Clusterbau: Hochverfügbarkeit mit pacemaker, OpenAIS, heartbeat und LVS [Gebundene Ausgabe]

Perhaps you have read the review I wrote back in June 2008 about the first edition.

In May 2008 the author published "Clusterbau mit Linux-HA Version 2" ISBN: 3897217791.
This new 2nd edition "Clusterbau: Hochverfügbarkeit mit pacemaker, OpenAIS, heartbeat und LVS" ISBN:3897219190 is basically an enhanced and extended version.

It aditionally includes openais and pacemaker to reflect the development that took place in the meantime. As a reflection of that the book size has increased from 256 to 355 pages.

The first chapter (Introduction) remains basically unchanged.

The second chapter (Basics) is now trying to cover the history and development of heartbeat, the CRM and Openais (corosync). This is unfortunately one of the chapters that remain to be quirky. I had to re-read a couple of paragraphes and take memos on the "evolution of cluster software" chart to be able to follow the text.

The installation chapter (third) did already cover Fedora, OpenSuse, Debian, SLES and RHEL in the 1st edition. Now the author even added instructions for CentOS, Ubuntu and openais (corosync).

What was The fourth chapter in the 1st edition became now the fourth and fifth chapter "Creating resources" and "administration of the cluster" covering 96 pages !!
Needless to say that that is more detail there than ever before. This now also covers the CRM's subshell and a note on Linbits new GUI.

Chapter 6 "Planning, setup and operation" and Chapter 7 "Agents" are basically unchanged.

Chapter 8 "Case scenarios" and Chapter 9 "LVS" now cover 62 pages. This allows coverage for DRBD, Firewall and LVS but also NFS server, Iscsi target and clients and (new) virtual machines.

Chapter 10 "Observation and Security" has doubled in size (16p) but still feels a bit short.

Summary:

Overall I have to say I am impressed how much the book improved within 19 months.
Author and editor team have effectively worked on some of the defects I "complained" about for the 1st edition. The quirky "neudeutsch" or "denglish" formulations are (almost) history.
The front index is also slightly improved.

As a result of that, the book reads more fluently, makes more sense.
In short, it allows the reader to do just more with it.

It becomes very obvious that the author and editor -team have worked hard to turn the book around. However, I think they still have some work left to do.

The 2nd editions focus is still strong on the configuration and parameter descriptions. It now covers the case scenarios to start your cluster setup with more detailed. The XML snippets that are included and explained throughout the book give the reader now a stronger idea of how to configure the cluster and how to manage it.

Part of this is down to improved documentation but also to a new quality of the management interfaces like the GUI's or the CRM subshell.

For a third edition I would like the suggest the below improvements:

- still more case scenarios (double)
- workshop like tutorials, templates
- more system charts and explanations

This 2nd edition has made a huge leap forward since 1st edition.
Its covering more background, better explained than ever before.

I can foresee that should the publisher decide to translate this book into English this must become an instant hit.
There is STILL absolute nothing comparable out there and it doesn't seem that other publishers have even started picking up this topic yet.

Overall I would give it 5 stars for beginners, 4 stars for professionals.

Contents quality: 5 stars
Contents quantity: 4 stars
Compactness: 5 Stars
Pricing: 4 stars

Prinzipiell findet man im Internet ja alles was man benötigt. In der Praxis kämpft man dann halt mit Konfigurationsdateien aus anderen Versionen die aktuell nicht mehr funktionieren, muss stundenlang suchen etc. - da bietet dieses Buch wirklich Hilfe. Es wird grob auf die wichtigsten Programmpakete eingegangen, es gibt ein paar sinnvolle Beispielkonfigurationen. Man kann es wirklich rasch durchlesen, es ist verständlich geschrieben. Einige Teile überfliegt man und kehrt dann später zurück wenn man sie benötigt.
Wenn es wirklich tief werden soll, dann kann man auch gezielt im Internet weiterforschen, das Basis-Verständnis hat man.

Nur absolute Anfänger in der Thematik sollten sich ein paar Grundlagen aus anderen Werken holen, wer unter Linux mit der Kommandozeile nicht umgehen kann oder ipdump, etc. zum ersten Mal hört sollte sich vorbilden. Ansonsten: uneingeschränkt empfehlenswert und einigermaßen aktuell.

In diesem Buch wird das Grundverständnis zum Aufbau von Clustersystemen sehr gut beschrieben. Dabei werden auch wichtige Faktoren berücksichtigt, die unmittelbar mit dem Betrieb von Clustersystemen zusammenhängen, wie z.B. die Vermeidung von Kosten durch Produktionsausfälle oder die Problematik inkonsistenter Datenbestände. Ein sehr übersichtlich gestaltetes Werk, voll vom fundiertem Wissen des Autors, dass sich gleichermaßen an Cluster-Einsteiger und an erfahrene Administratoren richtet. Das Buch kann ich nur empfehlen!

Marcel Sommer