If you will be using the local system as a managing server or a relay system, set items for tuning the CPU, memory, and networks. For an overview of tuning, see 6.1.7(2) Adjusting the maximum number of hosts that can be connected concurrently in the Description and Planning Guide.

Figure 5-5 Relay System Customization page

The following describes items unique to this page. The other settings are the same as on the Server Customization page for the Server setup for JP1/Software Distribution Manager. For details, see 4.2.4 Server Customization page.

Number of clients that can be connected at one time

You can limit the number of subsystems that can be connected at one time. Specifically, this value becomes the number of socket connections. Each socket connection is counted from the time it is established until the time it is released. Determine the value using the formula shown below:

Number of subsystems that can be connected at any one time =

number of directly connected systems + number of Packagers to be connected x 2

The minimum value is 4, the maximum is 9,999, and the default is 50.

If the network traffic is heavy, resulting in many packet collisions, you can reduce the traffic by reducing this value, thereby reducing the number of subsystems that can be connected at one time.

A connection request from a subsystem that exceeds this value will be denied, but once there is room in the number of connections, execution requests are sent again.

Max. number of relays or clients in which jobs can execute concurrently

Specify the maximum number of subsystems that can process jobs concurrently. Specifically, this value becomes the number of startup messages the managing server sends at one time to subsystems (in other words, the number of clients that can execute jobs concurrently). If users execute jobs for more clients than the value specified in this field, the system splits the jobs on the basis of the specified value and executes them.

The minimum value is 0, the maximum is 9,999, and the default is 20. If you specify 0, the managing server no longer sends startup messages to the lower systems, disabling job execution initiated by higher systems as well as target system startup using the client control facility. If the size of the files to be distributed is large (10 MB or greater), the LAN load may become high even when only a few clients are connected. Therefore, you should specify a value that matches the network performance. The result of maximum number of relays or clients in which jobs can execute concurrently x package size becomes the amount of data that flows over the network at one time.

Specify a value that does not exceed the value specified in Number of clients that can be connected at one time. If the subsystems are connected through a circuit-switched network, specifying a number greater than the number of agreement lines serves no purpose.

Maximum cache size for the management file

Specify the maximum size of the information about the jobs executed from the higher system or relay system (management files) that is to be cached in the relay system's memory. The minimum value is 0, the maximum is 1000000 (1 gigabyte), and the default is 100000 (100 megabytes). If you specify 0, the function for caching management files is disabled.

When a job is created, its management file is cached in memory until the job is deleted. If the total size of the cached management files exceeds the maximum size you have specified, job processing throughput will deteriorate. To avoid such deterioration of job processing throughput, you should take care to specify a maximum cache size for the management file that is appropriate to the actual scope of your operations. Determine the value based on the estimated value of each item in the following formula:

Cache size for the management files (kilobytes) =

Number of jobs stored in the relay system and that are executed from the higher system or this relay system

x number of destinations for each job

x number of packages for each job (applicable to remote installation jobs)

x 1 kilobyte

When the maximum size would be exceeded because of a need to cache another management file, the least frequently referenced management file is deleted so that the new management file can be cached.

Record the system/software information answered from the lower clients

Specify whether or not system information and installed-package information that are sent from a lower system when jobs are executed is to be recorded. By default, this check box is selected.

System information

Records system information that is reported from a lower system when jobs are executed. By default, this check box is selected.

Installed package information

Records installed package information that is reported from a lower system when jobs are executed. By default, this check box is selected.

Record the results of ID group jobs

Specify whether or not the results of ID group jobs that are executed by clients and that terminate normally are to be recorded. ID group jobs that result in an error are recorded regardless of this setting. To check the results of ID group jobs in the relay system, the client's execution results must be managed in the relay system that manages the ID group. The results from each client are deleted when the ID group job is invalid or cancelled. The default is that this check box is not selected.