3.11.5 Configuration using server exclusive for recovery (N-to-1 recovery system)

(1) Features of the system configuration

This is a system configuration for resolving the transaction and releasing the resources when there is a problem in certain J2EE server and a global transaction is used in a configuration with redundant J2EE servers. Use load balancer to create redundant J2EE servers for N machines.

The prerequisites of this configuration are as follows:

• Used in a system using global transaction

• The CORBA Naming Service is started as in-process

• The status file of transaction is saved to shared disk

• The operations are performed using Management Server

Application Server is subject to node switching in N-to-1 recovery system.

The resource adapters used in executing node are required for standby node as well when configuring N-to-1 recovery system. The following three types of configurations are described as an example for allocation of resource adapters and relation with system configuration:

• When the configuration of J2EE applications and resource adapters is the same for N executing node machines

• When the configurations of J2EE applications differ and all resource adapters are identical for N executing node machines

• When the configurations of J2EE applications and resource adapters differ for N executing node machines

The following are the examples of respective configurations:

Figure 3‒53: Example of system configuration of N-to-1 recovery system (when configuration of J2EE applications and resource adapters is same for N executing node machines)

Note: For other legend items, see 3.2 Description of the system configuration.

Figure 3‒54: Example of system configuration of N-to-1 recovery system (when configuration of J2EE applications differ and all resource adapters are identical for N executing node machines)

Note: For other legend items, see 3.2 Description of the system configuration.

Figure 3‒55: Example of system configuration of N-to-1 recovery system (when configuration of J2EE applications and resource adapters differ for N executing node machines)

Note: For other legend items, see 3.2 Description of the system configuration.

Features

• The executing node and standby node are in an N-to-1 ratio.

• A shared disk device is required. A volume group for the number of executing nodes (N) is required in the shared disk device.

• The CORBA Naming Service is started as in-process.

• The same resource adapters as in the executing node are required in the standby node. If a separate resource adapter is imported for each J2EE server in the executing node, you must import respective resource adapters for J2EE server in the standby node.

  Also, a J2EE application is not required in the J2EE server on the standby node.

• When the database has a cluster configuration and if Application Server identifies only the virtual address (Logical address), you can still establish the connection.

Node switching flow

When there is a failure in any of the J2EE servers started by Application Server in the executing node and when a J2EE server of the server exclusive for recovery is started by cluster software, the transaction used by the failed J2EE server is stopped. Thereafter, the cluster software of the J2EE server machine that failed and the cluster services of the corresponding standby nodes are stopped.

To Page Top

(2) Processes to be started on each machine

This subsection describes the required software and processes of the respective machines.

To Page Top