When JP1/Software Distribution Managers are used in a hierarchical configuration, a hierarchy with the relay manager of each department at its apex is treated as one of the JP1/Software Distribution systems. This provides the following benefits:

• Different client identification keys can be used for departments
  In a JP1/Software Distribution system, individual hosts must be identified using standardized host names or IP addresses. When JP1/Software Distribution Managers are configured in a hierarchy, the choice of host names or IP addresses can be made for each relay manager instead of at the system level.
  
• Each department can have unique management items
  JP1/Software Distribution has a facility that enables the system administrator to create unique management items for collecting needed information, such as user inventory and registry information. These management items can be set separately for the central manager and for the relay managers. In a large network, it may be desirable to manage some information on a system-wide basis and other information by department. Information appropriate to each level can be collected in this way.
  For details about managing registry information, see 2.2.1 Acquiring system information. For details about managing user inventory information, see 2.2.3 Acquiring user inventory information.
  

If a relay manager is placed under the central manager, the central manager cannot perform operations on system configuration components under the relay manager (for example, it cannot add, move, or remove lower relay systems or clients). Only the relay manager can perform operations on the system configuration components below it.

Consequently, if you build a large-scale JP1/Software Distribution system in which JP1/Software Distribution Managers are configured in a hierarchy, at each main site you should first build a basic system that has a relay manager at its apex, and then create system configuration information at each of those relay managers. After that, you can install the central manager and define the managing servers at all the main sites as relay managers.

After you set up, at the central manager, a JP1/Software Distribution Manager at each main site as a relay manager, you must collect information on the system configuration under each relay manager. To perform a batch operation that collects all this information from one relay manager, execute a Get system configuration information job with the relay manager as the job destination. The system configuration information managed by that relay manager is then transferred and displayed in the System Configuration window of the central manager. To collect, from a higher relay manager in a multiple-level hierarchy, the system configuration information of a lower relay manager, specify the lower relay manager as the destination.

The following figure shows the general procedure for collecting system configuration information from relay managers by executing a Get system configuration information job.

Figure 5-3 Collecting system configuration information from relay managers

After the information has been collected, all relay managers, relay systems, and clients within the network can be handled as normal destinations. Also, if you use the central manager to create IDs, you can set up all relay managers and relay systems within the network as relays that can manage the IDs.

To keep the information in the System Configuration window up to date, when you set up the central manager and relay managers you should specify the option that automatically registers and applies system configuration information. Changes to the system configuration under a relay manager will then be relayed automatically to the central manager.

From a managing server, you can normally specify any of three types of job destinations: a host group, an ID group, or a single client. However, if JP1/Software Distribution Managers are arrayed in a hierarchical configuration, the highest manager (the central manager) can execute a job specifying all lower clients (all-lower-clients job) as the destination. All lower clients means all hosts under the relay managers. If you execute an all-lower-clients job for a specific relay manager, the relay manager that received the job executes it for all its lower clients.

The central manager can check the execution results of an all-lower-clients job based on the number of clients (total count and count per set of execution results). You can also check the execution status of each client if you specify, during relay manager setup, that the execution results of all-lower-clients jobs are to be recorded.

If you re-execute an all-lower-clients job from the central manager, the relay manager automatically detects and re-executes the job for only those clients whose job status is Waiting for transmission or Error.