Synopses & Reviews
Master AFS, the Internet's most scalable, manageable distributed file system.
One technology already delivers the scalability, flexibility, power, and centralized file systems management the entire computer industry keeps promising: the Andrew File System (AFS). That's why hundreds of large institutions worldwide are already using AFS to reliably support their users' distributed file storage needs.
Managing AFS is the first book that brings together the critical information you need to build and maintain a reliable AFS-based distributed file system. Learn how AFS works, how it implements the client/server model, and why it's uniquely well-designed to serve the needs of today's organizations. Then, discover how to:
- Make the right architectural and topology choices to optimize AFS up front
- Take full advantage of AFS' powerful large-scale administration tools
- Slash desktop administration costs with AFS file caching and heterogeneous system support
- Set up and maintain user accounts(including how to create “dataless clients”
- Archive AFS data using built-in tools or third-party options
- Improve reliability through replication and automatic client failover
- Provide unified file services for UNIX® and Windows NT® desktops
Managing AFS includes detailed coverage of AFS security, including user authentication and Kerberos; AFS debugging and monitoring; solutions for providing access to Windows NT clients; and much more. You'll also find real-world AFS case studies, and a step-by-step guide to evaluating, piloting, and rolling out AFS throughout your organization and connecting to other AFS sites around the world.
If you're not looking for a better distributed storage solution yet, you will be soon. There's only one best-of-breed solution: AFS. And there's only one definitive guide to this revolutionary technology: Managing AFS.
Written for UNIX system administrators, this guide deals with AFS, a high-end UNIX filing system developed at Carnegie Mellon University and used widely in many industries such as banking and finance. This title explains how to manage AFS to its greatest effect including the installation of an adequate server setup to handle thousands of clients with a minimum of administrator and hardware overhead.
About the Author
RICHARD CAMPBELL has worked for the last five years as a distributed computing consultant on Wall Street. Previously, he was a senior researcher at the Institutional File System Project at the University of Michigan, which established AFS as the distributed file system for the campus of over 30,000 users.
Table of Contents
Each chapter concludes with a Summary.
1. Architectural Overview.
Beginnings. Benefits of AFS. Global Filesystems. Drawbacks. Other Sources of Information.
2. AFS Technology.
Client Configuration. Reading Data. Writing Data. Server Configuration. Volume Management. Server Processes. Network Protocol. AFS Extras.
3. Setting Up an AFS Cell
Preliminary Decisions. Hardware Requirements. Installation Overview. The Basic Overseer. File Services. Database Services. Server Machine Types. Adding More Machines. Ubik. Time Synchronization.
4. AFS Volumes and Files.
Volume Operations. An Example Volume. Backup Volumes. Moving Volumes. The Volume Location Database. Replication. Client Path Preferences. Conventional Namespaces. Deleting Volumes. Querying the Databases. Suggested Practices.
5. Client Administration.
Configuration. Cache Manager. Finding AFS Servers. Protocol Optimizations. Windows NT Clients. PC-Enterprise. Installing Clients and AFS Paths. Convenient Path Names. Set-User-Identifier Programs. Time Services. Messages.
6. Managing Users.
Kerberos. AFS's Kerberos. User Databases. Setting Up a User Account. Kerberos and Passwords. Ticket Lifetimes. The uss Utility.
7. Using AFS.
Authentication. Protection Data. Access Control Lists. Negative Rights. Groups and ACLs. More Group Management. Machine Groups. Volumes Revisited. Backup Volumes. Checking for Errors. AFS versus UNIX. Programming Issues. System Software. Using AFS on Windows NT.
8. Archiving Data.
The Archive System. Volume Sets. Dump Levels. The Tape Coordinator. Tape Labels. Running the Dumps. Querying the Databases. Getting the Data Back. Automating Dumps. Saving the Backup Database. Common Strategies. Volume Dumps. AFS Database Archiving. Third-Party Software. Legato's Networker. IBM's ADSM. PDC's BudTool
9. More AFS Administration.
Administrative Credentials. Server Management. Updating AFS Binaries. Job Notification. KeyFile Management. Changing the Cell Name. Database Servers. File Servers. Salvager Data. Networking Support. NFS-AFS Gateways. Administration Examples. E-Mail, NetNews, and the Web. Third-Party Software. Other Administration Tools.
10. Debugging Problems.
File Server Performance. File Server Problems. Salvaging Files. Database Servers. Ubik Debugging. The scout Monitor. Afsmonitor. AIX Auditing. Client Debugging. Cache Problems. Disaster Recovery. Crashing. Version Control. Security Issues.
11. Large-Scale Management.
Case Study: IBM. Case Study: Morgan Stanley. Case Study: University of Michigan. Case Study: Multiresident AFS.
12. Implementing AFS.
The Business Case. An AFS Project. Operational Buy-In. Ongoing Work. AFS Futures. DCE and DFS. The Competition.
12. Appendix A: AFS Command Suite