2.2.3 Configuration in cluster system

(1) Prerequisites of cluster system

When using JP1/AJS3 for Cloud Service Applications in a cluster system, the following prerequisites must be satisfied:

1. The cluster software must be a program that can control the following 2 to 4:

2. It must be able to use a shared disk that can be taken over from the execution system to the standby system.

   In particular, the following conditions must be satisfied:

   • The shared disk must be assigned before JP1/AJS3 for Cloud Service Applications is started.

   • The assignment of the shared disk must not be released while JP1/AJS3 for Cloud Service Applications is being executed.

   • The assignment of the shared disk must be released after JP1/AJS3 for Cloud Service Applications is stopped.

   • Exclusive control is carried out so that the shared disk can be accessed only from the execution system node.

   • Files must be protected by such a system as a file system with a journal function, so that the files will not be erased in case of a system failure.

   • Even in a case of failover, the contents written to files must be guaranteed and taken over.

   • Failover can be forcibly executed even if an ongoing process is using the shared disk.

   • Recovery processing that is performed if a failure in the shared disk is detected must be controlled by the cluster software and JP1/AJS3 for Cloud Service Applications does not have to be aware of the recovery processing. If starting or stopping JP1/AJS3 for Cloud Service Applications is necessary as an extension of the recovery processing, the cluster software must request JP1/AJS3 for Cloud Service Applications to execute starting or stopping.

3. The IP address used for a cluster operation must satisfy the following conditions:

   • Communication must be able to be performed using an IP address that can be taken over and is used for a cluster operation.

   • The IP address used for a cluster operation must be uniquely determined from the host name used for a cluster operation.

   • The IP address used for a cluster operation must be assigned before JP1/AJS3 for Cloud Service Applications is started.

   • The IP address used for a cluster operation must not be deleted while JP1/AJS3 for Cloud Service Applications is being executed.

   • Correspondence between the host name used for a cluster operation and the IP address used for a cluster operation must not be changed while JP1/AJS3 for Cloud Service Applications is being executed.

   • The IP address used for a cluster operation must be deleted after JP1/AJS3 for Cloud Service Applications is stopped.

   • Recovery processing that is performed if a network failure in the shared disk is detected must be controlled by the cluster software, and JP1/AJS3 for Cloud Service Applications does not have to be aware of the recovery processing. If starting or stopping JP1/AJS3 for Cloud Service Applications is necessary as extension of the recovery processing, the cluster software must request JP1/AJS3 for Cloud Service Applications to execute starting or stopping.

4. The host name used for a cluster operation must be set in the hosts file or name server and can be transmitted through TCP/IP communication.

Unless the above conditions are satisfied, a problem may occur in the operation of JP1/AJS3 for Cloud Service Applications. For example, the following problems may occur:

• If data that was written by the execution system break at the time of failover,

  A problem of an error, data loss, or failure in starting may occur and JP1/AJS3 for Cloud Service Applications cannot operate correctly.

• If recovery processing is not performed even when an NIC failure occurs,

  A communication error occurs and JP1/AJS3 for Cloud Service Applications cannot operate correctly until NIC is changed through control by the cluster software or the server fails over to another server.

To Page Top

(2) System configuration to be supported

An example of system configuration for operating JP1/AJS3 for Cloud Service Applications in a cluster system is shown below.

Figure 2‒9: Example of system configuration for operating JP1/AJS3 for Cloud Service Applications in a cluster system

Legend:

JP1/AJS3 for CSA: JP1/AJS3 for Cloud Service Applications

If a cloud service is supported, JP1/AJS3 for Cloud Service Applications created on a cloud can be configured in a cluster system.

Important

• If a logical host environment is created for each of JP1/AJS3 for Cloud Service Applications, JP1/AJS3 - Manager, and JP1/AJS3 - Agent, they will fail over, in case of a failure, without affecting one another. It is therefore recommended to construct JP1/AJS3 for Cloud Service Applications, JP1/AJS3 - Manager, and JP1/AJS3 - Agent on different logical hosts.

• When considering system configuration during a cluster operation, a host name in FQDN format cannot be used as a logical host name.

• JP1/AJS3 for Cloud Service Applications can be set up with one logical host per physical host. This means that plural JP1/AJS3 for Cloud Service Applications of different logical host names cannot be set up on one physical host.

• JP1/AJS3 for Cloud Service Applications supports an active-standby cluster configuration. It does not support an active-active cluster configuration.

• When constructing JP1/AJS3 for Cloud Service Applications in a cluster configuration in a cloud environment, an IP address cannot be used as the host name of the JP1/AJS3 for Cloud Service Applications server.

For details on constructing JP1/AJS3 - Manager and JP1/AJS3 - Agent in a cluster system, refer to the description on the configuration example of a cluster system in the JP1/Automatic Job Management System 3 System Design (Configuration) Guide.

To Page Top