19.5.1 Procedure for setting the N-to-1 recovery system

(1) Example of system configuration

The following figure shows an example of system configuration of the N-to-1 recovery system. Note that the subsequent points describe examples of building the system using this example of system configuration.

Figure 19‒6: Configuration example of the N-to-1 recovery system (In UNIX)

An overview of this example is described as follows:

• Deploy four executing node servers and one recovery server. The J2EE servers in the four executing nodes are configured so that the settings of the J2EE applications and resource adapters are the same. Only the alias IP address and the shared disk device to be mounted are different.

• Set the same name for the J2EE server in the executing node and the standby node. In this case, as you have deployed a J2EE server called MyServer in the executing nodes, also deploy a J2EE server called MyServer in the standby node.

• Build and perform operation management of the system using the management portal of Management Server.

• Operate Administration Agent in the executing node and not in the standby node.

When operating the N-to-1 recovery system, build an Application Server that satisfies the following conditions:

• Use a global transaction that uses OTS. Save the status file of the transaction in the shared disk device.

• Operate the Naming Service as an in-process.

• In the J2EE server of the standby node, import and deploy a resource adapter with the same settings as all resource adapters that you are using in the J2EE servers of the executing node.

• In the N-to-1 recovery configuration, build multiple virtual hosts within one cluster. The executing node and recovery server configure the virtual host with the J2EE server deployed in a cluster configuration. In the N-to-1 recovery configuration, the J2EE server operates within the virtual host rather than the cluster. Therefore, set the IP address and host name of the virtual host.

  To deploy Application Server in a cluster configuration, set the alias IP address, and specify settings so that the running node inherits the alias IP address and the client does not consider the nodes in the cluster. In the N-to-1 recovery configuration, the alias IP address becomes the IP address of the virtual host. A configuration example of the virtual host in the N-to-1 recovery configuration is as follows:

  Figure 19‒7: Configuration example of the virtual host in the N-to-1 recovery configuration

To Page Top

(2) Procedure for system setup

When integrating with the HA monitor, you must set the files of the management portal and HA monitor. The following figure shows the procedure for setting the N-to-1 recovery system:

Figure 19‒8: Procedure for setting the N-to-1 recovery system (In UNIX)

Steps 1 to 9 of the figure are described as follows:

1. Create N number of partitions in the shared disk to build a file system.

2. Allocate the shared disk to the system.

   When allocating the shared disk to the system, specify the same mount destination directory in the active node and spare node.

3. Set the environment of the cluster server with the management portal.

   Specify the settings for using the HA monitor in 'Settings of Cosminexus Management Server', 'Configuration definition of management domain' and 'Environment setup of a logical server' of the management portal. For details, see 19.5.2 Setting the environment of the cluster server.

4. Edit the configuration files.

   Specify various definition files of Administration Agent, Management Server, and Cosminexus HTTP Server. For details, see 19.5.3 Editing the configuration files.

5. Set the environment of the HA monitor.

   Define the environment of the HA monitor with the sysdef file. For details, see 19.5.4 Setting the environment of the HA monitor.

6. Create the shell script file.

   Create the shell script file for monitoring the Administration Agent, and for starting and stopping the logical servers on Application Server. For details, see 19.5.5 Creating a shell script.

7. Set the server-compliant environment.

   Using the servers file of the HA monitor, define the environment of the active-node server and spare-node server to be operated on the nodes. For details, see 19.5.6 Setting the server-compliant environment.

8. Set the status of LAN.

   Using the LAN status setup file of the HA monitor, specify the IP address of the LAN adapter, and define switching of LAN in the HA monitor. For details, see 19.5.7 Setting the LAN status.

9. Using the management portal and commands of the HA monitor, specify the settings for Application Server.

   Using the management portal and commands of the HA monitor, deploy Application Server in a cluster configuration, and set the J2EE applications and resource adapters. For details, see 19.5.8 Setting Application Server.

   Reference note

   • In the N-to-1 recovery system, if there is timeout for the recovery process in the standby node due to a failure (when the server is down or a deadlock has occurred) in the database server, you must execute the recovery process manually.

     After finding the cause for timeout, execute the monbegin command corresponding to the executing node host that has stopped, and then execute the monact command corresponding to the executing node host that has stopped, in the standby node host.

     When an error occurs in the N-to-1 recovery system, see 2.5.8 If a problem occurs in N-to-1 recovery systems in the uCosminexus Application Server Maintenance and Migration Guide.

The files that you must set in the N-to-1 recovery system are different for the executing node and the standby node. The following table describes the necessity of setting files in the N-to-1 recovery system:

Table 19‒4: Necessity of setting files in the N-to-1 recovery system

Category		File type	Necessity of setting the file		Reference
			(N number of) Executing nodes	(One) Standby node (dedicated for recovery)
Setting Application Server	Setting the Administration Agent	adminagent.properties	Necessary (Individual settings in N nodes)	N	19.5.3
	Setting the Smart Agent	localaddr
	Setting Cosminexus HTTP Server	httpsd.conf#
	Setting the properties of the J2EE server	usrconf.properties#	Necessary (Common settings in N nodes)	Necessary (Almost same as the executing node, except for alias IP address)
Setting the HA monitor	Setting the environment of the HA monitor	sysdef	Necessary (Individual settings in each host)	Necessary (Settings for each of the N nodes)	19.5.4
	Creating the shell script file	monitor.sh start.sh stop.sh	Necessary (Usage of a common command for N nodes)	N	19.5.5
	Setting the server-compliant environment	servers	Necessary (Individual settings in N nodes)	Necessary (Settings for each of the N nodes)	19.5.6
		recover.sh	N	Necessary (Usage of a common command for N number of executing nodes. Differences are passed with arguments)
	Setting the LAN status	server-identification-name.up server-identification-name.down	Necessary (Individual settings in each host)	Necessary (Copy the configuration file of the executing node)	19.5.7

To Page Top

(3) Setup procedure for adding or deleting an executing node

The setup procedure for adding or deleting (degenerating) an executing node in the N-to-1 recovery system is described as follows.

Adding an executing node

1. Create a mount directory in the executing node to be added.

2. In the servers file of the standby node, add an entry corresponding to the executing node to be added.

   You do not require this step if the entry corresponding to the executing node to be added already exists.

3. In the standby node, execute the monbegin command corresponding to the executing node that you have added.

4. In the executing node, execute the monbegin command.

5. Using the management portal, build the execution environment of Application Server in the executing node to be added.

6. Using the management portal, start the executing node Application Servers as a batch.

Deleting an executing node

1. Execute the monend command in the executing node.

To Page Top