10.1.1 Overview of a cluster system
A cluster system aims at achieving high availability, and is composed of an active server and a standby server. The active server is used for processing, and the standby server is available to take over processing should the active server fail. Thus, a cluster system can prevent interruptions of business operations and improve availability. This behavior of taking over processing when a failure occurs is called failover.
The software that controls the entire cluster system is called cluster software. The cluster software monitors whether the system is running normally, and when it detects an abnormality, it performs failover to prevent interruptions of business operations.
Remark: Some cluster systems might aim at load balancing, but that functionality is not described in this manual.
|
|
![[Figure]](GRAPHICS/ZU080000.GIF)
Applications, such as JP1/Service Support, run on a logical host. A logical host is a logical server controlled by the cluster software. The logical host is the unit of failover. A logical host uses a logical host name, a shared disk, and a logical IP address, all of which can be taken over by the standby server. Applications executed on a logical host store data on the shared disk, and communicate by using the logical IP address. Thus, failover is available independently from physical servers.
Using JP1/Service Support on a logical host environment of a cluster system is called cluster use.
- Note
-
Term "logical host"
In this manual, the term logical host is used as the term indicating the unit of failover, but some cluster software and applications use the terms group or package instead. Check the corresponding terms as necessary, referring to the manual for the cluster software or other materials.
Note that a physical server is called a physical host, while a logical server (which is a unit of failover) is called a logical host. The host name of the physical host (which is displayed when the hostname command is executed) is called the physical host name, and the IP address corresponding to the physical host name is called the physical IP address. Also, a physical host uses the local disk, which is the disk dedicated to the corresponding server and cannot be used by other servers.
- Note
-
Term "node"
In this manual, the server that processes business operations is called the active node, and the server that is prepared to take over in event of a failure is called the standby node. The active node and the standby node are switched each time a failover occurs.
In addition, in order to distinguish the two nodes when the system is configured or during environment setup, the node that first runs as the active node is called the primary node. The node that first runs as the standby node is called the secondary node. After operations start, the active node and the standby node can switch repeatedly due to occurrence of failures, but the relationship between the primary node and the secondary node does not change.