Managing Kubernetes: Operating Kubernetes Clusters in the Real World

by Brendan Burns, Craig Tracey

Overview

While Kubernetes has greatly simplified the task of deploying containerized applications, managing this orchestration framework on a daily basis can still be a complex undertaking. With this practical book, site reliability and DevOps engineers will learn how to build, operate, manage, and upgrade a Kubernetes cluster—whether it resides on cloud infrastructure or on-premises.

Brendan Burns, cofounder of Kubernetes, and Craig Tracey, staff field engineer at Heptio, dissect how Kubernetes works internally and demonstrate ways to maintain, adjust, and improve the cluster to suit your particular use case. You’ll learn how to make architectural choices for designing a cluster, managing access control, monitoring and alerting, and upgrading Kubernetes. Dive in and discover how to take full advantage of this orchestration framework’s capabilities.

• Learn how your cluster operates, how developers use it to deploy applications, and how Kubernetes can facilitate a developer’s job
• Adjust, secure, and tune your cluster by understanding Kubernetes APIs and configuration options
• Detect cluster-level problems early and learn the steps necessary to respond and recover quickly
• Determine how and when to add libraries, tools, and platforms that build on, extend, or otherwise improve a Kubernetes cluster

Product Details

ISBN-13:	9781492033912
Publisher:	O'Reilly Media, Incorporated
Publication date:	12/03/2018
Pages:	185
Sales rank:	523,610
Product dimensions:	6.90(w) x 9.10(h) x 0.50(d)

About the Author

Brendan Burns is a co-founder of the Kubernetes open source container management platform. He is currently a distinguished engineer at Microsoft running the Azure Resource Manager and Azure Container Service teams. Before Microsoft he was a senior staff engineer on the Google Cloud Platform. Prior to working in Cloud he developed web search backends that helped power Google search. Prior to that he was a Professor of Computer Science at Union College in Schenectady, NY. Brendan received a PhD in Computer Science from the Universityof Massachusetts Amherst and a BA from Williams College.

For the last 20 years Craig Tracey has helped build the infrastructure that powers the Internet. In this time he has had the opportunity to develop everything from kernel device drivers, to massive-scale cloud storage services, and even a few distributed compute platforms. Now as a Software Engineer turned Field Engineer at Heptio, he helps organizations accelerate their adoption of Kubernetes by teaching the principles of cloud native architectures through code.

Based in Boston, Massachusetts, in his free time, Craig loves playing hockey and exploring Europe. Craig holds a BS in Computer Science from Providence College.

Table of Contents

Preface ix

1 Introduction 1

How the Cluster Operates 2

Adjust, Secure, and Tune the Cluster 3

Responding When Things Go Wrong 3

Extending the System with New and Custom Functionality 4

Summary 5

2 An Overview of Kubernetes 7

Containers 7

Container Orchestration 9

The Kubernetes API 10

Basic Objects: Pods, ReplicaSets, and Services 10

Organizing Your Cluster with Namespaces, Labels, and Annotations 14

Advanced Concepts: Deployments, Ingress, and StatefulSets 15

Batch Workloads: Job and Scheduledjob 18

Cluster Agents and Utilities: DaemonSets 18

Summary 18

3 Kubernetes Architecture 21

Concepts 21

Declarative Configuration 21

Reconciliation or Controllers 22

Implicit or Dynamic Grouping 23

Structure 25

Unix Philosophy of Many Components 25

API-Driven Interactions 25

Components 26

Head Node Components 26

Components On All Nodes 28

Scheduled Components 29

Summary 30

4 The Kubernetes API Server 31

Basic Characteristics for Manageability 31

Pieces of the API Server 31

API Management 32

API Paths 32

API Discovery 33

OpenAPI Spec Serving 36

API Translation 37

Request Management 38

Types of Requests 38

Life of a Request 39

API Server Internals 46

CRD Control Loop 46

Debugging the API Server 46

Basic Logs 47

Audit Logs 47

Activating Additional Logs 47

Debugging kubectl Requests 48

Summary 48

5 Scheduler 49

An Overview of Scheduling 49

Scheduling Process 50

Predicates 50

Priorities 50

High-Level Algorithm 51

Conflicts 52

Controlling Scheduling with Labels, Affinity, Taints, and Tolerations 53

Node Selectors 53

Node Affinity 54

Taints and Tolerations 56

Summary 57

6 Installing Kubernetes 59

Kubeadm 59

Requirements 60

Kubelet 61

Installing the Control Plane 62

Kubeadm Configuration 63

Preflight Checks 64

Certificates 65

etcd 65

Kubeconfig 67

Taints 68

Installing Worker Nodes 68

Add-Ons 69

Phases 70

High Availability 70

Upgrades 71

Summary 73

7 Authentication and User Management 75

Users 76

Authentication 77

Kubeconfig 85

Service Accounts 87

Summary 89

8 Authorization 91

REST 91

Authorization 92

Role-Based Access Control 93

Role and ClusterRole 94

RoleBinding and ClusterRoleBinding 96

Testing Authorization 98

Summary 99

9 Admission Control 101

Configuration 102

Common Controllers 102

PodSecurityPolicies 102

ResourceQuota 104

LimitRange 106

Dynamic Admission Controllers 107

Validating Admission Controllers 108

Mutating Admission Controllers 110

Summary 113

10 Networking 115

Container Network Interface 115

Choosing a Plug-in 117

Kube-proxy 117

Service Discovery 119

DNS 119

Environment Variables 120

Network Policy 121

Service Mesh 123

Summary 124

11 Monitoring Kubernetes 127

Goals for Monitoring 127

Differences Between Logging and Monitoring 129

Building a Monitoring Stack 129

Getting Data from Your Cluster and Applications 129

Aggregating Metrics and Logs from Multiple Sources 131

Storing Data for Retrieval and Querying 133

Visualizing and Interacting with Your Data 134

What to Monitor? 134

Monitoring Machines 135

Monitoring Kubernetes 136

Monitoring Applications 136

Blackbox Monitoring 137

Streaming Logs 138

Alerting 138

Summary 139

12 Disaster Recovery 141

High Availability 141

State 142

Application Data 142

Persistent Volumes 143

Local Data 143

Worker Nodes 143

etcd 144

Ark 145

Summary 146

13 Extending Kubernetes 149

Kubernetes Extension Points 149

Cluster Daemons 150

Use Cases for Cluster Daemons 150

Installing a Cluster Daemon 151

Operational Considerations for Cluster Daemons 151

Hands-On: Example of Creating a Cluster Daemon 152

Cluster Assistants 152

Use Cases for Cluster Assistants 153

Installing a Cluster Assistant 153

Operational Considerations for Cluster Assistants 154

Hands-On: Example of Cluster Assistants 154

Extending the Life Cycle of the API Server 155

Use Cases for Extending the API Life Cycle 155

Installing API Life Cycle Extensions 156

Operational Considerations for Life Cycle Extensions 156

Hands-On: Example of Life Cycle Extensions 156

Adding Custom APIs to Kubernetes 158

Use Cases for Adding New APIs 158

Custom Resource Definitions and Aggregated API Servers 159

Architecture for Custom Resource Definitions 160

Installing Custom Resource Definitions 160

Operational Considerations for Custom Resources 161

Summary 161

14 Conclusions 163

Index 165