SYSTEM MODES.\ There are two primary modes of system operation: single-user and multi-user. Single-user is a system state designed for administrative and maintenance activities which require complete and unshared control of the system. Single-user mode is sometimes called the maintenance mode. Single-user mode is entered via manual intervention during the boot process. Sometimes, however, the system will enter single-user mode if there are problems in the boot process that the system cannot handle on its own. Multi-user allows many users to all log onto the same CPU. Users can access different applications simultaneously or even the same application simultaneously. The kernel manages the different users by scheduling the use of the processing time as well as swapping programs and data in and out of memory through virtual memory to disk. The most important fact to remember is that the number of concurrent users depends on the amount of memory installed in the computer. Each user has a certain amount of memory set aside for his or her work, unless everyone is willing to tolerate slow response time from the network.

System Shutdown

While there are many occasions when shutting down or rebooting the system is appropriate, neither operation should be performed indiscriminantly. While it is generally not something to worry about, there is a degree of hardware fatigue associated with turning a computer system off and on again, and it is often better to let it run 24 hours a day than to shut it down at night.

REBOOTING.\ There are only four common situations in which rebooting the system is called for:

If you make changes to any of the system software or configuration files that are examined or executed only when the system is booted, you must reboot for these changes to take effect.

Some devices, especially printer and modem ports, can become confused enough that resetting them is only accomplished by re-initializing the system.

If the system has been up and running constantly for over a week, it is wise to bring the system down to single-user mode and run fsck. If any fixes are made to the root partition, the system must be rebooted.

If the system console becomes irretrievably hung, the system must be rebooted.

SHUTTING THE SYSTEM DOWN.\ There are two proper ways to shut down the operating system: shutdown and reboot. As a last resort, the system can be shut down by turning off the power to the CPU. This method is recommended only under emergency conditions because of its detrimental impact on system files and certain types of hard disk drives. These disk drives expect their floating heads to be parked prior to shutdown. Powering off the system could cause the heads to crash and cause irreparable damage to the disk.

Shutdown.\ This command is the most often used method of initiating a orderly system shutdown. It is the safest, most considerate, and most thorough to initiate a halt, reboot, or return to single-user mode. The command will send messages to each user's terminal at progressively shorter intervals as the time for shutdown approaches. The messages tell the time of the shutdown.

Reboot.\ This command terminates all currently executing processes except those essential to the system, then halts or reboots the system. When invoked without arguments, reboot syncs all disks before rebooting the system. The command does not send a message out to the users, unless you use the message option.

MONITOR

Some people would ask, "Why do I have to expend energy on monitoring the network when I could be doing something more productive, like file server or workstation maintenance?" There are several reasons why you should monitor your network:

l l l

To maintain a history of the performance of your system. Studying this history could point out potential failures long before they occur.

To provide a statistical basis for new equipment requests. Management is more likely to purchase new equipment if you can demonstrate that the current equipment will not meet the company's needs.

To enable you to tune your network for optimum performance. This is especially true on larger networks with more than one file server. In some cases, you can provide a perceived increase in throughput by simply transferring tasks from one server to another.

Various network operating systems (NOSs) have their own utility programs to monitor what processing is taking place on their network. You can use these programs to monitor the status of your network, and some utilities give you the capability to monitor a particular job request.