Part I begins with a brief look at the philosophy of backup and an explanation of why backup and recovery systems are so important. An "executive summary" chapter then covers the why, what, when, and how of backups and their storing, testing, and monitoring. The rest of Backup and Recovery has the details behind this, but the focus throughout remains on the underlying ideas.
Looking at open source tools, Part II begins with a chapter on standard tools such as cpio, tar, and dump. Individual specialists then present separate chapters on Amanda, Bacula, and PC Backup. These are clearly written and informative, but a downside is that there is no overview or comparison to help a prospective user choose between them.
A chapter "Open-Source Near-CDP" has a dozen pages on rsync and also covers rsnapshot and rdiff-backup. This includes an explanation of how hard links work and also looks at special issues and possible gotchas: mail stores and databases, large filesystems and memory use. One omission is any treatment of ssh forced commands, which find a natural use in automated rsync backups.
Part III has two chapters. A long chapter on commercial utilities doesn't discuss specific utilities or systems. Rather, it considers possible extreme demands — huge datasets, "aggressive" requirements — and gives an overview of some of the features offered by higher-end systems, such LAN-Free Backup, Continuous Data Protection, archives and Hierarchical Storage Management. It ends with an overview of features to consider in evaluating solutions.
A chapter on hardware takes a similar approach, again proceeding without mentioning any specific products. It surveys the different kinds of tape drives and the options for midrange tape types. It describes the different kinds of optical recording methods. It explores the variety of possibilities for disk targets. And it evaluates the possibilities, advantages and disadvantages of virtual tape libraries.
Part IV covers bare-metal recovery, with separate chapters on Solaris, Linux and Windows, HP-UX, AIX, and Mac OS X, by different authors. Here only the Linux/Windows chapter was relevant to me: I found it useful, thought it is brief and more like a "howto" in its approach than most of the material in Backup and Recovery.
Part V deals with databases. An introductory chapter covers general issues with backing up databases, then there are separate chapters by specialists on Oracle, Sybase, DB2, SQL Server, Exchange, PostgreSQL, and MySQL. These chapters are pitched at system administrators who aren't specialist database administrators, and I found they improved my understanding of databases quite generally.
Part VI has two miscellaneous chapters. I skipped the chapter on VMware — a general chapter on virtual servers would have been more interesting. The final chapter places backups in their broader context, considering broader data protection issues, the differences between backups and archives, and planning for disaster recovery.
There's an immense amount in Backup and Recovery, which has seven hundred pages of text and no redundancy or padding. While some of it is specialised, a focus on underlying ideas and methodologies rather than "how to" instructions means that much of it will appeal to a broad audience. Pretty much every system administrator will find good value in it: even if someone else is responsible for backing up their systems, there are many ways in which an understanding of backups is still important.
May 2007
- External links:
-
- buy from Amazon.com or Amazon.co.uk
- details at O'Reilly & Associates
- information from W. Curtis Preston
- Related reviews:
-
- books about computing
- books published by O'Reilly & Associates