Interstage Shunsaku Data Manager Operator's Guide - Microsoft(R) Windows(R) 2000/ Microsoft(R) Windows Server(TM) 2003 - - UNIX -

8.1.1 Failover Function for Director Servers

A failover function for director servers can be used with cluster systems, in preparation for unexpected circumstances such as hardware faults or software error with director servers.

This section explains:

• What is the failover function for director servers?
• Cluster System
• Failover Operation Configuration

What is the Failover Function for Director Servers?

The failover function for director servers works by using cluster systems where, in addition to the server that is operating (hereafter referred to as the 'active server'), another server (hereafter referred to as the 'standby server') is prepared in case something goes wrong with the active server.

[Figure : Example of the Failover Function in Cluster Systems]

Using the failover function for director servers allows users to perform searches and updates to Shunsaku without being aware of which server is the active server and which is the standby server.

[Figure : Failover Function for Director Servers]

If something unexpected happens to the active server (such as a hardware fault or software error), the cluster system detects the error and performs a failover operation.

The failover operation minimizes system downtime by starting Shunsaku on the standby server and handing over control of jobs to the standby server. After jobs have been taken over, users can continue to perform searches and updates using Shunsaku in exactly the same way as before the server error occurred.

Additionally, because the server operating normally takes over the jobs from the failed server, the failure can be rectified without interrupting these jobs.

Cluster System

The failover function for director servers runs on the following cluster systems.

• Microsoft Cluster Service (MSCS)

• PRIMECLUSTER V4.1A30 or later

Using the HA functions requires knowledge of these cluster systems. Refer to the manuals for each cluster system for more information.

Failover Operation Configuration

The failover function for director servers can be used with the following operation configurations of cluster systems.

• 1:1 Active/Standby Configuration
• N:1 Active/Standby Configuration

1:1 Active/Standby Configuration

In this operation configuration, there is one standby server for each active director server.

If an abnormality occurs with the active server, the cluster system fails over to the standby server, so jobs do not need to be interrupted.

[Figure : 1:1 Active/Standby Configuration]

N:1 Active/Standby Configuration

In this operation configuration, there are multiple active servers for each standby server.

This operation configuration reduces the cost of standby servers because, unlike the 1:1 active/standby configuration, there is no need to prepare a standby server for each active server.

[Figure : N:1 Active/Standby Configuration]