Systemwalker Service Quality Coordinator Installation Guide

3.8 PRIMECLUSTER Cluster System Operation Model

This section explains how to install a Manager or Enterprise Manager on a PRIMECLUSTER cluster system.

Operating a Manager or Enterprise Manager in a cluster system improves management business availability because even if a problem causes one node to fail, management business will simply pass to the remaining normal node.

Execution environment

This function is available with Systemwalker Service Quality Coordinator Enterprise Edition.

It can only be installed for use with Enterprise Managers and Managers.

Required privileges

UNIX

System administrator (superuser) privileges are required.

Before performing this procedure

PRIMECLUSTER must be installed and its environment set up correctly before the procedure described in this section is performed.

For the procedure used to install PRIMECLUSTER and create a new cluster environment, refer to the PRIMECLUSTER Installation and Operation Guide as required.

Note

This section explains how to construct a Manager/Enterprise Manager environment in a PRIMECLUSTER cluster environment, and contains references to PRIMECLUSTER settings. For this reason a basic understanding and practical knowledge of PRIMECLUSTER is required.

Note that a Manager cannot be installed on a cluster system in the following situations:

• When the Manager is in a redundant configuration

The following cluster systems and types are supported.

[Cluster systems]

Solaris

• Fujitsu PRIMECLUSTER Enterprise Edition
  4.1/4.1A10/4.1A20/4.1A30/4.1A40/4.2A00/4.2A30/4.3A00/4.3A10

• Fujitsu PRIMECLUSTER HA Server
  4.1/4.1A10/4.1A20/4.1A30/4.1A40/4.2A00/4.2A30/4.3A00/4.3A10

• Fujitsu PRIMECLUSTER Clustering Base
  4.1/4.1A10/4.1A20/4.1A30/4.1A40/4.2A00/4.2A30/4.3A00/4.3A10

Linux

• Fujitsu PRIMECLUSTER Enterprise Edition 4.2A00/4.2A30/4.3A00/4.3A10

• Fujitsu PRIMECLUSTER HA Server 4.2A00/4.2A30/4.3A00/4.3A10

• Fujitsu PRIMECLUSTER Clustering Base 4.2A00/4.2A30/4.3A00/4.3A10

  
  

[Cluster types]

• 1:1 active/standby configuration (One node remains on standby for another node)

  In this section, the node where management business is carried out is referred to as the active node, the node that remains on standby to inherit business in the event of a problem is referred to as the standby node, and the process by which the business passes from the active node to the standby node in the event of an error is referred to as failover. Further, the disk that is shared by both the active and standby nodes is referred to as the shared disk.

  
  

Point

When setting the host name or IP address of a Manager involved in cluster operation (such as when installing an Agent, Proxy Manager or Operation Management Client), specify a logical host name or logical IP address. This will enable the operation to be performed without the need to make a distinction between the active node and the standby node.

The following sections explain the Enterprise Manager clustering procedure for two-tier systems, and the Manager clustering procedure: