Kubernetes cluster, you can automate
the removal process using the kubernetes.core.helm_plugin Ansible module. The full
Ansible Playbook is shown in Listing 6-13.

Listing 6-13. The helm_plugin_absent.yml File

---
- name: k8s helm plugin
  hosts: all
  vars:
    myplugin_name: "env"

  tasks:
    - name: helm plugin absent
      kubernetes.core.helm_plugin:
        plugin_name: "{{ myplugin_name }}"

state: absent

The localhost inventory executes your code in the Ansible Controller, as shown in
Listing 6-14.

Listing 6-14. Inventory

localhost ansible_connection=local

You can execute this code with the ansible-playbook command:

$ ansible-playbook -i inventory helm_absent.yml

214
Chapter 6 Ansible for K8s Management

A successful execution using the ansible-playbook command output includes the

following:

• Play recap status: ok=2 changed=1

• Task status:

TASK [helm plugin absent]

changed: [localhost]

You can verify the installed Helm plugin in the following section (see Report Helm
Plugin).

Report Helm Plugin

You can use the helm plugin list manual command to report the Helm plugins
deployed inside the cluster. This command lists all the plugins installed in the current
namespace, as well as their versions and other information. You can automate this
process using the kubernetes.core.helm_plugin_info Ansible module.
The Ansible Playbook shown in Listing 6-15 displays the version of the env plugin
installed in the ansible-examples namespace.

Listing 6-15. The helm_plugin_info.yml File

---
- name: k8s helm plugin
  hosts: all
  vars:
    myplugin_name: "env"
  tasks:
    - name: helm plugin info
      kubernetes.core.helm_plugin_info:
      register: output

- name: print plugin version

ansible.builtin.debug:
msg: "{{ (output.plugin_list | selectattr('name', 'equalto',
myplugin_name) | list)[0].version }}"

215
Chapter 6 Ansible for K8s Management

You can execute the code in your Ansible Controller using the localhost inventory,
as shown in Listing 6-16.

Listing 6-16. Inventory

localhost ansible_connection=local

You can execute your helm_present.yml Ansible Playbook using the ansible-
playbook command and specifying the -i parameter for the inventory file and the
playbook filename:

$ ansible-playbook -i inventory helm_plugin_info.yml

A successful execution using the ansible-playbook command output includes the

following:

• Play recap status: ok=3 changed=0

• Task status:

TASK [helm plugin info]

   ok: [localhost]
   TASK [print plugin version]
   ok: [localhost]

Deploy a Monitoring Tool

Prometheus is a popular monitoring tool based on time series data. Under the hood, it
streams timestamped values belonging to the same metric and the same set of labeled
dimensions. One of the strengths of Prometheus is its deep integration with Kubernetes.
You can choose between several different ways to deploy a monitoring tool inside
your Kubernetes clusters:

• Prometheus Operator and kube-prometheus

• Ansible role cloudalchemy.prometheus

• Helm chart

216
Chapter 6 Ansible for K8s Management

The Prometheus Operator is the Kubernetes-native solution that allows you to

deploy and manage Prometheus and its components, such as Alertmanager and
Grafana. It provides a declarative configuration that allows you to specify the desired
state of your monitoring stack, and the operator will ensure that it is always kept up-
to-date. In addition to deploying Prometheus, the operator can be used to configure
monitoring rules, manage alerting rules, and set up dashboards. To deploy the
Prometheus Operator, you can use different options, discussed next.

kube-prometheus
The kube-prometheus project is the easiest way to install end-to-end Kubernetes cluster
monitoring with Prometheus using the Prometheus Operator. It deploys the Prometheus
Operator by deploying it as part of kube-prometheus. The kube-prometheus project
deploys the Prometheus Operator and schedules a Prometheus job called prometheus-
k8s with alerts and rules by default. The default deployment includes multiple
Prometheus and Alertmanager instances, metric exporters such as node_exporter for
gathering node metrics, scrape target configurations linking Prometheus to various
metric endpoints, and example alerting rules for notification of potential issues in the
cluster. The setup is performed in three manual steps:

1. Download the kube-prometheus project:

$ git clone https://github.com/prometheus-operator/

kube-prometheus.git

2. Deploy kube-prometheus:

$ kubectl create -f manifests/setup

$ kubectl create -f manifests/

3. Forward the ports for Prometheus, Alertmanager, and Grafana:

$ kubectl --namespace monitoring port-forward svc/

prometheus-k8s 9090
$ kubectl --namespace monitoring port-forward svc/
prometheus-k8s 9090
$ kubectl --namespace monitoring port-forward svc/
alertmanager-main 9093
$ kubectl --namespace monitoring port-forward svc/grafana 3000

217
Chapter 6 Ansible for K8s Management

Ansible Collections
When you want to automate the installation of Prometheus, you can deploy it using
Ansible. You need to acquire the Ansible Collection called prometheus.prometheus from
the Ansible Galaxy repository. The collection requires the community.crypto collection
to be installed (resolved by the ansible-galaxy tool). The Prometheus role inside the
collection substitutes the previous cloudalchemy.prometheus Ansible role, which has
been deprecated in Ansible Galaxy.

1. Install the collection manually:

$ ansible-galaxy collection install prometheus.prometheus

2. Install the collection via requirements.yml, as shown in

Listing 6-17.

Listing 6-17. The requirements.yml File

---
collections:
- name: prometheus.prometheus

The role requires the jmespath Python library and gnu-tar when run on an Ansible
Controller, as demonstrated in Listing 6-18.

Listing 6-18. The prometheus.yml File

---
- name: install prometheus
  hosts: all
  roles:
    - prometheus.prometheus.prometheus
  vars:
    prometheus_targets:
      node:
        - targets:
            - k8s.ansiblepilot.com:9100
          labels:
          env: ansible-examples

218
Chapter 6 Ansible for K8s Management

The localhost inventory executes your code in the Ansible Controller

(see Listing 6-19).

Listing 6-19. Inventory

localhost ansible_connection=local

You can execute the code using the ansible-playbook command and specifying the
inventory file and the playbook filename:

$ ansible-playbook -i inventory prometheus.yml

A successful execution using the ansible-playbook command output includes the

following:

• Play recap status: ok=2 changed=1

• Many messages according to the prometheus.prometheus.

prometheus role

Helm Chart
Another way to install the Prometheus Operator is via a Helm Chart. You can reuse the
Ansible Playbook seen in the previous section (see The Helm Package Manager) with
different Ansible variables. You can specify the extra variables via the command line
using the -e parameter of the ansible-playbook command. The extra variables override
the values of the Ansible Playbooks.
Enable the Helm repository called prometheus-community by specifying the URL:

$ ansible-playbook -i inventory -e "helm_chart_name=prometheus-community"

-e "helm_chart_url=https://prometheus-community.github.io/helm-charts"
helm_repo_present.yml

Install the Helm chart package named kube-prometheus-stack from the

prometheus-community Helm repository as follows:

$ ansible-playbook -i inventory -e "chart_name=kube-prometheus-stack" -e

"chart_ref=prometheus-community/kube-prometheus-stack" helm_present.yml

219
Chapter 6 Ansible for K8s Management

Fetch Logs from Resources

To fetch logs manually from Kubernetes resources, you can use the kubectl logs
command. This command allows you to view the logs of a container in a pod or to
retrieve the logs of a resource such as a Deployment, StatefulSet, or CronJob. The syntax
for the command is as follows:

$ kubectl logs <resource-name> [<container-name>]

If a container name is not provided, the logs of the first container in the pod will be
fetched. For example, to view the logs of a Deployment, you can use the kubectl logs
my-deployment command. This will show the logs for all of the pods in the Deployment.
To view the logs of a specific container in a pod, you can use the $ kubectl logs
my-pod -c my-container command.
The Ansible Playbook shown in Listing 6-20 obtains the log from the ansible-
examples pod in the ansible-examples namespace.

Listing 6-20. The log_pod.yml File

---
- name: k8s log
  hosts: all
  vars:
    myproject: "ansible-examples"
    pod_name: "nginx-server"
  tasks:
    - name: get log from pod
      kubernetes.core.k8s_log:
        name: "{{ pod_name }}"
        namespace: "{{ myproject }}"
      register: log

- name: display log

ansible.builtin.debug:
var: log

220
Chapter 6 Ansible for K8s Management

Instead, the Ansible Playbook in Listing 6-21 can obtain logs from different
resources: Deployment, DeploymentConfig, StatefulSet, and CronJob.

Listing 6-21. The log_resource.yml File

---
- name: k8s log
  hosts: all
  vars:
    myproject: "ansible-examples"
    myname: "nginx-server"
    mykind: "Deployment"
  tasks:
    - name: get log from resource
      kubernetes.core.k8s_log:
        api_version: apps/v1
        kind: "{{ mykind }}"
        namespace: "{{ myproject }}"
        name: "{{ myname }}"
      register: log

- name: display log

ansible.builtin.debug:
var: log

The localhost inventory executes your code in the Ansible Controller, as shown in
Listing 6-22.

Listing 6-22. The Inventory File

localhost ansible_connection=local

You can execute your code with the ansible-playbook command (specifying the
log_pod.yml or log_resource.yml file):

$ ansible-playbook -i inventory log_pod.yml

221
Chapter 6 Ansible for K8s Management

Successful execution using the ansible-playbook command includes the following:

• Play recap status: ok=2 changed=1

• Log messages displayed onscreen

Apply a JSON Patch Operation

JSON Patch is a format for describing changes to a JSON document. To apply JSON patch
operations to existing Kubernetes objects manually, you can use the kubectl patch
command. This command takes two arguments—the object name and a JSON patch file.
The patch file should contain one or more operations, in either JSON or YAML format,
that you want to apply to the object. For example, to add a new label to an existing pod,
you can create a patch file containing the YAML shown in Listing 6-23.

Listing 6-23. The patch-file.yaml File

- op: add
path: /metadata/labels/new-label
value: new-value

Then, you can apply this patch to the nginx-server pod object using the following
command:

$ kubectl patch pod nginx-server --patch "$(cat patch-file.yaml)"

This will add the new label to the pod. You can use the same command with different
patch files to make changes to other objects in the cluster.
You can also automate the previous operation using the Ansible Playbook shown in
Listing 6-24.

Listing 6-24. The patch.yml File

---
- name: k8s patch
  hosts: all
  vars:
    mypod: "nginx-server"
    myproject: "ansible-examples"

222
Chapter 6 Ansible for K8s Management

  tasks:
    - name: patch a Pod
      kubernetes.core.k8s_json_patch:
        kind: Pod
        namespace: "{{ myproject }}"
        name: "{{ mypod }}"
        patch:
          - op: add
            path: /metadata/labels/new-label
            value: new-value

The localhost inventory executes this code in the Ansible Controller, as shown in
Listing 6-25.

Listing 6-25. The Inventory File

localhost ansible_connection=local

You can execute the patch.yml Ansible Playbook with the ansible-playbook
command:

$ ansible-playbook -i inventory patch.yml

A successful execution using the ansible-playbook command output includes the

following:

• Play recap status: ok=2 changed=1

• Task status:

TASK [patch a Pod]

changed: [localhost]

223
Chapter 6 Ansible for K8s Management

Copy Files and Directories to and from a Pod

It’s helpful to copy files and directories to and from a pod in Kubernetes, especially in
the early stage of the development or for some investigations. For manual operations,
you can use the kubectl cp command. This command allows you to copy files and
directories between a pod and the local file system. The syntax for the command is
kubectl cp <source> <destination>, where the source can be either a pod or the
local file system, and the destination can be either a pod or the local file system. For
example, to copy a file from a pod to the local file system, you can use this command:

$ kubectl cp <pod-name>:/path/to/file /local/destination/path

Similarly, to copy a file from the local file system to a pod, you can use this
command:

$ kubectl cp /local/source/path <pod-name>:/path/to/destination

You can automate the operation using the kubernetes.core.k8s_cp Ansible

module. The parameter state defines whether you want the default operation to copy
data from local to pod (the to_pod option) or from pod to local (the from_pod option).
The Ansible Playbook in Listing 6-26 copies data from the /data directory of the pod
named nginx-server in the ansible-examples namespace to the local folder called
/tmp/data.

Listing 6-26. The cp.yml File

---
- name: k8s copy
  hosts: all
  vars:
    mypod: "nginx-server"
    myproject: "ansible-examples"
    remote_path: "/data"
    local_path: "/tmp/data"
    direction: "from_pod"
  tasks:
    - name: copy data
      kubernetes.core.k8s_cp:

224
Chapter 6 Ansible for K8s Management

namespace: "{{ myproject }}"

        pod: "{{ mypod }}"
        remote_path: "{{ remote_path }}"
        local_path: "{{ local_path }}"
        state: "{{ direction }}"

You can execute the code in your Ansible Controller using the localhost inventory,
as shown in Listing 6-27.

Listing 6-27. Inventory

localhost ansible_connection=local

You can execute your cp.yml Ansible Playbook by using the ansible-playbook
command and specifying the -i parameter for the inventory file and the playbook
filename:

$ ansible-playbook -i inventory cp.yml

A successful execution using the ansible-playbook command output includes the

following:

• Play recap status: ok=2 changed=1

• Task status:

TASK [copy data]

changed: [localhost]

Manage Services on Kubernetes

To manage services on Kubernetes, you can use the kubectl command-line tool. This
tool lets you create, edit, delete, and view services in your Kubernetes cluster.
To create a service, you can use the kubectl expose command, which takes a
deployment or a pod as an argument.
To edit a service, you can use the kubectl edit command, which will open the
service definition in an editor.

225
Chapter 6 Ansible for K8s Management

To delete a service, you can use the kubectl delete command, which takes the
service name and namespace as arguments.
To view a service, you can use the kubectl get command, which takes the service
name and namespace as arguments.
You can also use the kubectl describe command to view detailed information
about a service.
You can automate the managing of services on the Kubernetes cluster using the
following Ansible Playbook (see Listing 6-28).

Listing 6-28. The service.yml File

---
- name: k8s service
  hosts: all
  tasks:
    - name: expose https port with ClusterIP
      kubernetes.core.k8s_service:
        state: present
        name: port-https
        namespace: default
        ports:
          - port: 443
            protocol: TCP
            selector:
              key: special

You can execute the code in your Ansible Controller by specifying the localhost in
the inventory, as shown in Listing 6-29.

Listing 6-29. Inventory

localhost ansible_connection=local

You can execute the code using the ansible-playbook command and specifying
the -i parameter for the inventory file and the playbook filename:

$ ansible-playbook -i inventory service.yml

226
Chapter 6 Ansible for K8s Management

A successful execution using the ansible-playbook command output includes the

following:

• Play recap status: ok=2 changed=1

• Task status:

TASK [expose https port with ClusterIP]

changed: [localhost]

Taint Nodes
Taints and tolerations are a flexible way to steer pods away from nodes or evict pods
that shouldn’t be running. A typical use case of the taint node is to take advantage of
particular nodes with special hardware (such as GPUs) to run only pods that require the
use of the specialized hardware and keep out the pods that don’t require it. Moreover,
you can use this tool when you need to execute maintenance in your Kubernetes cluster.
Kubernetes Taint is used to mark a node as unable to schedule any pods that do not
tolerate the taint. This is useful when you need to perform maintenance on a node, as
it ensures that any existing pods running on the node will remain running and any new
pods will be rescheduled onto other nodes in the cluster. The taint command works
by adding a label to the node, which is then used to identify which pods should not be
scheduled onto the node. This label can be removed when the maintenance is complete,
allowing the node to start accepting new pods again.
The Ansible Playbook in Listing 6-30 taints the node defined in the mynode variable
and sets the NoExecute and NoSchedule attributes’ keys and values. The playbook uses
the kubernetes.core.k8s_taint Ansible module, introduced in version 2.3 of the
kubernetes.core Ansible collection in March, 2022.

Listing 6-30. The taint.yml File

---
- name: k8s taint
  hosts: all
  vars:
    mynode: "k8s.ansiblepilot.com"

227
Chapter 6 Ansible for K8s Management

  tasks:
    - name: taint node
      kubernetes.core.k8s_taint:
        state: present
        name: "{{ mynode }}"
        taints:
          - effect: NoExecute
            key: "key1"
            value: "value1"
          - effect: NoSchedule
            key: "key1"
            value: "value1"

You can execute your Ansible Controller by specifying the localhost in the
inventory, as shown in Listing 6-31.

Listing 6-31. Inventory

localhost ansible_connection=local

You can execute the code using the ansible-playbook command and specifying the -i
parameter for the inventory file and the playbook filename:

$ ansible-playbook -i inventory taint.yml

A successful execution using the ansible-playbook command output includes the

following:

• Play recap status: ok=2 changed=1

• Task status:

TASK [taint node]

changed: [localhost]

228
Chapter 6 Ansible for K8s Management

Drain, Cordon, or Uncordon Nodes

Kubernetes Drain is a tool used to safely evict all of the pods from a node before
performing maintenance on the node. Note that it’s always a good practice to have a
planned maintenance window. The Drain operation is often combined with the Taint
operation.
When you run the drain command, Kubernetes will attempt to delete all of the pods
on the node except for mirror pods, which are created by replication controllers, jobs, or
daemon sets. The drain command also allows you to specify a PodDisruptionBudget,
which will ensure that the number of pods running on the node does not drop below the
specified minimum during the eviction process. Once the eviction process is complete,
you can then safely perform maintenance on the node, knowing that the pods will be
safely relocated to other nodes in the cluster.
Kubernetes Cordon marks a node as “unschedulable,” meaning that no new pods
will be scheduled onto the node. This is useful when you need to perform maintenance
on a node, as it ensures that any existing pods running on the node will remain running
and any new pods will be rescheduled onto other nodes in the cluster. The cordon
command works by setting the node’s unschedulable flag to true, which will prevent the
Kubernetes scheduler from placing any new pods on the node. Once the maintenance
is complete, the unschedulable flag can be reset back to false, allowing the node to start
accepting new pods again.
Kubernetes Uncordon is used to mark a node as schedulable, meaning that new
pods can be scheduled onto the node. This is the opposite of the kubectl cordon
command, which marks a node as unschedulable, thus preventing new pods from
being scheduled onto that node. The uncordon command works by setting the node’s
unschedulable flag to false, allowing the Kubernetes scheduler to place new pods on
the node. This is useful when maintenance has been completed on a node, and you want
to allow new pods to start running on the node again.
The Ansible Playbook shown in Listing 6-32 sets a node in the drain state but
aborts the operation if there are pods not managed by a ReplicationController, Job, or
DaemonSet. It uses a grace period of ten minutes.

229
Chapter 6 Ansible for K8s Management

Listing 6-32. The drain.yml File

---
- name: k8s drain
  hosts: all
  vars:
    mynode: "k8s.ansiblepilot.com"
    grace_period: 600
  tasks:
    - name: drain node
      kubernetes.core.k8s_drain:
        state: drain
        name: "{{ mynode }}"
        delete_options:
          terminate_grace_period: "{{ grace_period }}"

The Ansible Playbook in Listing 6-33 marks the specified node as unschedulable as
part of the Cordon status.

Listing 6-33. The cordon.yml File

---
- name: k8s cordon
  hosts: all
  vars:
    mynode: "k8s.ansiblepilot.com"
  tasks:
    - name: cordon node
      kubernetes.core.k8s_drain:
        state: cordon
        name: "{{ mynode }}"

In the same way, you can perform the uncordon operation to mark the selected node
as schedulable in the Kubernetes cluster. The Ansible Playbook is shown in Listing 6-34.

230
Chapter 6 Ansible for K8s Management

Listing 6-34. The uncordon.yml File

---
- name: k8s uncordon
  hosts: all
  vars:
    mynode: "k8s.ansiblepilot.com"
  tasks:
    - name: uncordon node
      kubernetes.core.k8s_drain:
        state: uncordon
        name: "{{ mynode }}"

The localhost inventory shown in Listing 6-35 executes the code in the Ansible
Controller.

Listing 6-35. Inventory

localhost ansible_connection=local

You can execute the code using the ansible-playbook command and specifying the
inventory file and the playbook filename (drain.yml, cordon.yml, uncordon.yml):

$ ansible-playbook -i inventory drain.yml

A successful execution using the ansible-playbook command output includes the

following:
• Play recap status: ok=2 changed=1

• Task status:

TASK [drain node]

changed: [localhost]

Many IT professionals chain the three Ansible Playbooks when they upgrade a new
version of Kubernetes in their clusters or expand with a list of hostnames and loops
between them.

231
Chapter 6 Ansible for K8s Management

Kubernetes Dynamic Inventory

Ansible Dynamic Inventory is a feature of Ansible that allows you to automatically generate
and update inventory information from external sources such as cloud providers, CMDBs,
and inventory management systems. This allows users to deploy infrastructure and
applications dynamically and automate them quickly. The inventory is stored in a YAML
file, which is then used by Ansible to determine which hosts to target for tasks. Ansible
Dynamic Inventory can be used for several different use cases, such as dynamic scaling,
provisioning, and application deployment. It can be used to manage hybrid data center
environments and ensure that the right resources are allocated for each environment. See
the “Ansible Dynamic Inventory” section in Chapter 5 for a practical example.

Roll Back Deployments and DaemonSets

Deployments can fail to deploy for a number of reasons. Your Plan B should be to
execute a rollback of the previous working version of the application or service. You can
execute this operation, thus triggering the rollback operation to the deployments and
DaemonSets. The Ansible Playbook in Listing 6-36 starts a rollback on the nginx-server
deployment in the ansible-examples namespace.

Listing 6-36. The rollback.yml File

---
- name: k8s rollback
  hosts: all
  vars:
    myproject: "ansible-examples"
    kind: "Deployment"
    name: "nginx-server"
  tasks:
    - name: rollback
      kubernetes.core.k8s_rollback:
        api_version: apps/v1
        kind: "{{ kind }}"
        name: "{{ name }}"
        namespace: "{{ myproject }}"

232
Chapter 6 Ansible for K8s Management

The localhost inventory shown in Listing 6-37 executes the code in the Ansible
Controller.

Listing 6-37. Inventory

localhost ansible_connection=local

You can execute the code using the ansible-playbook command and specifying the
inventory file and the playbook filename:

$ ansible-playbook -i inventory rollback.yml

A successful execution using the ansible-playbook command output includes the

following:

• Play recap status: ok=2 changed=1

• Task status:

TASK [rollback]
changed: [localhost]

et a New Size for a Deployment, ReplicaSet,

S
Replication Controller, or Job
You can automate the setting of the number of replicas for a Deployment, ReplicaSet, or
Replication Controller, or the parallelism attribute of a Job using the kubernetes.core.
k8s_scale Ansible module. See the “Scale Your App” section in Chapter 4.

Security
Security is a broad concept, especially when many components are interconnected,
like in a Kubernetes cluster. The Linux kernel has a number of overlapping security
extensions (capabilities, SELinux, AppArmor, and seccomp-bpf) that can be configured
to provide the least privilege to programs.

233
Chapter 6 Ansible for K8s Management

The reference for Kubernetes in this field are the four Cs of cloud-native security:

• Cloud

• Cluster

• Container

• Code

To begin with, the cloud infrastructure that hosts your cluster needs to be adequately
secured and enable access to ports on the cluster only by trusted networks. By default,
Container Engines are open from anywhere. A network firewall could be helpful in
mitigating this behavior. This prevents a potential attacker from running port scans to
detect potential open ports in your systems and connect to them. This is the first “C”
in cloud-native and it impacts the security of the entire infrastructure, even when the
cluster is in a private or public cloud data center.
The next is cluster security, which involves the ability of a potential attacker to
execute commands in your cluster. A potential attacker could have the ability to access
running Container Engines such as Docker Daemons, exposed publicly, as well as the
Kubernetes Dashboard without the proper authentication or authorization mechanism.
This can be prevented by using network policies and security in ingress.
The next “C” is containers. A potential attacker can run any container of choice
without restriction, even in privileged mode, no matter what repository it came from or
what tag is used.
Moreover, the attacker can install any application when it can run a malicious
container in your cluster without restriction. A few years ago, the Dirty Cow
CVE-2016-5195 exploited a bug in the Linux kernel’s memory subsystem copy-on-
write (COW) to gain write access from the container to the host system, increasing the
privileges.
You can prevent this behavior by enabling only running containers from a secure
internal registry, disallowing running in privileged mode, and sandboxing isolating
each other.
The best way to prevent supply chain attacks and minimize microservice
vulnerability is acting in the code. Avoid hard-coding database credentials, thereby
passing critical pieces of information through environment variables or exposing
applications without TLS/SSL. These are bad coding practices.
The DevSecOps movement applies security as soon as possible in the code design
phase of your development.
234
Chapter 6 Ansible for K8s Management

Another important tool is to scan for vulnerabilities as soon you create your
container image, when you push to the container image. You must apply this
methodology as soon as possible during the development process.
You can also reduce the attack surface of container images using the smallest images
possible, removing all unnecessary binaries. Reducing the attack surface should be
your mantra.
A good security practice is to build your images using scratch. Add a USER directive to
run as a non-root user inside a container. You can sign your images in order to run only
trusted containers. Most of the time, you derive your containers using the FROM directive.
The most secure way is to combine the scratch with a multi-stage build. In this way,
you obtain a shirked image with only your application. This approach is demonstrated in
Listing 6-38.

Listing 6-38. The Containerfile File

FROM fedora:latest as build

COPY hello.go /app
WORKDIR /app/
RUN go build -o hello

FROM scratch
COPY –from=build /app/hello /app/
ENTRYPOINT ["/app/hello"]

The resulting container has only the executable /app/hello. Hence, it’s super
compact and very secure because it doesn’t have unnecessary binaries and doesn’t
require all the unused libraries of the operating system.

AAA
Every time you make a request, Kubernetes determines if it is allowed or denied using
Authentication, Authorization, and Accounting (AAA). This is how your Kubernetes
cluster verifies who you are and what you are authorized to execute. Policies define
permissions according to role-based access control (RBAC) or attribute-based access

235
Chapter 6 Ansible for K8s Management

control (ABAC). Associated policies define Kubernetes resources. Kubernetes RBAC is

the most commonly used method to perform AAA tasks. Policies are in place to make
decisions to allow or deny an action. The following acronyms are used in the AAA context:

• authn: Authentication (who are you?)

• authz: Authorization (what are you authorized to do?)

• auth: Accounting (what area are you authorized to access?)

You can use Kubernetes RBAC to provide authorization to resources in your

Kubernetes cluster.

OpenID Identity Provider

An identity provider stores and authenticates the identities of users, ensuring the
validity of their login credentials across multiple platforms. In the past, Kerberos was
the preferred authentication system, but it’s sometimes complicated, so nowadays,
especially for API authentication, other solutions are preferred.
OpenID Connector (OIDC) is an authentication layer on top of OAuth 2.0. It is a
secure mechanism allowing an application to contact an identity service, retrieve user-
specific data, and return some information to the application in a secure way. You can
see this with Facebook and Google login buttons in many mobile apps and websites.
They use the authentication information stored in your account, so you don’t have to
retype information into every service. GitHub, Microsoft, and Azure ID are other popular
identity providers.
Kubernetes hosts a per-cluster, public OpenID Connect (OIDC) endpoint that
contains signing keys for JSON web tokens allowing external systems, like Amazon
IAM, to validate and accept the Kubernetes-issued OIDC tokens. OIDC federation
access allows you to assume RBAC roles via the Secure Token Service (STS), enabling
authentication with an OIDC provider and receiving a JSON Web Token (JWT), which is
used to assume a role. OpenID Connect allows single sign-on (SSO) for logging into your
Kubernetes cluster. You can use an existing public OpenID Connect Identity Provider
(such as Google, GitHub, Microsoft, Amazon, and so on), or you can run your own
identity providers, such as dex, Keycloak, CloudFoundry UAA, or Tremolo Security’s
OpenUnison.

236
Chapter 6 Ansible for K8s Management

Calico
You can secure your network communications using Calico around your pod. Calico
acts like a firewall in your pod. It enables you to use network policies within a pod and
apply the zero-trust security model. This means that it considers all the traffic dangerous
by default, and you need to manually set an allow list. The easiest way to install Calico
is via the Tigera operator. The Tigera operator provides lifecycle management for Calico
exposed via the Kubernetes API, defined as a custom resource definition.

Key Takeaways
Automating the management of your Kubernetes clusters using Ansible technology
enables a faster maintenance window and fewer human errors. You can apply the
automation to small and large clusters and many use cases. Managing software with
the Helm package manager simplifies the adoption of the standard software lifecycle
maintenance process. Ansible reduces the cluster administration toil, automating every
operation of cluster maintenance and software lifecycle management. Creating another
cluster is becoming easy and very affordable nowadays, especially for Cloud providers
(Amazon, Google, Azure, and so on). You explore Kubernetes and its cloud computing
opportunities in the next chapter.

237
CHAPTER 7

Ansible for Kubernetes
Cloud Providers
Cloud providers are a great resource for propelling your business to a new level. The
need for more resources is just a few clicks away. Cloud computing is a commercial
offer from a vendor that offers a pool of on-demand computing, storage, database, and
network shared resources sold as a service that you can rapidly deploy at scale. In recent
times also, Machine Learning and Artificial Intelligence resources are becoming popular.
Worldwide analysts see growth in the adoption of cloud provider services year
after year. Forrester report predicts1 that in 2023 cloud-native adoption will rise to 50
percent of worldwide enterprise organizations. The previous adoption rate was 42
percent in 2021 and 33 percent in 2020. Gartner forecasts2 end-user spending on public
cloud services of $591.8 Billion in 2023. The reason behind this is that maintaining
on-premise infrastructure might drain a lot of resources from your IT department.
An affordable Total Cost of Ownership (TCO) and the possibility of obtaining a fully
managed service have attracted organizations in recent years to move their applications
to cloud providers. From the application point of view, it changes nothing, as they
perceive the underlying layer as platform-independent because they rely on container
technology. You have the ability to run your pods in the Kubernetes cluster on-premise,
via cloud computing, or via a hybrid infrastructure without any changes. Ansible
Dynamic Inventories are very useful in conjunction with the cloud provider because the
environment can be fast-paced. As the number of machines and systems under your
control increases, you need a powerful tool to comply with regulations. Ansible Dynamic
Inventories interact with the cloud provider API and retrieve identifiers and IP addresses

1
https://www.forrester.com/report/predictions-2023-cloud-computing/RES178164
2
https://www.gartner.com/en/newsroom/press-releases/2022-10-31-gartner-forecasts-
worldwide-public-cloud-end-user-spending-to-reach-nearly-600-billion-in-2023

239
© Luca Berton 2023
L. Berton, Ansible for Kubernetes by Example, https://doi.org/10.1007/978-1-4842-9285-3_7
Chapter 7 Ansible for Kubernetes Cloud Providers

of the machines in the IT real estate. Ansible Dynamic Inventories are heavily used to
query cloud providers or Kubernetes cluster infrastructures (see the “Ansible Dynamic
Inventory” section in Chapter 5).
One of the main benefits of adopting cloud computing in your organization is that
you are outsourcing the management of the Kubernetes control plane because the
vendor manages it. Adopting the shared responsibility model displayed in Table 7-1,
this complies with the most up-to-date international security and process management
standards. Usually, you need to set up the network configuration for your applications
or services. After the initial configuration, you can spin up global-scale cloud-native
applications in minutes.

Table 7-1. Cloud Computing Shared Responsibility Model

Responsibility On-premise IaaS PaaS SaaS

Application configuration Customer Customer Customer Customer

Identity and access control Customer Customer Shared Shared
Application data storage Customer Customer Shared Vendor
Application Customer Customer Customer Vendor
Operating system Customer Customer Vendor Vendor
Network flow control Customer Shared Vendor Vendor
Host infrastructure Customer Vendor Vendor Vendor
Physical security Customer Vendor Vendor Vendor

Like everything in life, you can’t apply a “one size fits all” approach. Cloud
computing is a great tool that you can apply to your IT infrastructure when the on-
premises resources are not enough, your stakeholders are located worldwide, or
you need to handle a temporary spike of traffic. In this context, a hybrid deployment
makes much more sense. On the other hand, complexity arises because you need
to interconnect your local on-premise infrastructure with infrastructure, machines,
applications, and cloud-based resources. This context is where the Ansible automation
tool shines!

240
Chapter 7 Ansible for Kubernetes Cloud Providers

Cloud Architecture
Cloud providers organize their data center architecture in regions and availability zones.
Figure 7-1 shows the Amazon Web Services (AWS) global infrastructure map, currently
divided into 99 availability zones within 31 geographic regions worldwide at the moment
of writing this book. A region directly impacts the performance and services of the
final users. When your users are in only one part of the world, reducing the latency for
the final user deploying compute and storing the data as closely as possible is more
convenient. A region is a set of data centers connected to a low-latency network. Each
region has more than a data center. An availability zone is a data center independent
in the region. You need to distribute between more availability zones to implement a
high availability and redundant application. Implementing this application architecture
makes the service more reliable and tolerant of some resources or availability zone
failures.

Figure 7-1. The Amazon Web Services (AWS) global infrastructure map

The availability zone is a totally independent data center within the region, with
its electric system, cooling system, and network devices. Each region has at least two
availability zones. Multiple availability zones protect from data center failures. Services
for redundant zone storage automatically replicate your data and services across zones.
Note that some features might not be available in all regions. Review your cloud
provider documentation when you need a specific feature. You need the least amount of
latency as possible, so the data will be sent as fast as possible to your final users. Usually,
the price of services varies from region to region.

241
Chapter 7 Ansible for Kubernetes Cloud Providers

Most providers provide a paired region in the same country. For example, us-west-1
and us-west-2. This guarantees that if your primary choice region fails because of
an outage, the second region will fail over the workload. If the outage affects multiple
regions, the vendor will prioritize at least one region per area for repair. Planned
maintenance is always scheduled in one region at a time. Service replication is often
provided in the paired region as well.
There are three architectural approaches in the cloud providers:

• Infrastructure as a Service (IaaS)

• Platform as a Service (PaaS)

• Software as a Service (SaaS)

Virtual machines are the building blocks inside cloud providers. They have
different names, but the concept is the same and sometimes hidden behind totally
managed services by the vendors. The virtual machine was the first service offered by
cloud providers, sharing resources of common hardware. They are often referred to
by Infrastructure as a Service (IaaS). You take care of managing everything except the
hardware and the networking for the virtual machines. Nowadays, all the cloud providers
offer a marketplace of virtual machine images with different operating systems on them.
Many organizations calculate the Total Cost of Ownership (TCO) of managing the
hardware and plan some migration when it is convenient or efficient. Cloud providers
provide portals and tools to spin up, monitor, and manage a large number of virtual
machines and even hybrid clouds. They also provide blueprints to comply with company
guidelines and offer recommendations to optimize your environment for security,
high availability, and great performance. Pricing might be complicated because it is
calculated hourly, but is also flexible enough to spin up the testing machine for a limited
amount of time. The maintenance of virtual machines (operating system updates,
patches, security, storage, and so on) might impact your IT department and you might
decide to move to a completely managed service offered by the cloud providers. These
types of services are called Platform as a Services (PaaS) because you delegate the
management of the resources to the vendor and take care only of the application to
deliver to your audience. This model is also sometimes called a fully managed platform.
Containers are an important part of modern cloud computing. The major benefits
of containers are their self-contained dependency, which increases portability between
operating systems and hardware platforms and increases efficiency. Containers reduce
the overhead related to maintenance.

242
Chapter 7 Ansible for Kubernetes Cloud Providers

Most of the vendors offer a PaaS service to spin up a limited amount of containers via
the Portal, API, and CLI via client SDK. When the number of containers increases, you
need Kubernetes.
Kubernetes is the common ground for deploying containers in the cloud because
it is widely deployed and adopted by all modern vendors. Containers standardized
the experience between different suppliers, enabled hybrid cloud IT data centers, and
enabled a consistent user experience between vendors. When the number of containers
increases, you need a system to keep track of the moving parts, as well as to make
sure that containers are configured correctly, the file system is consistent, and works
together. As you learned in the previous chapters, Kubernetes takes care of the automatic
application deployment and automatic scaling to meet the load and the customer traffic
demand. It’s battle-tested by Google and many worldwide organizations.
The third architectural design is called Software as a Service (SaaS). Some vendors
also refer to this category as “serverless” or “functions.” It’s the smallest computing
service and can be triggered via a standard web address (URL). It executes once and then
stops. It’s a serverless computing service because you don’t need to worry about virtual
machine toil, maintenance, or processes to manage; you just focus on the functionality.
The most famous serverless services are Amazon Lambda and Azure Function. This
architectural design enables great growth and great flexibility and saves money because
it reduces wasted resources when nobody is using them.
Storing secrets is another important task in a modern IT infrastructure. The most
convenient way is to use external services such as AWS Key Management Service (AWS
KMS), Google Cloud Key Management Service (Cloud KMS), Azure Key Vault, or Vault
by Hashicorp. These types of services help store and retrieve Kubernetes secrets. Keys
are stored in a customer-owned key vault customer-managed key (CMK) or a hardware
security module (HSM).
The Envelope Encryption technique (or Digital Envelope) allows you to store keys
inside an external service and use them using a double encryption mechanism. Using
this method, only the provider can read the envelope, and only the customer can read
the content of the message. You need to encrypt the encryption key with the key stored
in the Key Management Service. Key rotation is also good advice in order to maintain a
high-security level.
You can grant temporary access to keys; this feature is helpful for audit account
use cases.

243
Chapter 7 Ansible for Kubernetes Cloud Providers

Amazon Web Services (AWS)

Amazon Web Services is probably the biggest cloud provider on the market at the
moment, according to international analysts. Amazon Web Services spans 31 regions
with 99 availability zones and is available for users of 245 countries and territories.
Amazon Elastic Compute Cloud (EC2) provides Infrastructure as a Service (IaaS),
many options for Platform as a Service (PaaS), and serverless options with the Amazon
Lambda service.
The network is managed by Virtual Private Clouds (VPC) that you can configure
as public and private segments, and traffic between is regulated by gateways. There
are multiple storage services (detailed in Table 7-2). The most famous is Elastic Block
Storage (EBS), which offers persistent block-level storage and the flexibility to attach
from one instance of EC2 to another. Amazon Simple Storage Service (S3) instead is
an object storage service designed to provide 99.999999999 percent durability, 99.99
percent availability, and incredible scalability and security.

Table 7-2. Amazon Web Services IaaS, PaaS, SaaS Services

IaaS PaaS SaaS

Amazon Elastic Compute Cloud (AWS EC2) Amazon Elastic Kubernetes Service (EKS) AWS Fargate
Amazon Elastic Container Service (ECS)

You can try the Amazon Web Services with a one-year trial, which includes some
resources for the most popular products, like 750 hours per month of Linux, Red Hat
Enterprise Linux (RHEL), or SUSE Linux Enterprise Server (SLES), t2.micro or t3.micro
EC2, and RDS instances (managed Relational Database Service for MySQL, PostgreSQL,
MariaDB, or SQL Server).
You can generate an Ansible Dynamic Inventory of the resources in your Amazon
Web Services infrastructure. The aws_ec2 inventory plugin is included in the amazon.aws
collection. You can also interact with Amazon Web Services and create inventory groups
automatically with your machines inside based on Amazon EC2 instance metadata (see
the section in Chapter 5 titled “Ansible Dynamic Inventory”). First of all, you need to install
the amazon.aws collection in your Ansible Controller. If it’s not already installed, you can
proceed using the ansible-galaxy command included in all the Ansible installations:

$ ansible-galaxy collection install amazon.aws

244
Chapter 7 Ansible for Kubernetes Cloud Providers

You can verify the successful installation of the amazon.aws collection using the same
command with the list option:

$ ansible-galaxy collection list amazon.aws

The amazon.aws collection requires the botocore and boto3 Python libraries to be
installed:

$ pip3 install botocore boto3

The minimal parameter of the aws_ec2 inventory plugin is the “regions” that
specify which regions to query in AWS. The most powerful feature is the ability to create
dynamic groups. This feature reads the host variables according to the keyed_groups
option. The keyed_groups option also defines the prefix and a key format. The full
inventory configuration file of the plugin is shown in Listing 7-1.

Listing 7-1. The inventory.aws_ec2.yml File

plugin: amazon.aws.aws_ec2
regions:
  - us-east-1
filters:
  tag:env: ansible-examples
keyed_groups:
  - key: 'architecture'
    prefix: arch
  - key: instance_type
    prefix: instance_type
  - key: tags.type
    prefix: tag_type
hostnames:
  - ip-address

You need to enable the aws_ec2 Ansible inventory plugin in your Ansible
configuration file, as shown in Listing 7-2.

Listing 7-2. The ansible.cfg File

[inventory]
enable_plugins = amazon.aws.aws_ec2
245
Chapter 7 Ansible for Kubernetes Cloud Providers

It is important to verify that the following environment variables are defined

containing the AWS_ACCESS_KEY_ID and AWS_SECRET_ACCESS_KEY:

export AWS_ACCESS_KEY_ID='AK123'
export AWS_SECRET_ACCESS_KEY='abc123'

You can interact with the Ansible Dynamic Inventory plugin using the ansible-
inventory or ansible-playbook command line:

$ ansible-inventory -i inventory.aws_ec2.yml --graph

This ansible-inventory command generates a graph-style output with all the

resources in the cloud provider.

• Tag: ansible-examples

• graph view

You can execute the code using the ansible-playbook command combined with the
Ansible Dynamic Inventory for the ansible-examples tag:

$ ansible-playbook -i inventory.aws_ec2.yml ping.yml

Amazon Elastic Kubernetes Service (Amazon EKS) offers many options to run
a Kubernetes cluster based on the amount of control that you want to have and the
Kubernetes experience of your teams:

• AWS fully managed nodes: AWS Fargate is fully managed by Amazon

Web Service if you want only to focus on the data plane and quickly
spin up the pods, services, and so on.

• Managed nodes: You can manually add under Amazon Elastic

Compute Cloud (Amazon EC2) instances in Amazon VPC if you need
some special customization nodes.
• Self-managed nodes: For maximum customization, you can also
add customer-managed nodes using Amazon Elastic Compute Cloud
(Amazon EC2) instances in Amazon VPC.
246
Chapter 7 Ansible for Kubernetes Cloud Providers

Amazon provides a shared responsibility model, providing tools and infrastructure

for creating the infrastructure. Choosing the fully managed service (AWS Fargate), AWS
takes direct accountability for any failure of the Kubernetes data plane. In another type
of service, the responsibility might be wary only if there is a failure on the underlying
cloud infrastructure instances. You can focus on developing an application, data, and
support.
AWS Fargate is powered by microVM (Firecracker) technology, a secure and fast
microVM for serverless computing. Amazon EKS-optimized Linux AMI is the default
AMI for Amazon EKS workloads. The AWS Build Specifications are available on GitHub
for creating custom Linux AMI. For example, you could use Bottlerocket, a popular
open-source lightweight Linux-based OS designed to run containers.
AWS Identity and Access Management (Amazon IAM) manages Authentication
Authorization Accounting (AAA) permissions. RBAC roles allow mapping using
ConfigMap to map roles within Kubernetes and IAM, specifying the mapRoles attribute.
You can map the IAM role to a Kubernetes RBAC for authentication.
For example, you can create an aws-auth ConfigMap currently configured to map
the EksNodeRole IAM role to the system:bootstrappers and system:nodes groups. The
EksNodeRole IAM role is applied to the control plane that serves the Kubernetes cluster.
The username in RBAC for one of the nodes would be similar to system:node:ip-node1.
example.com. If a node does not have the EksNodeRole IAM role attached to it, it does not
have the correct permissions in RBAC to be included as a cluster node.
Amazon EKS requires an Amazon VPC network in order to operate. The Kubernetes
cluster node hosts (in an auto-scaling group) are deployed in the private subnet.
Traffic is granularly enabled to the public subnet via a NAT gateway that interconnects
the private subnet with the public subnet. The public subnet inside the AWS cloud is
connected to the Internet via the Internet gateway. Some use cases require a Bastion
host (in an auto-scaling group) configured in the public subnet. You can control inbound
(ingress) and outbound (egress) traffic to and from nodes in a security group and/or
network security group.
Amazon also created an Amazon EKS API that you can use to interact with the
control plane of your Kubernetes cluster. Table 7-3 shows a comparison between the
Kubernetes API and the Amazon EKS API.

247
Chapter 7 Ansible for Kubernetes Cloud Providers

Table 7-3. Kubernetes API vs Amazon EKS API

Kubernetes API • kubectl command-line utility
• Kubernetes objects: Pods, deployment, namespace
• Labelling
Amazon EKS API • eksctl command-line utility
• Cluster management
• Managed node groups
• Fargate
• Tagging

eksctl is an additional command-line tool created by Amazon to interact with

Amazon EKS API. It requires Amazon command-line tools (AWS CLI), kubectl
command-line tools, and IAM permission for Kubernetes (eksClusterRole). Check out
the AWS Prescriptive Guidance for more details about AWS security best practices.
The following eksctl command deploys a small Amazon EKS cluster of two nodes in
the eu-west-1 region as a control plane:

$ eksctl create cluster --name test-cluster --nodegroup-name test-nodes

--node-type t3.small --nodes 2 --nodes-min 1 --nodes-max 5 --managed
--version 1.21 --region eu-west-1 -–zones eu-west-1a,eu-west-1b --role-arn
arn:aws:iam:1234567890:role/eksClusterRole

Each parameter of the eksctl command uses the following:

• name: The name of the Kubernetes cluster, test-cluster, in

this example

• nodegroup-name: The Amazon EKS node group name,

test-cluster, in this example

• node-type: The size of the EC2 instances for the nodes

• nodes: The number of nodes, 2 in this example

• nodes-min: The minimum number of nodes in the Amazon EC2

auto-scaling configuration node group, 1 in this example

• nodes-max: The minimum number of nodes in the Amazon EC2

auto-scaling configuration node group, 5 in this example

248
Chapter 7 Ansible for Kubernetes Cloud Providers

• managed: Create an Amazon EKS managed node group; it is an AWS

managed node group in this example

• version: The version of Kubernetes of your Amazon EKS cluster, 1.21

in this example

• region: The AWS region of the deployment of Amazon EKS cluster

and node group, eu-west-1 in this example

• zones: (optional): Limit the AWS availability zones of the deployment

of Amazon EKS cluster and node group, eu-west-1a and eu-west-1b
in this example

• role-arn: (optional) The Amazon IAM role ARN identifier using the
eksClusterRole or eksNodeRole

You can verify that the successful cluster was generated using the kubectl
command:

$ kubectl get nodes

For each Kubernetes node, you obtain the release version:

NAME                                          STATUS   ROLES    AGE
  VERSION
ip-172-16-11-31.eu-west-1.compute.internal    Ready    <none>   2m
  v1.24.1-eks-fb459a0
ip-172-16-53-234.eu-west-1.compute.internal   Ready    <none>   2m
  v1.24.1-eks-fb459a0
ip-172-16-82-159.eu-west-1.compute.internal   Ready    <none>   2m
  v1.24.1-eks-fb459a0

As expected, there are three instances: (ip-172-16-11-31.eu-west-1.compute.

internal, ip-172-16-53-234.eu-west-1.compute.internal, and ip-172-16-82-159.
eu-west-1.compute.internal) in a Ready status from two minutes on the version 1.24,
specifically Kubernetes version v1.24.1-eks-fb459a0.
When you request a large number of resources and nodes in Amazon, your
cluster might fail to build because the current Amazon availability zone (AZ) doesn’t
have sufficient resources to serve your request. You will receive a ROLLBACK_FAILED
error message. You can perform a more detailed root cause investigation from the
CloudFormation Console under the Events tab.
249
Chapter 7 Ansible for Kubernetes Cloud Providers

Note Amazon Web Services Cloud9 (AWS Cloud 9) is a full online IDE that makes
it easy to create AWS cloud resources to create, execute, and troubleshoot code
using only your favorite browser and CloudFormation stacks.

Another interesting service is Amazon Elastic Container Registry (Amazon ECR),

which provides a fully managed registry for Docker containers and Open Container
Initiative (OCI) artifacts. Major advantages are 99,99 percent SLA, high availability
and scalability, encryption at rest (requires an AWS Key Management Service [KMS]
key), vulnerability scanners, and IAM integration. You can publish publicly or privately
visible images and manage permissions using the IAM repository policy. Do not confuse
Amazon Elastic Container Registry (ECR) with Amazon Elastic Container Service
(ECS). Amazon Elastic Container Service is usually more affordable than the Amazon
Elastic Kubernetes Service (EKS). Moreover, it can run images using the three launch
type models:

• AWS Fargate

• Amazon EC2

• Amazon ECS on AWS Outposts

With the AWS Fargate launch type model, you pay for the resources that your
containerized application utilizes. The amount of vCPU and memory is provided in
sizes—from the tiniest of a minimum of 0.25 vCPU and memory of a minimum of
512 MB to a maximum of 2 GB to the biggest of 16 vCPU with memory from 32 GB and a
maximum of 120 GB. This is in 8 GB increments.
Using the standard UNIX toolchain, you can create and publish files to the private
SCM repository (git), generate a Docker image (docker build), and push (docker
push) to the Container Registry, in this case, Amazon ECR.
AWS CodeCommit is a managed source control repository compatible with Git. It
enables secure coding team collaboration and integrates all the popular features of its
competitors (GitHub, GitLab, and so on).
Amazon provides the service CodePipeline to release software via Continuous
Integration via an intuitive workflow-style step-by-step user interface. Through this
service, you can create a CI/CD pipeline, a set of automated processes and tools that
allow your organization’s developers and operations teams to work cohesively to build
and deploy code to a production environment.

250
Chapter 7 Ansible for Kubernetes Cloud Providers

You can define a pipeline in AWS CodePipeline that stores code in the AWS
CodeCommit, triggering an image build that populates the AWS Elastic Container
Registry. For example, the popular delivery pipeline in AWS uses AWS CodePipeline,
Wave, and Flux.
AWS CloudWatch is the Amazon way to use meaningful metrics data in your cluster.
AWS CloudWatch Container Insights is specialized for container data collection. It
collects metrics and logs, aggregates at the cluster, node, pod, task, and service level,
and visualizes metrics in a dashboard (CPU, memory, disk, network, container data).
The EKS Control Plane can be configured to send logs to CloudWatch and the Data
Plane Fluentd service. The Fluentbit is more lightweight than the Fluentd service,
so most of the time it is preferred. Note that you need to enable the AIM policy
CloudWatchAgentServerPolicy in order to be able to push the log from your EKS cluster
to the CloudWatch.
Each API metric is collected in the dashboard inside a log group. This is a way to
acquire data inside your cluster. Moreover, you can enable AWS X-Ray for tracing the
service from the backend to the frontend in order to debug and analyze any latency in
your system. The traditional debug process is difficult to troubleshoot and involves a
combination of different components.
AWS X-Ray analyzes the traffic requests and organizes them into segments and
subsegments to pinpoint bottlenecks and perform the service graph. It then drills down
on the response code of your API, database, and so on, to understand what exactly
is happening. Usually, you keep track of utilization, latency, duration acquiring raw
metrics, performance metrics, and error metrics and combine them during every
investigation and resolution.
It might be useful to consult the Amazon Well-Architected Framework in order to
review the set of architectural best practices for delivering operational excellence for
better ROI. This will allow you to secure your data and networks, deliver reliability, and
recover after an incidents.

Google Cloud Platform (GCP)

The most successful Google services, such as Google Maps, Google Photos, YouTube,
and Gmail, run on containers. Some statistics estimate a real estate of 4 billion
containers in their infrastructure.

251
Chapter 7 Ansible for Kubernetes Cloud Providers

Google Cloud spans 35 regions with 106 availability zones and is available for users
of more than 200 countries and territories, as shown in Figure 7-2. You can try Google
Cloud Platform using the $300 trial, which is enough to start a laboratory and try the
basic tools.

Figure 7-2. The Google Cloud Platform (GCP) global infrastructure map

As shown in Table 7-4, Google offers the Infrastructure as a Service (IaaS) called

Google Compute Engine (GCE), which allows you as a customer to access powerful
virtual machines. GCP Cloud Storage might be handy for organizing the entire database,
raw video streams, and even a matrix for machine learning models. GCP Networking
is grouped into one easy-to-use dashboard. The Google Cloud Identity and Access
Management (IAM) manages access control by defining who has what access to which
cloud resources.
Google Containers Registry (GCR) is the container registry in the GCP. You can
interact with it by specifying gcr.io in your Kubernetes manifest files as a source of
containers.

Table 7-4. Google Cloud Platform IaaS, PaaS, SaaS Services

IaaS PaaS SaaS

Google Compute Engine (GCE) Google Kubernetes Engine (GKE) Google Cloud Run

252
Chapter 7 Ansible for Kubernetes Cloud Providers

You can interact with GCP via the Google Cloud Shell directly or via the Google
Cloud Console. The Shell appears at the bottom of your Console window when you
choose the Console button on the dashboard. When creating your account, the first
step is to enable the API of the services that you want to use—in this case, the Google
Container API. This API enables the GCP service. In the dashboard, simply search for
API & Services and search for the relevant API to enable.
Google Kubernetes Engine (GKE) is the Google-managed offer of Kubernetes for
managing your deployments and scaling your containerized applications using the
Google infrastructure. It integrates the latest release of Kubernetes plus all the specific
patches for the Google infrastructure. The service is fully compatible with the Kubernetes
API, four-way autoscaling, official release channels, multi-cluster support, and scales up
to 15,000 nodes. The major advantage of using Google Kubernetes Engine (GKE)
is that it closely follows the latest changes in the Kubernetes open-source project.
Google Kubernetes clusters are formed by a virtual machine running on the
Google Compute Engine. At the moment of writing this book, there are two differences
between AKS and GKE services. GKE supports only a few operating systems, and
the SLA is different in the zonal clusters. The first weakness is that there are only two
available operating systems for the server nodes: Container-Optimized OS and Ubuntu.
Container-Optimized OS is an operating system image for the Google Compute Engine
virtual machines optimized for running Docker containers. It is maintained by Google
and based on the open-source Chromium OS project.
Note the Service-Level Agreement (SLA). Amazon Elastic Kubernetes Service (EKS)
offers an SLA of 99.95 percent, whereas GKE offers 99.95 percent only for its regional
clusters. The GKE for its zonal clusters has a 99.5 percent SLA. GKE has a lot of flexibility
based on your workload and budget requirements. The smallest is the single control
plane in a single zone “zonal cluster,” but you can increase to multi-zonal and regional
options.
One of the benefits of running in GCP is that it is fully integrated with Cloud
Controller Manager to manage accesses on the cluster. As a managed service, it
outsources the data plane that is managed by Google,. You can then focus on deploying
your application or service. A cluster is the foundation of the GKE service. A cluster
consists of at least one control pane and multiple nodes. The cluster uses some
connected GCP services such as VPC networking, persistent disk, load balances, and
cloud operations. Each node is managed by the control plane and runs your Pods (aka
user pods). A typical workload completely in Google Cloud could use services like Cloud

253
Chapter 7 Ansible for Kubernetes Cloud Providers

Code as IDE, Code Repository ad SCM, and Cloud Build to create container images
to store in Google Container Registry before being deployed in GKE. You can interact
with your cluster using Cloud Console UI, the gcloud command-line interface, or
Google APIs. One of the main benefits of GKE is the native integration between cloud
monitoring and cloud logging Google services. These services are performed via cloud
operations for GKE, are enabled by default, and provide a monitor dashboard specifically
tailored for Kubernetes. You can also collect application logs for easy troubleshooting
and further analysis.
Another option for running your cloud-native application is using the Google Cloud
Run service. It is the Google serverless offer used by companies like MailChimp, Airbus,
and MediaMarktSaturn. It supports the most popular computer languages (GO, Python,
Java, Node.js, .NET, and Ruby) on a fully managed platform to create containerized and
scalable applications.
You can retrieve the Google Cloud Compute Engine list of resources using the gcp_
compute Ansible inventory plugin included in the google.cloud collection.
You can install the google.cloud Ansible collection using the ansible-galaxy
command included in all Ansible installations:

$ ansible-galaxy collection install google.cloud

The following command verifies a successful installation of the google.cloud

collection:

$ ansible-galaxy collection list google.cloud

The google.cloud collection requires the requests and google-auth Python

libraries installed in your Ansible Controller:

$ pip3 install requests google-auth

The configuration file in Listing 7-3 returns the list of resources in an Ansible
Dynamic Inventory format.

Listing 7-3. The ansible.cfg File

[inventory]
enable_plugins = google.cloud.gcp_compute

Meanwhile, Listing 7-4’s inventory.gcp.yml lists the resources in the ansible-

examples project.

254
Chapter 7 Ansible for Kubernetes Cloud Providers

Listing 7-4. The inventory.gcp.yml File

plugin: gcp_compute
projects:
- ansible-examples
auth_kind: serviceaccount
service_account_file: /home/ansible/credentials.json

The credentials.json file is the service account credential file that contains the
access token downloaded via the Google Cloud Console.
Google, like with the other competitors, has plenty of services to simplify your
workload in cloud computing.

Microsoft Azure Cloud Services

Microsoft Azure provides similar building blocks as Amazon Web Services and Google
Cloud Platform, plus some specific services that might be useful for IT professionals.
Azure spans more than 60 regions with more than 90 availability zones and is available to
users of more than 200 countries and territories. Azure offers some free services the first
year, and a selection of more than 55 services is always free. The free trial at the moment
of writing this book includes 750 hours of B1s burstable virtual machines, 100 GB storage
in one standard tier registry, and 10 webhooks in the Azure Container Registry.
Azure Compute is their Infrastructure as a Service (IaaS). There are many options
for Platform as a Service (PaaS) and serverless with the Azure Functions service. See
Table 7-5.

Table 7-5. Google Cloud Platform IaaS, PaaS, SaaS Services

IaaS PaaS SaaS

Azure Compute Azure Kubernetes Service (AKS) Azure Containers

Azure Compute deploys Windows and Linux virtual machines (VMs) using the
powerful Microsoft Hyper-V hypervisor technology. Azure Hybrid Benefit is the licensing
model that enables you to use software assurance-enabled Windows Server and SQL
Server licenses and Red Hat and SUSE Linux subscriptions on virtual machines in Azure.
Azure Spot Virtual Machines is an option to use heavily discounted, unused computing

255
Chapter 7 Ansible for Kubernetes Cloud Providers

capacity to run your workloads. Azure uses Azure Virtual Network (VNet) to connect
virtual machines to public and private segments. Routing and access control lists can
also be configured dynamically for enhanced security control. When you’re embarking
on your journey to the cloud, you might find Azure Migrate useful. It helps migrate Linux
and Windows, including Virtual Desktop Infrastructure (VDI) resources. It enables a
comprehensive security and compliance overview of your hybrid infrastructure. Azure
Compute Virtual Machines also supports the latest Ampere Altra ARM–based processors
(a complete multi-core server processors system on chip solution built for cloud-native
workloads) since 2022.
The first specific Azure service is a resource group. Resources groups are essential
to every architecture on Azure. They are logical buckets where you include all the
resources needed by your application or project, and you can manage them as a group.
Each resource exists in only one resource group, can be moved to another one, and
can be from different regions. You can apply a common access control policy to all the
resource groups.
Azure Resource Manager (ARM) is the deploying manager on Azure when it comes
to creating, updating, and deleting resources. You can interact with Portal, PowerShell,
Azure CLI, REST APIs, and client SDKs using the Azure Resource Manager API. The best
practice is to deploy, monitor, and manage your resources through the Azure Resource
Manager as a group rather than manually start one after another, individually. The major
advantages are as follows:

–– Consistency: Each deployment triggers the same workflow on each

deployment

–– Dependency: Ensures that all the necessary resources are in place at

the right time

–– Access control: Applies a common policy

–– Tagging: Easily identifies and applies a label

–– Billing: Simplifies audit and understanding

Virtual machines in an Azure cloud provider are called Azure compute and are
widely used all over the platform.
Azure calls a group of identical, load-balanced virtual machines a scale set. The
baseline is a single image that is deployed by the scale set. If one virtual machine in the
scale set stops or fails, the others continue working and take over the load. The scale set

256
Chapter 7 Ansible for Kubernetes Cloud Providers

enables automatic matching of the workload and adding or removing virtual machines
to try to match the needs. The service is free, but you pay for the virtual machines. It can
manage thousands of resources. Azure Service Bus is a message bus interface that you
can use as a source event with Event Driven Ansible (see Chapter 8).
As the previous cloud providers, you can retrieve the list of resources using an
Ansible Inventory plugin. First of all, you need the azure.azcollection Ansible
collection in your Ansible Controller. The installation can be performed using the
ansible-galaxy command included in all Ansible installations:

$ ansible-galaxy collection install azure.azcollection

You can obtain the full list of the installed Ansible collections using the same
command with the list option:

$ ansible-galaxy collection list azure.azcollection

The azure.azcollection collection requires the Python plugins listed in the

requirements-azure.txt file, which you can retrieve from the GitHub repository of the
project3 installed:

$ pip3 install -r requirements-azure.txt

The azure_rm Ansible inventory plugin of the azure.azcollection collection

retrieves the resources included in the ansible-examples Azure Resource Manager, as
demonstrated in Listing 7-5.

Listing 7-5. The inventory.azure.yml File

plugin: azure_rm
include_vm_resource_groups:
- ansible-examples
auth_source: auto

You need to enable the dynamic inventory in your Ansible project with the ansible.cfg
Ansible configuration file, as shown in Listing 7-6.

3
https://github.com/ansible-collections/azure

257
Chapter 7 Ansible for Kubernetes Cloud Providers

Listing 7-6. The ansible.cfg File

[inventory]
enable_plugins = azure.azcollection.azure_rm inventory

As with other Ansible Dynamic Inventory plugins, you can interact with it using the
ansible-inventory or ansible-playbook command line:

$ ansible-inventory -i inventory.azure.yml --list

The output of the ansible-inventory command is a JSON file with all the resources
of the cloud provider:

• Namespace: ansible-examples

• list view

You can execute the code using the ansible-playbook command combined with the
Ansible Dynamic Inventory for the ansible-examples resource group:

$ ansible-playbook -i inventory.azure.yml ping.yml

On the PaaS offering, the easiest way to deploy containers is via the Web App
for Containers. Web App is a managed platform of Azure. It can host your existing
container images. The container supports web applications written in popular computer
programming languages (Python, Node.js, .NET, Java, PHP, and Ruby). There are also
API apps that host and simplify the data backend services.
Azure Container Instances (ACI) is the primary service of Azure for managing and
deploying your applications in a container. The service takes care of managing your
workload on containers without the complexity of virtual machines. On batch workload,
for example, you can spin up the container on demand, process the data, and turn off
the containerized application. Using this workflow, you reduce costs and resources. ACI
supports Azure Portal, CLI, and PowerShell.
When the workload becomes bigger, you can use the Azure Kubernetes Service (AKS)
to orchestrate it. The setup is quick as the service already includes the standard Azure
services, identity and access management, elastic provisioning, integration with
Microsoft coding tools, and global reach using Azure Stack.
Azure Container Registry (ACR) is the Google-managed Container Registry that
manages container image files and artifacts.

258
Chapter 7 Ansible for Kubernetes Cloud Providers

The major advance of using Azure Kubernetes Service (AKS) is its rich integration
with other Azure services, for example, Azure Active Directory (Azure AD).
Azure Kubernetes Service (AKS) is an Azure-managed Kubernetes service with
hardened security and fast delivery. Microsoft handles the control plane, so you
can focus on deploying your applications. AKS also supports NVIDIA GPU-enabled
node pools.
You can interact with the AKS cluster using Azure CLI, Azure PowerShell, Azure
Portal, and template-driven deployment (Terraform, Azure Resource Manager
templates, and Bicep).

Other Vendors
With many vendors offering Kubernetes as part of their offer of cloud services, you
need a way to compare the main features among them. The Certified Kubernetes
Conformance Program, by CNCF, enables multiple Kubernetes cluster interoperability.
At the moment, 90 certified Kubernetes offerings passed the Certified Kubernetes
Software Conformance by CNCF (Sonobuoy). Software conformance certifies that every
vendor’s version of Kubernetes supports at least the community APIs of the open-source
edition.
The benefits of the Certified Kubernetes Conformance Program are as follows:

• Consistency: Works with any installation of Kubernetes

• Timely updates: Updates at least once per year to the latest version

• Confirmability: All end users can run the certification conformance

application (Sonobuoy)

According to IDC, Forrester, and Gartner international analysis, the current

Kubernetes in the cloud computing marketplace is divided between Akamai, Alibaba,
Amazon Web Services, DigitalOcean, Google, Huawei, IBM, Microsoft, Oracle,
OVHcloud, Tencent, Vultr, Zadara, and Linode.

• Rancher, acquired by SUSE in July 2020, is an enterprise-grade and

user-friendly Kubernetes management platform with more than
37,000 active users. SUSE is the company behind the SUSE Enterprise
Linux operating system.

259
Chapter 7 Ansible for Kubernetes Cloud Providers

• Alibaba Cloud Container Service for Kubernetes (ACK) was one of

the first certified platforms and integrates virtualization, security,
storage, and networking capabilities.

• DigitalOcean Kubernetes (DOKS) from DigitalOcean Cloud is a

managed control plane like in PaaS services such as Amazon Elastic
Kubernetes Service (EKS). It includes the DigitalOcean CSI plugin
and the Container Storage Interface (CSI) Driver for DigitalOcean
Block Storage.

• IBM Cloud Kubernetes was one of the first fully-managed Kubernetes

in the cloud, since May 2017 (formerly known as IBM Cloud
Container Service). It runs on the IBM Cloud.

• Linode Kubernetes Engine (LKE) is a service known for its simple

pricing model and ease of use based on the Linode infrastructure.

• Oracle Container Engine for Kubernetes (OKE) offers the same

service in the Oracle Cloud Infrastructure (OCI).

With such a multitude of offerings, it can be difficult to make the right decision.
Competition, in this case, is a good driver for the final user because companies are
competing with each other to offer better services, at a better price, a better SLA, and
a better infrastructure. Before making your decision, consider the alternatives. The
Certified Kubernetes Conformance Program guarantees that you can always move your
application between providers, thus avoiding vendor lock-in mechanisms.

Key Takeaways
Ansible, Kubernetes, and cloud providers are an incredible combination of powerful
tools, technologies, and cost-efficient ways to deploy applications or services at a world
scale. Organizations all over the world are considering hybrid clouds and learning
how to deploy container images faster and scale as needed. You learned how to easily
maintain a remote data center in a cloud environment with Ansible and the right tools
You already know to manage your local laboratory or on-premise data center. Ansible
Dynamic Inventory is the critical tool for homogenizing, scaling, and maintaining your
fleet of machines when the scale is simply too much. In the next chapter, you learn about
Ansible for Enterprise.

260
CHAPTER 8

Ansible for Enterprise
One of the main benefits of Ansible is that it standardizes infrastructure automation
among different technologies and programming styles (Perl, bash, Puppet, Chef, and
so on). It is the Swiss Army Knife tool of modern operations engineers and can easily be
used in Infrastructure as Code (IaC), Configuration as a Code (CaC), Policy as a Code,
Code pipelines, orchestration (K8s), and event-driven automation.
Ansible runs literally everywhere. It enables organizations to speed up processes
and become more efficient and productive. It is the rise of platform engineering tools:
user-driven, self-service infrastructure and deployments that extend the principle of
continuous integration and delivery, furthering infrastructure and operations (I&O)
agility, faster, more efficient, safe, and compliant with international standards.
When you are part of an enterprise organization, you use Ansible in a slightly
different way. Many organizations use Ansible because the amount of resources
to manage is simply too big to handle manually. You can also use tools that make
collaboration within your team more effective and dynamic. You can manage resources
and relationships at a planetary scale nowadays. Also, organizations embrace the change
of internal processes for fast-pacing decision-making. One of the main advantages of
Ansible automation is that it produces code that you can store in a common source
control repository and can be shared between your teams. The Red Hat Ansible
Automation Platform grew up around the idea of creating a web service API and
web user interface that teams can use every day. Some colleagues advantage of the
uncomplicated web interface to trigger the automation, whereas all the operations can
be performed via the RESTful API.

The Ansible Automation Platform

The Ansible Automation Platform (AAP) is the commercial offering from Red Hat, the
company leading the Ansible project. The Ansible Automation Platform subscription
includes the latest software and official support, documentation, and additional

261
© Luca Berton 2023
L. Berton, Ansible for Kubernetes by Example, https://doi.org/10.1007/978-1-4842-9285-3_8
Chapter 8 Ansible for Enterprise

resources from Red Hat. The AAP is a platform of Ansible-based IT automation products,
including Ansible Controller (formerly Ansible Tower), Ansible Core (formerly Ansible
Engine), Automation Hub, and other enterprise features, including:

• Ansible Certified Collections

• Single sign-on

• Role-based access control

• Multi-cloud support

• Red Hat support and subscription

It allows organizations to automate IT operations, deploy and scale applications,

and manage Infrastructure as Code (IaC) across multiple environments, such as on-
premises, cloud, and hybrid.
Ansible Automation Platform includes Ansible Automation Controller (formerly
Ansible Tower). This web-based management console provides a graphical user
interface, role-based access control, and other features to make it easier to manage and
scale Ansible automation across an organization.
It includes a RESTful API and enterprise Ansible collection to interact with it. It
also includes Ansible Core (formerly Ansible Engine), which is the core of the platform.
The release of the Ansible language provides the ability to automate provisioning,
configuration management, application deployment, and many other IT needs. It
includes enterprise-grade support in the subscription services. The platform includes
more products that interact with the Ansible Automation Controller. The most important
is the Ansible Automation Hub. In the same way that the Ansible Galaxy registry
downloads community content, the corporate world relies on the Ansible Automation
Hub to provide Ansible resources and content that they download from it. The Ansible
Automation Hub manages resources from Red Hat product teams, ISV partners, and
community and private contributors.
The partner list in Ansible Automation Hub is getting lengthier day by day and
includes, at the moment, the following: Amazon, Arista, Aruba Networks, Check Point,
Cisco, Citrix, CyberArk, Dell EMC, Dynatrace, F5 Networks, Fortinet, HPE, IBM, Juniper
Networks, Kubernetes, Microsoft, NetApp, NVIDIA, 1Password, SAP, ServiceNow,
Splunk, TrendMicro, and VMWare.

262
Chapter 8 Ansible for Enterprise

The Ansible Automation Platform provides centralized management, visibility,

and scalability to the IT organization and improved collaboration among teams and
departments.
The Ansible Automation Platform has a bi-yearly software release cycle, usually
in May and November. The release 2.x+ fully embraces the Container software design
paradigm with the full support of Kubernetes as an execution environment and
operating system.
The latest release at the time of writing this book is version 2.3. Its main features are
as follows:

• Automation Execution Environments

• Automation Mesh

• Automation Services Catalog

Currently, the Red Hat Ansible Automation Platform is supported on Red Hat
Enterprise Linux versions 8 and 9 and Red Hat OpenShift. The Ansible Automation
Platform Operator provides cloud-native, easy-to-install deployment of Ansible
Automation Platform instances in your OpenShift environment. The Ansible Automation
Platform Operator deploys and manage instances of the Automation controller and
Private Automation hub in your OpenShift cluster. At the moment of writing this book,
the Ansible Automation Platform Operator for installing Ansible Automation Platform
2.3 is available on OpenShift Container Platform 4.9+. It deploys Ansible Automation
Platform instances with a Kubernetes native operator. The two types of instances are
Ansible Controller and Private Automation Hub.
The easiest way to install the Ansible Automation Platform Operator is via the
OperatorHub in the web console. Just search for the Ansible Automation Platform
provided by Red Hat. The latest release is 2.3.0+0.1674778407, as shown in Figure 8-1.
The Operator takes care of all the necessary dependencies, such as PostgreSQL 13.

263
Chapter 8 Ansible for Enterprise

Figure 8-1. The Ansible Automation Platform Operator

You can specify some installation parameters, as shown in Figure 8-2.

Figure 8-2. AAP operator installation parameters

264
Chapter 8 Ansible for Enterprise

Here is a full list of Ansible Automation Platform Operator parameters:

• Update Channel: The exact version to install

• Installation mode: Install the operator cluster-wise or only in a

specific namespace

• Installed namespace: Customize the default “aap” namespace

• Update approval: Manual or automatic

Once it’s successfully installed, you receive a confirmation like the one shown in
Figure 8-3.

Figure 8-3. AAP operator was installed

You can see the Operator Dashboard in Figure 8-4, which enables you to:

• Automation Controller (AC): Deploy a new instance of

AutomationController.

• Automation Controller Backup (ACB): Back up deployment of the

controller, including jobs, inventories, and credentials.

265
Chapter 8 Ansible for Enterprise

• Automation Controller Restore (ACR): Restore a previous controller

deployment from an AutomationControllerBackup. The deployment
name you provide will be the name of the new AutomationController
CR that will be created.

• Automation Controller job template (JT): Define a new job template

in the controller.

• Automation Controller job (AJ): Launch a new job via Controller.

• Automation Hub (AH): Deploy a new instance of Automation Hub.

• Automation Hub Backup (AHB): Back up deployment of the hub,

including all hosted Ansible content, secrets, and the database. By
default, a persistent volume claim will be created using the default
StorageClass on your cluster to store the backup.

• Automation Hub Restore (AHR): Restore a previous hub deployment

into the namespace.

Figure 8-4. The AAP Operator Dashboard

266
Chapter 8 Ansible for Enterprise

Figure 8-5. AAP Operator running instances

When all your instances are fully running, you can access the Ansible Automation
Platform Dashboard (see Figure 8-6), and your automation journey begins. Inside
the interactive web dashboard, you can create users applying RBAC access criteria,
fetch Ansible projects from the most common SCM, and run your automation using
the Job Templates. Every execution of an Ansible Playbook is called a job execution.
A job execution can incorporate Credentials and Vaults and can be executed against
Inventories or Hosts. One of the biggest benefits is that the dashboard centralizes your
experience. You can store credentials, Ansible vaults, keys, and security tokens between
the IT department team, customizing accesses using RBAC and integrating AAP with the
most common enterprise-grade authentication systems.

267
Chapter 8 Ansible for Enterprise

Figure 8-6. AAP Controller Job Dashboard

Ansible Controller executes your automation via Container Groups and Instance
Groups. The Ansible Controller allows you to execute jobs directly on an instance
member, in a virtual environment, in a Kubernetes namespace, or a in a project in an
OpenShift cluster.
When implementing your Infrastructure as Code (IaC), creating an IaC Ansible
pipeline can be handy. It is a Git-centric continuous deployment pipeline using Ansible.
DevOps teams create the Ansible resources and store them in a Git repository. Those
commits are picked up by a CI/CD tool, typically Jenkins, or trigger an Ansible Controller
WebHook to launch associated jobs (playbooks) or workflow.
All the book’s code can be used in the Ansible Automation Platform for easy
interaction between your IT department team.

Event-Driven Ansible
Modern businesses use many different IT systems and services, with the consequence
of increased complexity in managing integration and mode API interfaces. This often
results in a “spaghetti infrastructure.” Accessing a centralized event streaming service
that acts as a message bus with event-driven architecture can achieve easier integration
between different systems and services. The result is a more flexible system that can
adapt to changes in the business environment.

268
Chapter 8 Ansible for Enterprise

An exciting evolution of automation is Event-Driven Ansible (EDA). It’s highly

scalable with a flexible automation capability. At the moment of writing this book,
the EDA was released in October 2022 as a technological preview by Red Hat. Event-
Driven Ansible contains a decision framework that is built using Drools. In a nutshell, it
implements the ability to react to a system event and trigger automation. In this way, you
can work smarter by applying Event-Drive Ansible and Ansible Rulebooks.
Think about the ability to apply “If-This-Than-That” logic or to minimize the Mean-
Time-To-Resolution (MTTR) for outages in your systems. This is the latest evolution of
smart automation, an understanding of how to execute automated scripts to mitigate
and quickly remediate any event in your IT infrastructure. The key component is the
Event-Driven Ansible Server, which runs in the system as a service and listens for events.
The Ansible Rulebook (see Figure 8-7), similar to the Ansible Playbook, explains to the
EDA Server what the playbook executes, specifying the relevant parameters.

source

event rule conditio action

nal

Figure 8-7. Workflow of an EDA rulebook

The building blocks of Event-Driven Ansible automation are as follows:

• Event: Any changes from the source

• Rule: Evaluated for every event

• Conditional: The event expression that triggers the action

• Action: What to execute (playbook, modules, tasks)

In the rulebook, you can define the sources of events: hooks, alert managers,
URLs, and files to listen to. For each of them, you can define and perform some action
according to the rule that you write. Each rule defines the condition that triggers the
action. Each rule has some conditionals. Conditionals evaluate some event status
logically and define when to trigger an action. For example, it could be an event from the
alert manager or when a specific HTTP status code is obtained, a newly created file, or
a message from the webhook. When a condition is reached, one or more actions can be

269
Chapter 8 Ansible for Enterprise

triggered. The most common action is run_playbook. Each source of event determines
the type of event in the rulebook. A very interesting source of events is Apakce Kafa as a
“message bus” or webhook.
At the moment of writing this book, the following source plugins are available:

• alertmanager: Receives webhook events from alertmanager

• Azure_Service_Bus: Receives Azure Service Bus events

• kafka: Receives Apache Kafka topic messages

• url_check: Polls URLs and sends events based on their statuses

• watchdog: Sends events when a file status changes

• webhook: Receives events from a webhook

• file: Loads facts from YAML files and reloads them when a
file changes

• tick: Generates events with an increasing index

• range: Generates events with an increasing range

The installation of Ansible Event-Driven requires two components:

• The Ansible ansible.eda collection

• The Python ansible-rulebook command-line utility

You can manually install these using the ansible-galaxy

command or using a requirements.yml file:

$ ansible-galaxy collection install ansible.eda

$ pip3 install ansible-rulebook

The utility requires the Java 11+ runtime environment (Oracle, OpenJDK, GraalVM,
and so on) and the JAVA_HOME environment variable, set accordingly. You can verify
a successful installation using the --version parameter of the ansible-rulebook
command to verify the current version of the software:

$ ansible-rulebook --version

The rulebook in Listing 8-1 exposes a webhook on port 8000 that is listening for
messages. Each time you receive a non-empty message (a condition), it triggers the
execution of the playbook.yml playbook, shown in Listing 8-2.
270
Chapter 8 Ansible for Enterprise

Listing 8-1. The eda.yml File

---
- name: Listen for webhook events
  hosts: all
  sources:
    - ansible.eda.webhook:
        host: 0.0.0.0
        port: 8000
  rules:
    - name: webhook
      condition: event.payload.message != ""
      action:
        run_playbook:
          name: playbook.yml

Listing 8-2. The playbook.yml File

- name: eda demo

  hosts: localhost
  tasks:
    - name: print a message
      ansible.builtin.debug:
        msg: "Event Driven Ansible"

Listing 8-3. The inventory File

localhost ansible_connection=local

You can execute your rulebook using the following command:

$ ansible-rulebook --rulebook eda.yml -i inventory --verbose

You can add the optional parameter --verbose when you want more debug
information.
You can also create a container image with the Ansible rulebook and deploy it in a
pod to implement a scalable Event-Driven Ansible on Kubernetes.
The latest release of Ansible Automation Platform 2.3 integrated Event-Driven
Ansible as a Developer Preview release.

271
Chapter 8 Ansible for Enterprise

IT Trends
More and more applications are moving from monolith devices to microservices
software design patterns. This is one of the reasons for the increased use of containers
and the adoption of the Kubernetes container orchestration tool. An orchestrator helps
you deploy high-availability applications with no downtime. They scale up or down
when the traffic demands more resources, making applications more flexible. Latency
is becoming the key metric to evaluate the quality of your applications. Kubernetes also
has the ability to self-heal infrastructure to run your application when an event happens,
such as a failure in the underlying server, storage, or network. The lowest metrics for
Recovery Point Objective (RPO) and Recovery Time Objective (RTO) are available for a
disaster recovery or data protection plan.
IT infrastructure is becoming the heart of organizations. International analyst
agencies predict that the complexity and type of threats to rise yearly. The container
and Kubernetes Security Market is in expansion. We often hear on the news about
ransomware or supply chain attacks. The most impactful event on this side was
the SolarWinds hacking in June, 2022. The hackers breached the Onion network in
September 2019, malicious code had been injected into Orion for seven months. FireEye,
the detection and response Cyber Security Experts and Solution Provider, was the first
firm to publicly report the attack, called the SolarWinds hack “UNC2452.” The attack
was performed via a “Sunburst” backdoor, which connects to an external command-
and-control server. The Orion software was turned into a weapon, thus gaining access to
several global IT infrastructures.
According to the reports, the malware affected 18,000 SolarWinds customers and
impacted many public and private companies. Famous organizations such as FireEye,
Microsoft, Intel, Cisco, and Deloitte were impacted by this attack, as were government
departments such as the U.S. Homeland Security, State, Commerce, and Treasury. It also
significantly damaged SolarWinds’ reputation, probably taking them years to heal and
recover.
This event raised public opinion and consensus that companies needed to pay more
attention to privacy issues. Corporations are obliged by law to report data breaches, and
in the upcoming years, we are going to see a rise in data breach compensation claims.
Starting in 2018, with the implementation of the European General Data Protection
Regulation (GDPR), many data protection frameworks were created to protect data as an

272
Chapter 8 Ansible for Enterprise

asset. For example, there is the EU–US Privacy Shield, California’s Consumer Privacy
Act (CCPA), Brazil’s General Personal Data Protection Law/Lei Geral de Proteção de
Dados (LGPD), and South Africa’s Protection of Personal Information Act (PoPIA).

Ansible Trusted Code

To minimize the risk of supply chain attacks, it’s important to use only trusted code in
your organization. In November 2022, Red Hat released the following:

• Ansible Trusted Ansible Collections

• Signed Ansible Projects

Ansible Trusted Ansible Collections is a feature available in the Ansible Automation

Platform, Automation Hub, that enables the download of only digitally certified content
signed via a GPG key. GNU Pretty Good Privacy, or GPG, is a popular program that
authenticates contents with digital signatures of stored files.
In the same way that Ansible Certified Contents are digitally signed, you can use
the same technology for in-house content. The ansible-sign command-line utility is
designed precisely for project signing and verification. You can install the utility via the
popular operating system package manager or the Python pip installer.
As general habits, we store our project in an SCM repository, usually Git. The
list of files for signature is specified in the MANIFEST.in file, where we can specify
which files to include or exclude. This is a famous format used in many open-source
projects nowadays. You can include or exclude single files or recursive-include or
recursive-exclude directories or only match some criteria with global-include and
global-exclude.
Listing 8-4 is an example of a MANIFEST.in file for this project that excludes the .git
directory, includes the inventory file, and includes all the files with the yml extension
under the playbooks directory.

Listing 8-4. The MANIFEST.in File

recursive-exclude .git *
include inventory
recursive-include playbooks *.yml

273
Chapter 8 Ansible for Enterprise

The full command to sign this project is as follows:

$ ansible-sign project gpg-sign PROJECT_ROOT

In the same way, you can verify the project using this command:

$ ansible-sign project gpg-verify PROJECT_ROOT

A signed project has an additional .ansible-sign directory with two files inside.
The first is the sha256sum.txt file, which contains the SHA256 hash of each file. This
is already an industry standard. The second file is sha256sum.txt.sig, which contains
the GPG signature of the previous file. With the contents signature and verification, you
obtain a success when the signature verification is valid. Two files exist, the checksum is
valid and the signature is valid.
Here are the possible scenarios when something is not correct:

• The checksum is invalid when files are tampered with or are invalid

• The signature is invalid when the signature is tampered with

or invalid

• There is a signature mismatch when the signature is wrong, expired,

or inaccessible

The benefits of the feature outweigh the extra overhead in the effort. I’m happy
with the outcome. Now you know how to use the ansible-sign utility for your projects.
Taking a closer look at the utility by design, it already supports the expansion with more
options and more hashing technologies in the future. Sigstore is a new standard for
signing, verifying, and protecting software that might be embraced in the future.
Ansible enterprise customers can use the Signed Ansible Projects features every day
as part of the Ansible Automation Platform 2.3 (November 2022) or the open-source
AWX 21.7.0 with AWX Operator v0.30.0 (October 2022). The Signed Ansible Projects
feature is implemented in AWX and Ansible Automation Controller. You can enable the
feature of creating a credential using the GPG Public Key type and selecting it in your
projects. Content verification will automatically occur upon future project sync. Every
time your project syncs, the signature will be evaluated, and if it doesn’t match, the
execution will be forbidden. Under the hood, it uses the ansible-sign utility to validate
the project.

274
Chapter 8 Ansible for Enterprise

What’s Next?
Ansible and Kubernetes are two open-source products that benefit from a vibrant
community of people worldwide. Somehow, the holistic view is more powerful than
each part of it. Kubernetes has more than 18,000 GitHub contributors, whereas Ansible
has 5,100 contributors.
New trends emerge every day in this fast-paced IT world. Artificial intelligence-
powered automation is already on the rise, and automated machines are expected to
take over a growing portion of professions in the coming years. Sometimes machines
can execute a task better, faster, and cheaper than humans. Other times, machines create
big mistakes. Applying artificial intelligence to infrastructure could be a great boost in
efficiency for the Infrastructure and Operations (I&O) of projects.
Since the OpenAI ChatGPT release announcement on November 30, 2022, the
ChatGPT tool has seen great hype on the Internet. People around the globe tested
in the most variant situations. The application of the GPT-3 (Generative Pre-trained
Transformer) large language model is impressive with a massive amount of data.
Applying this type of technology to your daily workflow could be interesting. At the
moment of writing this book, you can interact with the tool in a chatbox user interface.
The tool can generate some Ansible code that you need to revisit before running it.
Announced at AnsibleFest 2022, Project Wisdom will be released later in 2023 as
the result of a close collaboration between the Ansible Engineering Team and IBM
Research. The goal was to create an Ansible AI (Ansible Lightspeed with IBM Watson
Code Assistant) integrated into the VSCode editor to create code based on the Ansible
task text.
This type of technology scenario creates opportunities for IT professionals to step
up, upskill themselves, and be part of the machine evolution. The unique ability to
connect data, be creative, and imagine the future is a game changer for humans. “We are
the master of our fate: We are the captain of our souls,” as William Ernest Henley said in
Invictus.
Intuition is the unarticulated, almost imperceptible gist that you get in a situation.
That feeling that sometimes it’s quite right, or that sometimes it is about to happen,
that we can’t explain, but know to be true. It seems like magic, but it is humans’ unique
ability that differentiates us from machines. Gut instinct is a valuable tool that we can
use for the present to identify what to improve in our life without really thinking about it.
It’s the human essence.

275
Chapter 8 Ansible for Enterprise

Thank You
Kubernetes is a very powerful but also complex technology for container orchestration.
It solves the problem of managing many containers to deploy microservices of cloud-
native applications and services. You learned how to automate and simplify many
day-to-day operations using Ansible automation technology. These applications and
services are used by people around the world—they interact every day with websites,
bank accounts, and e-commerce, they reserve a table at a restaurant, order a delivery/
pickup, a taxi, or a flight, and many other examples. The microservice application design
has become very pervasive in just a few years at a world scale.
Briefly, Ansible and Kubernetes are evolving open-source products. Check out the
official websites for the latest news and updates.
I’m so grateful to live in the information age, where ideas become businesses
overnight, and as individuals, we have equal opportunities for success.
Using Ansible automation with Kubernetes for container orchestration with cloud
providers opens up unlimited computing resources. The use of artificial intelligence and
deep learning models is growing faster and faster. Altogether, this world enables a new
class of cloud-native applications of tomorrow that is going to change our lives.
The future is uncharted territory, and the sky is the limit. I look forward to hearing
your success automation story.
“Simplicity is the ultimate sophistication.” —Leonardo da Vinci
Please be nice. Life is sweeter that way.

276
Index
A inventory, 33, 37
localhost, 38
AppArmor profiles, 193
open-source vs. commercial
AAP Operator, 263
options, 29
AAP Operator parameters, 265
Playbook, 40–42
Alpine Linux distribution, 15
conditional statements, 45
Amazon availability zone (AZ), 249
facts, 48
Amazon command-line tools (AWS
filters, 44
CLI), 248
handler, 46
Amazon EKS API, 247, 248
Amazon Elastic Container Registry loop, 46, 47
(Amazon ECR), 11, 250 magic variables, 48, 49
Amazon Elastic Container Service plugins, 50, 51
(Amazon ECS), 11, 250 templates, 49, 50
Amazon Elastic Kubernetes Service variables, 43
(Amazon EKS), 11, 246, 250, 260 Vault, 49
Amazon Machine Image (AMI), Playbook/Inventory, 31
11, 84, 247 plugins, 28
Amazon Web Services (AWS), 3, 11, 56, 87, provisioning, 28
99, 109, 153, 201, 244 Role, 51, 52
Ansible, 7, 169 runner process, 60
benefits, 27 Ubuntu Linux systems, 35
builder process, 57, 58 UNIX target node, 31, 32
code language, 37 Windows target node, 32
collection, 54, 55 YAML format, 39, 40
configuration file, 245 Ansible Automation Controller, 262
configuration management, 29 Ansible Automation Hub, 262
control node, 34 Ansible Automation Platform (AAP), 126,
DevOps, 27 261, 263
Execution Environment, 55, 56, 59 ansible-builder command, 59
Galaxy, 53, 54 ansible-builder command-line tool, 60
INI format, 39 ansible-builder tool, 59
installation, 34 ansible.cfg file, 198

277
© Luca Berton 2023
L. Berton, Ansible for Kubernetes by Example, https://doi.org/10.1007/978-1-4842-9285-3
INDEX

Ansible Community package, 35 ansible-playbook command, 69, 82,

Ansible configuraion 231, 246
error, 111 ansible-rulebook command, 270
installation, 111, 112 ansible-runner tool, 61
kubectl, 113 ansible-sign command, 273, 274
kubernetes.core collection, 109, Ansible troubleshooting
111, 112 kubeconfig file, 118, 119
kubernetes.core.k8s module, 111 401 unauthorized
OpenShift cluster, 113 fatal error message, 114
PyYAML jsonpatch Kubernetes, 114–116
kubernetes, 111 Kubernetes authentication, 114
requirements.yml file, 112 OpenShift, 116, 117
resources, 109, 110 tokens, 113
setup.py install method, 111 x509 error, 118
Ansible Controller, 82, 126–128, 130, 134, Ansible Trusted Ansible Collections, 273
136, 160, 197, 205, 244 AnsibleTurboModule class, 112
Ansible Dynamic Inventory, 194, 195, Apache webserver container, 78
197, 232 AppArmor
Ansible Execution Environment, 123 Linux security tool, 192
ansible builder, 123, 125 vs. SELinux, 192
Ansible Playbook, 125 Application deployment, 29
Ansible Runner, 123, 125 Application.java, 173, 174
bindep.txt file, 124 Applications
context/Containerfile file, 125 containerized, 87
execution-environment.yml file, Kubernetes networking
123, 124 ClusterIP service, 144
localhost connection, 125 ExternalName service, 147
my_kube, 123 ingress object, 146, 147
pip package manager, 123 LoadBalancer service, 145
Red Hat, 125 nodeport property, 145
requirements.txt file, 124 NodePortservice service, 144
requirements.yml file, 124 pod status, 143
ansible-galaxy command, 270 YAML files, 148
ansible-inventory command, Kubernetes Service
38, 246, 258 Ansible Playbook, 142
Ansible modules, 198 inventory file, 143
Ansible Playbook, 30, 31, 37, 40–43, 49, 52, kubectl command, 142
185, 197, 198 nginx-example service, 142

278
INDEX

OpenShift cluster, 142 Automation Controller Backup (ACB), 265

pod statuses, 141 Automation Controller job (AJ), 266
service.yaml document, 141 Automation Controller Restore
stakeholders, 141 (ACR), 266
resilience, 148 Automation Hub (AH), 29, 262, 263,
scaling, 149 266, 273
Ansible, 150 Automation Hub Backup (AHB), 266
Ansible Playbook, 150–152 Automation Hub Restore (AHR), 266
auto-scaling, 153, 154 AWS fully managed nodes, 246
changed status, 152 AWS Identity and Access Management
get parameter, 149 (Amazon IAM), 247
inventory file, 151 AWS Key Management Service (AWS
k8s.gcr.io registry, 154 KMS), 243
kubectl command, 149 AWS X-Ray, 251
nginx deployment, 149 Azure Active Directory (Azure AD), 259
pod count, 152 Azure compute, 255, 256
READY column, 149, 150 Azure Container Instances (ACI), 258
scale parameter, 149 Azure Container Registry (ACR), 255, 258
updation Azure Kubernetes Service (AKS), 258, 259
ansible-playbook command, 157 Azure Resource Manager (ARM), 257
automation, 156, 157 advantages, 256
Kubernetes, 156 azure_rm Ansible inventory plugin, 257
latest tag, 155
logs, 156
nginx deployment, 155 B
pod instances, 155 Base Container Images (BCI), 16
rollback option, 155 Bitnami, 204–207
rolling updation, 154, 155 Borg, 88
scaling, 155 Business-critical services, 148
Array variables, 43
Artificial intelligence (AI), 163, 239,
275, 276 C
Attribute-based access control Calico, 237
(ABAC), 235 California’s Consumer Privacy Act
Authentication, Authorization, and (CCPA), 273
Accounting (AAA), 235, 247 Certified Kubernetes Conformance
Automation Controller (AC), 29, 30, 262, Program, 259, 260
263, 265, 266, 274 Charts, 202, 204, 207, 211

279
INDEX

Cloud computing, 61, 88, 148, 237, 239, lifecycle, 5

240, 242, 255, 259 Container Storage Interface (CSI), 90, 260
Cloud Key Management Service (Cloud Content verification, 274
KMS), 243 Continuous Delivery (CD), 94
Cloud-native applications, 87, 171, 199, Continuous Integration and Continuous
201, 202, 240, 254, 276 Delivery (CI/CD), 7
Cloud Native Computing Foundation Continuous Integration (CI), 93, 94
(CNCF), 82, 87 Cordon, 229, 230
Kubernetes, 9 credentials.json file, 255
Kubernetes distribution Customer-managed key (CMK), 243
Amazon EKS, 11
OCP, 9, 10
public cloud, 11 D
Cloud-native environments, 87, 95 DaemonSets, 91, 232, 233
Cloud providers, 87, 96, 99, 109, 239 Debian Security Advisories (DSA), 18
architecture, 241, 242 Deep learning (DL), 163, 276
Cloud storage services, 90 Deployment, 232, 233
CloudWatch, 251 container, 4
Cluster Autoscaler (CA), 153, 154 traditional, 2
CMD instruction, 13, 19 virtualized, 3
CodePipeline, 250, 251 Developers, 93–95
Common Vulnerabilities and Exposures Devmode, 75
database (CVE), 18 Digital Envelope, 243
Conditionals, 269 Disaster Recovery (DR), 148, 272
Configuration as a Code (CaC), 261 Distribution Specification, 5
Configuration class, 171 Docker, 63, 88
Configuration files, 5, 27, 39, 49, 82, 92, Debian Linux, 65, 67
113, 183, 186, 197, 245, 254, 257 Fedora/Red Hat Linux, 67, 68
Configuration management (CM), 7, platforms, 64
27–29, 99, 262 Windows, 70, 71
Containerfile, 5, 12, 123, 176, 235 DockerCon, 88, 89
Container Registry, 5, 6, 13, 22, 77, 178, 179 Dockerfile, 5, 12, 13, 18–19, 59
Containers, 4 Drain, 229, 230
analysis tools, 17
benefits, 6
business applications, 6 E
definition, 11 echo command, 12
deployment, 4 eksctl command, 248

280
INDEX

Elastic Block Storage (EBS), 90, 244 Google Containers Registry (GCR), 252
Elastic Compute Cloud (EC2), 244, 246 Google Kubernetes Engine (GKE), 89, 253
Elastic Container Registry (ECR), 250, 251 Gradle build tool, 172, 175, 176
Elastic Kubernetes Service (EKS), 11, 244, Graphics Processing Units (GPU), 163
246, 250, 253, 260 Guestbook application, 180–184
Enterprise customers, 274
Enterprise Linux distribution, 16, 17, 64
Envelope Encryption technique, 243 H
Event-Driven Ansible (EDA), 268–271 Hardware, 2, 88
Extra variable, 43, 126, 219 Hardware security module (HSM), 243
HashiCorp, 82, 243
Hello App, 12
F Alpine Linux, 15
Fault tolerant (FT), 148 application, 19–21
Fedora-like Linux operating system, 76 browser, 20
Fedora Linux, 14, 82 building, 22
Fetch logs, resources, 220, 221 deploy in Kubernetes, 24
Flatpak, 63, 71–74, 85 deploy in Operator, 25
Fluentbit, 251 Docker, 22, 23
FROM instruction, 13 Dockerfile, 18
Full command, 41, 52, 85, 112, 125, 127, 274 ADD instruction, 19
CMD instruction, 19
EXPOSE instruction, 19
G FROM instruction, 18
gather_facts variable, 48 LABEL instruction, 19
General Data Protection Regulation pip command, 19
(GDPR), 272 RUN instruction, 19
General Personal Data Protection Enterprise Linux distribution, 17
Law, 273 Fedora Linux, 14
Generative Pre-trained Transformer Linux base images, 13
(GPT-3), 275 RHEL, 15, 16
GitOps continuous deployment, 93–94 source files, 18
Google, 87–89 SUSE, 17
Google Cloud Platform (GCP), 3, 11, 87, Ubuntu Linux, 14, 15
99, 109, 251, 252 Hello application, 19–22, 176, 177
Google Compute Engine (GCE), 3, Hello container, 177
252, 253 HelloController class, 174
Google Container Engine (GKE), 11 HelloController.java file, 173

281
INDEX

Hello-deployment.yaml Manifest, 179 I

Hello-service.yaml Manifest, 180
Idempotency, 38
Helm, 10, 109
Identity and Access Management (IAM),
Helm Chart, 207, 211, 219
247, 252, 258
Helm package manager, 111
Image Specification, 5
advantages, 202
Infrastructure and operations (I&O),
charts, 202
261, 275
commands, 202, 203
Infrastructure as a Service (IaaS), 242, 244,
directory, 202
252, 255
glib Python library, 203
Infrastructure as Code (IaC), 1, 25, 27, 61,
helm command-line tool, 202
261, 262, 268
helm utility, 203
Integrated development environment
kubernetes.core Ansible collection, 203
(IDE), 95
Microsoft, 202
Internet of Things (IoT), 74, 107
modules, 203
IT infrastructure, 1, 25, 28, 243, 246,
tasks, 203
269, 272
uses, 202
Helm packages
installation, 207, 209 J
removal process, 209, 210 Java microservice
reporting, 210–212 configure, 171, 172
Helm plugins Demo Java web application, 172
commands, 212 Jenkins Pipeline, 94
installation, 212–214 Job, 233
removal process, 214, 215 Job execution, 267
reporting, 215, 216 Job template (JT), 266, 267
Helm repositories JSON Patch operation, 222, 223
addition, 204, 206 JSON Web Token (JWT), 236
Bitnami, 204
helm repo command, 204
removal process, 206 K
Heroku platform, 170 keyed_groups option, 245
Hetzner-ocp4, 103 Key Management Service (KMS), 243, 250
High availability (HA), 6, 99, 104, 148, 155, Kubeadm, 9, 102, 104, 106–108
184, 241, 242, 250, 272 kubectl, 92
Horizontal Pod Autoscaler (HPA), 153 kubectl command, 249
Hybrid-cloud, 4 Kubernetes
Hypervisor software, 96 manage services, 225, 227

282
INDEX

Kubernetes’ annotation, 193 Linux Containers (LXC), 2, 11

kubernetes.core.kustomize, 185 localhost inventory, 209, 211, 213, 214,
Kubernetes data plane, 171 216, 221, 223, 225, 231, 233
Kubernetes (K8), 8
advantage, 89
Ansible, 201 M
benefits, 89 Machine learning (ML), 163, 239, 252
Container Image, 133 Magic variables, 48, 49, 61, 66
control plane vs. data plane, 91, 92 main() method, 175
definition, 87 Managed nodes, 246
feature, 89 Manual work, 7
Google, 87, 88 Maven or Gradle build tool, 176
K3s, 107 Mean-Time-To-Resolution
objects, 90, 91 (MTTR), 269
self-healing capabilities, 87 Mesos, 88, 89
timeline, 87 Message bus, 257, 258, 270
upgrades, 107, 108 method parameter, 72
uses, 87 Microservices, 4, 6, 7, 87, 171, 172,
YAML files, 166, 167 272, 276
Kubernetes Operations (kOps), 109 Microservices software design
Kubespray, 96, 99–100 patterns, 272
Kustomization directory, 185, 191 microVM, 247
Kustomization file, 184, 185 Minikube, 96, 104
kustomization.yaml file, 111, 186 kubectl command, 105
kustomization.yml file, 185 memory/CPUs settings, 105
Kustomize tool, 184, 185 memory parameter, 104
minikube command-line tool, 104
minikube dashboard command,
L 105, 106
LABEL instruction, 13, 19 node, 105
Laboratory output, 105
creation, 95 packages, 104
hetzner-ocp4, 103 resources, 105
Kubespray, 99, 100 Ubuntu 22.04 LTS, 104
OpenShift Local, 100, 102, 103 virtual machine, 104
Raspberry Pi, 97, 98 Modern operating systems, 7, 8, 34, 96,
virtual machines, 96, 97 104, 202
Lei Geral de Proteção de Dados (LGPD), 273 Modern processor chips, 2

283
INDEX

Monitoring tool deployment, 216 OpenID Connect (OIDC), 236

Ansible Collections, 218, 219 OpenID identity provider, 236
Helm Chart, 219 OpenShift, 10, 89, 101, 113
kube-prometheus project, 217 OpenShift Container Platform (OCP), 9,
Multi-cloud, 4, 9, 262 10, 101, 103, 104, 263
mysql.yaml file, 186, 187 Operator SDK, 25, 191
Oracle Container Cloud Service
(OCCS), 11
N Orchestration technologies, 89
Namespaces Orchestration tools, 7, 272
creation, 126, 128, 129 Organizations, 1, 2, 4, 30, 89, 202
reporting, 129–131 Over-the-air (OTA), 74
reporting deployments
ansible-galaxy, 135
Ansible Playbook, 133, 134 P, Q
applying deployment, 132 Packer, 82, 84, 85
changed statuses, 134 PersistentVolume, 186
deployment_list.yml file, 135, 136 PersistentVolumeClaims, 186
deployment.yaml file, 131 PHP Guestbook application, 180, 181
inventory file, 134, 136 ping.yml file, 197
jmespath library, 135 pip utility, 35, 111
json_query filter plugin, 135 Platform-as-a-service (PaaS), 4, 89, 95,
kubectl namespace, 132 242, 244, 255
Kubernetes deployment, 131 Playbook, 38
message, 132 PodDisruptionBudget (PDB), 148, 229
metadata, 135 Podman, 5, 63
nginx deployment, 132 Podman Container Engine, 80
output, 132 Pods, 12, 253
pods, 132 Ansible Playbook, 137
query lookup plugin, 134 configuration, volumes, 163–165
Network storage protocols, 90 copying files/directories, 224, 225
nginx container, 77 definition, 136
Nginx web server, 92, 155, 207 inventory file, 138
kubectl, 138
Kubernetes file, 137, 138
O OpenShift cluster, 138
Open Container Initiative output, 138
(OCI), 5, 11, 15, 250 resources

284
INDEX

CPU, 158, 160 Redis Service YAML file, 182

GPU, 163 Registered variables, 43, 44
memory, 161, 162 ReplicaSet, 91, 150, 233
metrics, 158 Replication Controller, 150, 229, 233
namespace, 163 Role-based access control (RBAC),
resource utilization, 150 29, 191, 235
service scenario, 148 Runtime Specification, 5
Pod security, 191, 192
Pod Security Admission (PSA), 191–192
PodSecurityPolicy, 191 S
Pod security standard policy, 192 Scratch, 89, 96, 97, 235
Pokémon Go, 89 Seccomp, 193, 194
Prometheus, 216 seccomp profile audit.json file, 194
Prometheus Operator, 217 Secrets, 139, 141
Promise Theory, 88 Secure Token Service (STS), 236
Provisioning, 28 Security, 233
Python AAA, 235
daemon, 112 Calico, 237
dependencies, 111, 113 cloud infrastructure, 234
library, 111, 121–124 cluster security, 234
package manager, 111 code, 234
virtual Environment, 120–122 container images, 235
containers, 234, 235
directives, 235
R OpenID identity provider, 236
Raspberry Pi, 8, 97–98 Security pod resources, 192, 193
Recovery Point Objective (RPO), 272 Self-managed nodes, 246
Recovery Time Objective (RTO), 272 Service-Level Agreement (SLA), 253
Red Hat Ansible Automation Platform, Simple Storage Service (S3), 244
29, 261, 263 Single-board computer (SBCs), 97
Red Hat Enterprise Linux (RHEL), 9, 15, Single sign-on (SSO), 236, 262
16, 35, 64, 68, 123, 244, 263 Single source of truth (SSOT), 94
Red Hat OpenShift, 263 Snap, 74, 75, 77
Red Hat OpenShift Local, 10 Snapcraft, 74
Red Hat OpenShift Service on AWS snapd daemon, 74, 77
(ROSA), 9 Software as a Service (SaaS), 170, 242, 243
Red Hat Security Bulletins (RHSB), 18 Software development lifecycle (SDLC)
Redis leader deployment, 181 methodology, 169, 170

285
INDEX

Source control management system Uncordon, 108, 229–231

(SCM), 94 Universal Base Image (UBI), 15, 16
Spring Boot, 172, 175 UNIX toolchain, 250
state parameter, 72 Uptime, 2
Stateful applications, 163, 186, 191 User-defined variables, 43
Storage devices, 90
SUSE Linux Enterprise Server (SLES), 244
Syscalls, 193, 194 V
System Integration Test (SIT), 6, 201 Vagrant, 82, 83
Vagrantfile, 83, 84
Vendors, 4, 11, 242, 243, 259–260
T Vertical Pod Autoscaler (VPA), 153, 154
Taint nodes, 227, 228 Vibrant community, 9, 87, 89, 107, 275
Technology, 193 Virtual Desktop Infrastructure (VDI), 256
Total Cost of Ownership (TCO), 239, 242 Virtualized deployment, 3
Traditional deployment, 2 Virtual machines (VMs), 2–4, 83, 84,
Traditional infrastructure, 1 96–97, 255
Twelve-factor application, 170, 171 Virtual Network (VNet), 256
Virtual Private Clouds (VPC), 244, 253
VMWare Tanzu Application
U Platform, 95
Standard ubi UBI images, 16
Multi-service ubi-init UBI images, 16
Micro ubi-micro UBI image, 16 W, X, Y, Z
Minimal ubi-minimal UBI images, 16 Web-generated ZIP file archive, 173
Ubuntu Linux distribution, 14, 15 Webhook, 255, 268–270
Ubuntu Security Notices (USN), 18 wordpress.yaml file, 189

286

Learn Ansible Quickly RHCE
100% (3)
Learn Ansible Quickly RHCE
200 pages
Kubernetes Basic To Advance End To End
100% (4)
Kubernetes Basic To Advance End To End
295 pages
Ansible Automation Workshop
100% (1)
Ansible Automation Workshop
127 pages
Hands-On Kubernetes On Azure
100% (4)
Hands-On Kubernetes On Azure
330 pages
(Smtebooks - Com) Managing Kubernetes - Operating Kubernetes Clusters in The Real World 1st Edition PDF
100% (5)
(Smtebooks - Com) Managing Kubernetes - Operating Kubernetes Clusters in The Real World 1st Edition PDF
167 pages
Docker Certified Associate (DCA) Practice Tests - Four Full Length Practice Exams
No ratings yet
Docker Certified Associate (DCA) Practice Tests - Four Full Length Practice Exams
453 pages
Kubernetes User Roadmap
No ratings yet
Kubernetes User Roadmap
20 pages
Udemy Docker Advanced PDF
0% (1)
Udemy Docker Advanced PDF
71 pages
Application Observability with Elastic: Real-time metrics, logs, errors, traces, root cause analysis, and anomaly detection
From Everand
Application Observability with Elastic: Real-time metrics, logs, errors, traces, root cause analysis, and anomaly detection
Navin Sabharwal
No ratings yet
Kubernetes Security - Michael Hausenblas, Liz Rice PDF
100% (2)
Kubernetes Security - Michael Hausenblas, Liz Rice PDF
85 pages
Kubernetes CKA 0100 Core Concepts PDF
No ratings yet
Kubernetes CKA 0100 Core Concepts PDF
77 pages
Implementing Infrastructure As A Code Using Terraform: Case Study
100% (1)
Implementing Infrastructure As A Code Using Terraform: Case Study
4 pages
Kubernetes Tutorial
100% (11)
Kubernetes Tutorial
83 pages
D2 Ansible Playbooks Documentation PDF
100% (1)
D2 Ansible Playbooks Documentation PDF
150 pages
Terraform Practice Guide
100% (14)
Terraform Practice Guide
109 pages
Ansible For Containers and Kubernetes By Examples
From Everand
Ansible For Containers and Kubernetes By Examples
Berton
No ratings yet
Kubernetes: Build and Deploy Modern Applications in a Scalable Infrastructure. The Complete Guide to the Most Modern Scalable Software Infrastructure.: Docker & Kubernetes, #2
From Everand
Kubernetes: Build and Deploy Modern Applications in a Scalable Infrastructure. The Complete Guide to the Most Modern Scalable Software Infrastructure.: Docker & Kubernetes, #2
Jordan Lioy
No ratings yet
Hands-on DevOps with Linux: Build and Deploy DevOps Pipelines Using Linux Commands, Terraform, Docker, Vagrant, and Kubernetes (English Edition)
From Everand
Hands-on DevOps with Linux: Build and Deploy DevOps Pipelines Using Linux Commands, Terraform, Docker, Vagrant, and Kubernetes (English Edition)
Alisson Machado de Menezes
No ratings yet
Ansible For Kubernetes PDF
100% (6)
Ansible For Kubernetes PDF
172 pages
Ansible For Aws Sample
No ratings yet
Ansible For Aws Sample
106 pages
Advanced Ansible Best Practices
No ratings yet
Advanced Ansible Best Practices
66 pages
Serverless On Kubernetes Gorilla Guide
No ratings yet
Serverless On Kubernetes Gorilla Guide
43 pages
YAML-its Usage in Kubernetes
100% (1)
YAML-its Usage in Kubernetes
46 pages
Kubernetes Practicals Ebook
67% (3)
Kubernetes Practicals Ebook
187 pages
ShowIT Openshift - Mini OCP - DevOps 20180131 - Jiri - Kolar
No ratings yet
ShowIT Openshift - Mini OCP - DevOps 20180131 - Jiri - Kolar
59 pages
Terraform From Bigginer To Master
100% (3)
Terraform From Bigginer To Master
90 pages
Kubernetes For Fullstack Developers
100% (4)
Kubernetes For Fullstack Developers
637 pages
Kubernetes II Deep Dive
No ratings yet
Kubernetes II Deep Dive
74 pages
Learn Kubernetes 5 Minutes at A Time
No ratings yet
Learn Kubernetes 5 Minutes at A Time
187 pages
Deploying and Scaling Kubernetes With Rancher - 2nd Ed
No ratings yet
Deploying and Scaling Kubernetes With Rancher - 2nd Ed
66 pages
Getting Started With Kubernetes PDF
100% (1)
Getting Started With Kubernetes PDF
15 pages
Ansible Automation For Sysadmins v2
50% (2)
Ansible Automation For Sysadmins v2
23 pages
Learn+DevOps+ +kubernetes
No ratings yet
Learn+DevOps+ +kubernetes
420 pages
Kubernetes Succinctly
No ratings yet
Kubernetes Succinctly
121 pages
CICD With Docker Kubernetes Semaphore
No ratings yet
CICD With Docker Kubernetes Semaphore
92 pages
Running Airflow Reliably With Kubernetes
100% (1)
Running Airflow Reliably With Kubernetes
47 pages
Ansible Best Practices: How To Write, How To Execute, and How To Use in Real Life
100% (1)
Ansible Best Practices: How To Write, How To Execute, and How To Use in Real Life
67 pages
TESYHYG
100% (1)
TESYHYG
248 pages
Ansible For Aws Sample
No ratings yet
Ansible For Aws Sample
105 pages
Container Networking Docker Kubernetes
100% (8)
Container Networking Docker Kubernetes
72 pages
Terraform Associate
100% (9)
Terraform Associate
465 pages
Sysadmin
No ratings yet
Sysadmin
64 pages
OpenShift Container Platform-4.6-Installing On GCP-en-US
No ratings yet
OpenShift Container Platform-4.6-Installing On GCP-en-US
248 pages
Terraform Certified
100% (3)
Terraform Certified
121 pages
THE Definitive Cheat Sheet: Kubectl
No ratings yet
THE Definitive Cheat Sheet: Kubectl
23 pages
Openshift Tutorial
80% (5)
Openshift Tutorial
110 pages
Istio-Succinctly-Infraestructura Cloud
No ratings yet
Istio-Succinctly-Infraestructura Cloud
150 pages
Red Hat Service Mesh Deep Dive
100% (1)
Red Hat Service Mesh Deep Dive
120 pages
Study Guide PDF
No ratings yet
Study Guide PDF
58 pages
Masteringopenstack PDF
100% (2)
Masteringopenstack PDF
462 pages
Kubernetes Security
100% (1)
Kubernetes Security
156 pages
CIS Kubernetes Benchmark v1.7.0 PDF
No ratings yet
CIS Kubernetes Benchmark v1.7.0 PDF
330 pages
Cka PDF
75% (4)
Cka PDF
58 pages
Cloud Migration Mastery: Complete Guide To Seamless Cloud Integration With AWS, Microsoft Azure, VMware & NaviSite
From Everand
Cloud Migration Mastery: Complete Guide To Seamless Cloud Integration With AWS, Microsoft Azure, VMware & NaviSite
Rob Botwright
No ratings yet
Mastering Kafka Streams: From Basics to Expert Proficiency
From Everand
Mastering Kafka Streams: From Basics to Expert Proficiency
William Smith
No ratings yet
Mastering DevOps in Kubernetes: Maximize your container workload efficiency with DevOps practices in Kubernetes (English Edition)
From Everand
Mastering DevOps in Kubernetes: Maximize your container workload efficiency with DevOps practices in Kubernetes (English Edition)
Soumiyajit Das Chowdhury
No ratings yet
Ultimate VMware NSX for Professionals
From Everand
Ultimate VMware NSX for Professionals
Vinay Aggarwal
No ratings yet
Docker: Build, Test, And Deploy Applications Fast
From Everand
Docker: Build, Test, And Deploy Applications Fast
Rob Botwright
No ratings yet
Ceph Cookbook
From Everand
Ceph Cookbook
Singh Karan
4/5 (1)
Certified Kubernetes Application Developer (CKAD) Exam Success Guide: Ace your career in Kubernetes development with CKAD certification (English Edition)
From Everand
Certified Kubernetes Application Developer (CKAD) Exam Success Guide: Ace your career in Kubernetes development with CKAD certification (English Edition)
Sethumadhavan Kishore
No ratings yet
Office Automation Systems (OAS)
100% (1)
Office Automation Systems (OAS)
12 pages
M 6 Call Setup Optimization Lte Optimization Ra 47040 V 20 B H9L5gkSj
No ratings yet
M 6 Call Setup Optimization Lte Optimization Ra 47040 V 20 B H9L5gkSj
41 pages
Prathamesh Kalantri Resume PDF
No ratings yet
Prathamesh Kalantri Resume PDF
1 page
Artificial General Intelligence
No ratings yet
Artificial General Intelligence
6 pages
Compiled (SAAD)
No ratings yet
Compiled (SAAD)
49 pages
linuxcommands_advanced
No ratings yet
linuxcommands_advanced
13 pages
Candidate Name: Exam Title: Delivery Date: Grade: Candidate ID: Test Center ID
No ratings yet
Candidate Name: Exam Title: Delivery Date: Grade: Candidate ID: Test Center ID
1 page
SoMachine HVAC Quick Start
No ratings yet
SoMachine HVAC Quick Start
109 pages
Chapter: 7.5 Operation Topic: 7.5.1 Creation of Database and Table
No ratings yet
Chapter: 7.5 Operation Topic: 7.5.1 Creation of Database and Table
17 pages
Neq Sim User Manual
No ratings yet
Neq Sim User Manual
58 pages
20 - Custom Collection Development Part 8 - Core Java Tutorial - Mr. Hari Krishna
No ratings yet
20 - Custom Collection Development Part 8 - Core Java Tutorial - Mr. Hari Krishna
5 pages
Pyspark With Docker
100% (1)
Pyspark With Docker
15 pages
Technical Guidelines For Web Application and Website Security
No ratings yet
Technical Guidelines For Web Application and Website Security
37 pages
Cv-Sap MM
No ratings yet
Cv-Sap MM
5 pages
Assignment 2-Java
No ratings yet
Assignment 2-Java
11 pages
How To Use Digitrace
No ratings yet
How To Use Digitrace
4 pages
Introduction to GW
No ratings yet
Introduction to GW
4 pages
Web Scraping C18
No ratings yet
Web Scraping C18
35 pages
Hardware/Software Design of Embedded Systems: Sudeep Pasricha
No ratings yet
Hardware/Software Design of Embedded Systems: Sudeep Pasricha
50 pages
Microservices Vs Monolithic 13th Jan
No ratings yet
Microservices Vs Monolithic 13th Jan
16 pages
The Optimizing Information Leakage in Multicloud Storage Services
No ratings yet
The Optimizing Information Leakage in Multicloud Storage Services
7 pages
VT82C42 Keyboard Controller For PC Board
No ratings yet
VT82C42 Keyboard Controller For PC Board
13 pages
40G To 4x10G AOC Cable Data Sheet by JTOPTICS
No ratings yet
40G To 4x10G AOC Cable Data Sheet by JTOPTICS
5 pages
30 Nov 23
No ratings yet
30 Nov 23
3 pages
CVAbhiman
No ratings yet
CVAbhiman
2 pages
1676-Texto Del Artículo-7983-1-10-20210420
No ratings yet
1676-Texto Del Artículo-7983-1-10-20210420
16 pages
Linked List: Traversal Insertion Deletion
No ratings yet
Linked List: Traversal Insertion Deletion
89 pages
Product Specifications: Dr-Id 1200 (FDR D-Evo Ii)
No ratings yet
Product Specifications: Dr-Id 1200 (FDR D-Evo Ii)
32 pages
AP Changes in R12
No ratings yet
AP Changes in R12
3 pages
Getting Started With Walabot
No ratings yet
Getting Started With Walabot
43 pages