10.1 Overview of cluster systems
A cluster system consists of multiple server systems operated as a single system. The purpose of a cluster system is to provide uninterrupted service even if an error occurs. The server that is currently executing tasks is the primary node and the other systems that are on standby are secondary nodes. If an error occurs in the current system, a secondary node takes over and continues the operation. This prevents operation from being interrupted even if an error occurs.
A cluster system is also referred to as a node switching system because the primary node that executes tasks is switched over to a secondary node.
In JP1/AJS3, in addition to errors occurring in the system, cluster operation can be used for errors occurring in JP1/AJS3 service processes. JP1/AJS3 supports cluster operation by virtually defining multiple logical hosts on a single host so that JP1/AJS3 services can be started on each logical host. Furthermore, cluster operation in which shared disks and IP addresses are inherited by each logical host is also supported.#
The following figure shows operation when multiple JP1/AJS3 services run on a single host in a cluster system.
|
This section describes the prerequisites for cluster systems supported by JP1/AJS3, system configurations, processing procedures for a cluster configuration, and actions to be taken if a failover occurs. For details about the setup for using a cluster system, see 8. Setup for Operation in a Cluster System in the JP1/Automatic Job Management System 3 Configuration Guide (on a Windows host) or see 17. Setup for Operation in a Cluster System in the JP1/Automatic Job Management System 3 Configuration Guide (on a UNIX host).
- #
-
For JP1/AJS3 - Web Console, multiple logical hosts cannot be virtually defined for one host. This is because only one JP1/AJS3 - Web Console service can run on one host. JP1/AJS3 - Web Console supports cluster operation by allowing the cluster nodes to inherit the processing of the service, as well as the shared disk resources and IP addresses, from each other.
For details about the prerequisites and system configurations of the cluster system supported by JP1/AJS3 - Web Console, see 10.1.3 Cluster configurations supported by JP1/AJS3 - Web Console.
- Supplementary note on defining jobs that use user programs
-
When you define a job that uses a user program, you must be careful about re-execution. For example, if you re-execute a database update job, the same update is performed twice. If you re-execute a print job, the same data is printed twice. We recommend that you adjust the jobs with user programs for failovers, such as outputting the execution log or defining a judgment job, to prevent multiple executions of the same job.