When you set up a client, you can use the Specify Higher Systems dialog box to specify multiple relay systems and assign a priority to each of them:

Figure 6-18 Specify Higher Systems dialog box

Add button

Adds a higher system to be a connection target. You can specify a maximum of eight higher systems. In the Add/Change Higher System dialog box, specify the product type and host name or IP address of the higher system that is to be added.

Update button

Updates the information about a selected higher system. In the Add/Change Higher System dialog box, which is displayed when this button is chosen, specify the product type and host name or IP address.

Delete button

Deletes the selected higher system from the list of higher systems.

Up button

Increases by 1 the priority of the selected higher system.

Down button

Decreases by 1 the priority of the selected higher system.

When you finish specifying the settings, choose the OK button. The Connection Destination page appears again.

A client checks with the higher relay system to see whether there are any jobs addressed to it. This operation is called polling. The method of polling the relay system is set up on the Default Running Status/Polling page of the Client Setup dialog box. The following two polling methods are available:

Hot standby:

With the hot-standby method, only one relay system is polled. Normally, the relay system with the highest priority, as set in the Specify Higher Systems dialog box, is the one that is polled. Only if that relay system cannot be connected is the relay system with the second-highest priority polled. If the second relay system cannot be connected, the relay system with the next-highest priority is polled, and so on, until connection is established with a relay system.

Multiple hosts:

With the multiple-hosts method, all relay systems specified in the Specify Higher Systems dialog box are polled. Polling begins with the relay system with the highest priority.

Note the following points about operating a client in a multi-polling environment:

• In the case of a client operating in a multi-polling environment, you can register ID groups in the first-priority relay system only.
• Under the hot-standby method, if the relay system with the highest priority recovers from an error, it is possible to have the client poll that relay system again. This method is explained below:
  

  If the client is resident:

  If the client is resident, either stop the relay system that is connected or execute a job such as Get system information from client via the relay system with the higher priority. If the relay system that is connected is stopped, the relay system to be polled will change the next time polling occurs. If a job is executed, the relay system to be polled will change after the job is executed.

  If the client is nonresident:

  If the client is not resident, set the server-polling format at system startup to Perform polling according to priority when the system starts. After the client is restarted, the relay system with the highest priority will be polled.

• When jobs are executed on all paths in an environment in which hot standby is used, the job on the path of the relay system to which the client connected is executed, but the jobs on all other paths remain in unexecuted status in the relay system. Therefore, you must periodically delete the jobs remaining in the relay system and the completed jobs from the managing server. If a large number of jobs remain in the relay system, its performance might suffer.
• When jobs are executed on all paths in an environment in which multiple hosts are used, the same job is received every time a client polls. In this case, to execute a remote installation job, we recommend that you deselect the Replace existing package check box under the job installation condition. If this check box is selected, a package is sent each time the job is executed, and this might place an undue load on the network. If the check box is not selected, any job that the client has already received normally terminates without a package being sent again. The job maintenance code in this case is 900090009000.