The book provides a general coverage of Kafka's architecture and internal working.
It details many configuration parameters that affect clustering, replication, message delivery.
It also offers valuable material for system administrators who need to manage and monitor a running cluster.
It's not very good for programmers: Java API coverage is partial and inadequate.
The chapters are uncoordinated and poorly integrated with some repeated material.
Many errors in the code samples and the text.
Chapters 3, 4, and 11 discuss the programming API.
The first two deal with message producers and consumers, present the Java API for publishing and consuming messages and discuss delivery semantics. They also detail configuration options that can be used to customize message producers and consumers. Security and access control is mentioned but never really discussed. There is not a single full program that can be run but several snippets full of errors.
Chapter 11 offers a tutorial introduction to stream processing: what it is and what problems it solves. Three code examples illustrate Kafka Streams, the Stream framework that comes with Kafka and provides a high level abstraction for manipulating data streams. The chapter gives you a taste of what you can do with Kafka Streams but doesn't do much to teach how to use it.
Chapter 2 gives a tutorial on Kafka installation and discusses several configuration options that may help in tuning a Kafka cluster. Basic ZooKeeper knowledge can help understand.
Chapter 5 delves into the internals of replication, partitions, request processing, and message storage on physical files.
Chapter 6 discusses data delivery guarantees. It revisits producer and consumer issues related to message delivery, and how to configure brokers and topics. It also explains how "at least once" delivery is easily achievable while "exactly once" delivery is not.
Chapter 7 briefly explores the Kafka Connect architecture: a producer/consumer alternative to exchange data between Kafka and another data storage system.
Chapters 8-10 have a more sysadmin-oriented content.
Chapter 8 explores cross cluster data mirroring, why you may need it, available alternative architectures/models and issues of lost or duplicated data you may come across. It also introduces Kafka's own cluster mirroring tool MirrorMaker, its configuration and tuning.
Chapter 9 covers command line tools to create and manage topics and partitions.
Chapter 10 is on monitoring a Kafka cluster and explores JMX metrics exposed by brokers, producers and consumers that can help in monitoring and detecting problems. Basic JMX knowledge is required to follow along.
Kafka - The Definitive Guide (英语) 平装
- ISBN : 9781491936160
- 语种： : 英语
- ASIN : 1491936169
|5 星 (0%)||0%|
|4 星 (0%)||0%|
|3 星 (0%)||0%|
|2 星 (0%)||0%|
|1 星 (0%)||0%|
美国亚马逊： 17 条评论
Acceptable for administrators, bad for programmers, and tons of errors2017年12月16日 - 已在美国亚马逊上发表
Waste of money and time2018年11月12日 - 已在美国亚马逊上发表
You might better purchase Kafka in action or Kafka stream in action
Comprehensive and accessible2018年10月2日 - 已在美国亚马逊上发表
There is enough information here to both get you started and keep you out of trouble. Anyone seriously looking at using Kafka should read this book.