5.1.1 Overview of a cluster system
A cluster system contains multiple server systems, which work together as a single system. If a failure occurs on one server, job processing can continue on another server.
A cluster system consists of a host that performs processing and a host that is on standby to take over processing if a failure occurs. Servers that execute jobs are called primary servers. Servers that are ready to take over a job if a failure occurs on a primary server are called secondary servers. If a failure occurs, the secondary server takes over for the primary server to prevent operations from being disrupted. This is called a failover.
Failovers are performed in units of logical servers, called logical hosts. Any applications running in a cluster system must operate in a logical host environment to enable failovers for continuous operations. Applications running on a logical host are independent of physical servers and can operate on any server.
A logical host consists of three elements: an application running as a service, a shared disk, and a logical IP address. An application running as a service, such as JP1, stores data on a shared disk and uses a logical IP address for communication.
The following table shows the components of a logical host.
|
Logical host component |
Description |
|---|---|
|
Service |
An application, such as JP1, that runs in a cluster system. If the logical host for the primary node fails, the logical host for the secondary node starts the service using the same name, in order to take over. |
|
Shared disk |
A disk device connected to both the primary and secondary nodes. Information that will be inherited if a failover occurs (definitions, execution states, and so on) is stored on the shared disk. If a failure occurs on the primary logical host, the secondary server takes over the connection to the shared disk. |
|
Logical IP address |
An IP address assigned while a logical host is operating. If the primary server fails, the secondary server takes over the same logical IP address. This allows clients to access the same IP address as if a single server is always running. |
The following figure shows access during normal operations and after a failover.
|
|
![[Figure]](GRAPHICS/ZU030100.GIF)
While the primary server is running, on that server the shared disk and logical IP address are assigned and the services operate. If a problem occurs on the server, the secondary server takes over the shared disk and logical IP address, and restarts the same services that were on the primary server. Thus, although the physical server changes during a failover, since the secondary server takes over the shared disk and logical IP address, the change is transparent to clients.