Technology · Ebook

Kanister: Application-Aware Data Management on Kubernetes

by Shriira Press

Kanister is a CNCF sandbox project, born from the Veeam Kasten team, that solves one of the hardest problems in cloud-native operations: protecting the data of real, stateful applications. A volume snapshot captures bytes, but it cannot guarantee a database quiesced its writes or recorded a consistent point in time. Kanister closes that gap by letting domain experts encode application-specific operations as declarative Blueprints made of ordered Phases, run through a Kubernetes controller using ActionSets and Profiles. This book starts with the problem and the idea of application consistency, then unpacks each building block: Blueprints, Kanister functions, ActionSets, Profiles, the Go-template parameter system, and the kanctl and kando tools, closing with real-world adoption and where it fits the ecosystem.

Contents

1. 1Preface
2. 2Chapter 1 — The Stateful Problem on Kubernetes
3. 3Chapter 2 — Application Consistency and the Kanister Idea
4. 4Chapter 3 — Architecture and the Controller
5. 5Chapter 4 — Blueprints, Actions, and Phases
6. 6Chapter 5 — Kanister Functions: The Primitives
7. 7Chapter 6 — ActionSets and Profiles
8. 8Chapter 7 — Templating, Parameters, and Artifacts
9. 9Chapter 8 — Tooling and Operations
10. 10Chapter 9 — Kanister in Practice