Managing AFS: The Andrew File System 1st Edition

4.0 out of 5 stars 1 rating
ISBN-13: 978-0138027292
ISBN-10: 0138027293
Why is ISBN important?
This bar-code number lets you verify that you're getting exactly the right version or edition of a book. The 13-digit and 10-digit formats both work.
Scan an ISBN with your phone
Use the Amazon App to scan ISBNs and compare prices.
Have one to sell?
Loading your book clubs
There was a problem loading your book clubs. Please try again.
Not in a club? Learn more
Amazon book clubs early access

Join or create book clubs

Choose books together

Track your books
Bring your club to Amazon Book Clubs, start a new book club and invite your friends to join, or find a club that’s right for you for free.
More Buying Choices
10 used from $22.00

Black lives matter. Books to read right now.
click to open popover

Enter your mobile number or email address below and we'll send you a link to download the free Kindle App. Then you can start reading Kindle books on your smartphone, tablet, or computer - no Kindle device required.

  • Apple
  • Android
  • Windows Phone
    Windows Phone
  • Click here to download from Amazon appstore

To get the free app, enter your mobile phone number.


Special offers and product promotions

Editorial Reviews Review

Designed to handle terabytes of data and thousands of users distributed across large networks, the Andrew File System (AFS) works as a Unix and Windows NT add-on that replaces the standard Network File System (NFS). The basic idea of AFS is that all users of a network get a common picture of the file system even though it represents data that's stored on many different machines. Originally developed at Carnegie Mellon University and now commercially promoted by Transarc Corporation, AFS has a solid presence in very large commercial networks. Its popularity seems likely to trickle down to midsize organizations as businesses of all kinds become more network-intensive.

Campbell's book takes the ponderous set of AFS documentation and distills it into a collection of task- and question-oriented segments. The book opens with an analysis of AFS architecture that will prove useful to those thinking about deploying AFS. Managing AFS then covers hardware requirements, volume structure decisions, and client administration in depth. Following a discussion of AFS's implementation of Kerberos security, Campbell gets into task-oriented discussions, covering groups, rights, archive procedures, and disaster recovery. One interesting chapter presents AFS case studies, highlighting how IBM, Morgan Stanley, and the University of Michigan are using the file system. An appendix details the various AFS command suites in a sort of abbreviated man-page format.

Throughout Managing AFS, the text is clear and readable--even entertaining. If you're wondering how AFS might work on your network or you want to know the easiest way to set up a user account, this book meets your needs.

From the Inside Flap



This book describes the implementation, administration, and use of Transarc Corporation's AFS(r), the Andrew File System. This distributed system has several attributes which make it ideally suited for use in organizations trying to manage the constantly growing amount of file data needed and produced by today's operating systems and applications.

Even though most of the academic emphasis on AFS concerns its performance and semantics, much of the benefit of the system derives from its integrated approach to administration. The ability of AFS to support many more clients per server, to decrease network loads, to provide location-transparent access to data, and to automatically fail over to replica sites is reason enough to investigate the system. But the collection of tools supplied to administrators which provides behind-the-scenes control of the distributed system is equally important.

Over the last decade, AFS's use has steadily increased so that by now many hundreds of sites around the world are full-fledged members of an on-line, global, distributed file system. These sites tend to be quite large because AFS is particularly optimized to support the file storage needs of thousands of users. Yet given the ever-increasing sales of computers and their voracious disk appetites, mature solutions to medium- and large-scale computing sites will be needed by more and more people.

Audience and Scope
When you purchase AFS, you'll receive several manuals that come with the package. This book is not a replacement for serious reading of that official documentation. There are a multitude of options for almost all processes and utilities in the AFS command suite. Rather than catalog each argument, I hope to provide the reasons behind some of the design decisions which otherwise may appear arbitrary or counterintuitive. How and why the AFS clients and servers work as they do is the scope of this book. The following chapters describe the newer features of the system and highlight the latest advances with some explanations of their purpose; they do not describe the precise syntax or option to every possible command. For that, you can use the manual set supplied by Transarc, the developer of the commercial AFS product. The examples and suggestions for managing AFS are shown using the standard UNIX(r) command-line interface; the AFS server programs are available only for UNIX servers, so knowledge of basic UNIX commands and operations is a prerequisite for using the system.

If you've used NFS(r) or NetWare(tm) before and have managed desktops, file servers, and disks, you may be wondering what all the fuss is about; after all, these services have been around for years. But AFS is designed quite differently from these other file systems for reasons that make sense but take some explaining.

There's no simple linear way to discuss the mechanisms to manage and use AFS because most parts of the system are dependent on each other. I will try to create some order out of this interdependence at the risk of using some terms that are not fully described until later. For example, access control lists-used to detail who can access which files-are mentioned early on but are not defined until the middle of the book. The definition isn't required in the early stages, as long as you can trust that eventually the specifics will be explained.

Structurally, we'll begin with a broad overview and gradually introduce more and more detail, moving from the central servers to the distributed client desktops.

The first two chapters describe AFS in general terms.
Chapter 1 provides a general overview of the benefits of AFS, why it was developed in the first place, its particular design and drawbacks, and information about Transarc. Administrators, managers, and users will gain an understanding of the system from Chapter 1.
Chapter 2 introduces much of the technical vocabulary of AFS and describes its client/server architecture, including the caching protocol and file server organization. Here, the architecture described in Chapter 1 is put into concrete terms of the protocol used by clients and servers.
The next four chapters are devoted to the administrative specifics required to install and run an AFS cell. Chapter 3 discusses basic server management by itemizing the processes, operations, and systems that make the servers work, with special emphasis on the internal distributed database protocol. These issues are of concern when setting up an AFS cell.
Chapter 4 discusses AFS volume management issues, the very heart of AFS administration. While AFS is primarily thought of as a protocol by which client workstations access file data, the system's real value is in its support for large-scale administration. This chapter describes how AFS volumes are administered to create the global file namespace, to provide access to replicated data, and to provide users with on-line access to a backup version of their files.
Chapter 5 describes client configurations and includes suggestions for optimizing desktop access to AFS. The chapter also introduces the package client administration tool as well as two products that allow PCs to access AFS-Transarc's port to Windows NT(tm) and Platinum Technology's PC-Enterprise(tm).
Chapter 6 describes the user authentication process, management of the Kerberos database, and the use of standard MIT Kerberos, versions 4 and 5.
Chapter 7 details the user and developer view of AFS, including logging in to AFS, access controls, group management, and the slight differences in file system semantics. There's also a brief description of how PC users manage their AFS files with Transarc's NT port. Users not interested in administration and management can refer to this chapter for examples of how to use AFS.
Chapter 8 focuses on Transarc's built-in archiving system, ad hoc volume dump tools, and mechanisms to save other critical configuration data. It also includes information on integrating AFS with commercial backup systems.
Chapter 9 returns to the subject of overall administration, with more information on AFS server administrative responsibilities, insights into the Kerberos authentication system, providing access for NFS clients, issues regarding the installation of third-party software, and comments on a few of the public and commercial AFS administrative tools now available.
Chapter 10 continues the subject of management with a discussion of AFS's debugging and monitoring tools. The chapter focuses in depth on advanced issues of administering UNIX-based AFS servers.
Chapter 11 describes large-site administration and presents four case studies of successful use of AFS by global or otherwise interesting sites.
Chapter 12 concludes the book with some thoughts on how to evaluate and begin using AFS at your organization, and what alternative file systems are up to. Managers and administrators will gain a greater appreciation for the potential scope of an AFS implementation project.
Appendix A is a command-by-command description of the major AFS tools including all subcommands. As you read through the book, you can refer to this appendix for a reminder of the purpose of any commands. This listing also includes information on the precise authentication required to run each command.
Typographic Conventions
System operations are shown as command-line input and output, such as:
$ fs mkm alice user.alice
$ fs lsmount alice
'alice' is a mount point for '#user.alice'
For people unfamiliar with UNIX, $ is the standard UNIX Bourne shell prompt. Following UNIX convention, when you have logged in as the superuser or root, the shell prompt is #. These command sequences were executed on a Sun(r) workstation running the Solaris(tm) operating system with AFS version 3.4a and pasted into the text of this book. Output from AFS commands on other systems should be identical apart from dates and other ephemeral data; output from standard UNIX commands may differ more or less depending on your particular desktop operating system. Liberal use is made of standard programs such as date, touch and shell output redirection to demonstrate file creation. The cat program is often used to demonstrate that the files in question were created successfully.
The hypothetical organization used as an example throughout the examples, HQ, and its domain name, hq.firm, were non-existent at the time this book was written.
In writing this book I am indebted to the software engineers who put AFS together, first (and still) at Carnegie Mellon University and later at Transarc Corporation; to the computing community which has freely shared experiences and workarounds; to the University of Michigan's Center for Information Technology Integration, where some of my best friends work and play; and to the Union Bank of Switzerland for their support. Many people in these groups helped review this book; over the years, I've learned much abo

Product details

  • Paperback : 496 pages
  • ISBN-10 : 0138027293
  • ISBN-13 : 978-0138027292
  • Item Weight : 1.76 pounds
  • Product Dimensions : 6.9 x 1.1 x 9 inches
  • Publisher : Prentice Hall; 1st Edition (February 13, 1998)
  • Language: : English
  • Customer Reviews:
    4.0 out of 5 stars 1 rating