Linux High Availability HOWTO

Harald Milz, hm@seneca.muc.de

1. About this document

2. Introduction

• 2.1 Who Needs Highly Available Systems?
• 2.2 Linux In The HA Environment
• 2.3 Identifying Single Points Of Failure
• 2.4 Relation To Commercial Solutions
• 2.5 General Information About Linux

3. Disclaimer

4. Marketing Aspects

5. High Availability Terminology

• 5.1 Sample Configuration

6. Overall Linux-HA Structure

• 6.1 Hardware Aspects
• 6.2 Software Considerations
• 6.3 Configuration File Contents
• 6.4 Event Structure and Sequence

7. Attaching Shared Storage

• 7.1 Disk Technologies -- SCSI
• 7.2 Disk Technologies -- Serial Storage Architecture (SSA)
• 7.3 Other Disk Technologies
• 7.4 Raw Disk Devices
• 7.5 Accessing Disks Safely
• 7.6 Redundant Disk Setups
• 7.7 Filesystems
• 7.8 Other internal storage devices

8. Setting Up Redundant Networking

• 8.1 IP Address Takeover (IPAT)
• 8.2 MAC Address Takeover
• 8.3 TCP vs. UDP
• 8.4 Cluster Node's Lifesign: Heartbeat
• 8.5 Keeping Track Of The Network Status
• 8.6 Non-IP Heartbeat

9. Using PCMCIA Cards for Hotswap

• 9.1 Use Cardbus Adapters!
• 9.2 Fast Ethernet
• 9.3 Fast/Wide/Differential SCSI

10. To-Do List

11. Future Extensions

12. Other Changes to the Linux Operating System

• 12.1 Booting off mirrored root filesystems

13. Linux-HA At Linux Kongress May 21-23, 1997

14. Commercial Linux HA Packages

15. Credits

16. References

17. Changes

18. Glossary