Linux on HP Integrity Servers: A System Administrators Guide

ServiceGuard for Linux (what I'll call ServiceGuard in this chapter) is a High Availability (HA) application. It is used in conjunction with other HA techniques to produce a loosely coupled cluster. An application running on a primary system can failover to a backup system should a failure or other switch over condition exist on the primary system. ServiceGuard can detect failures of a CPU, LAN, application, and other components.

All the systems and all the application packages have IP addresses. When a failure occurs, the another system runs the high availability package and the IP address of the package moves with the package to the backup system. If only a LAN card fails in a primary system, ServiceGuard will fail over to a standby LAN card rather than the backup system.

You can have many systems in a ServiceGuard cluster, all of which can be active. You don't need to have systems in a standby state waiting for a failure to occur.

This chapter provides background on ServiceGuard to give you an overview of its operation. This is not a detailed setup guide for Service-Guard. There are excellent ServiceGuard documents on www.docs.hp.com including Managing ServiceGuard for Linux.