Servers provide users with access to shared resources such as email, data and printers. This ability to share resources with centralized control and administration makes servers more attractive to companies than the alternative of peer-to-peer networking, which can become like the "wild West" administratively. One disadvantage to the server approach is evident when the server goes down for some reason, depriving users of resources for a time.
As users put more and more information on servers, and companies cut costs by eliminating many desktop printers in favor of a few shared printers, the role of the server becomes more important. Uptime rules in centralized server environment and users are quick to yell when anything goes wrong. It is critical that server administrators devise procedures that maximize uptime, which results in higher productivity for users.
No matter how good the operational procedures, server administrators must perform routine maintenance on the server occasionally. Software updates must be installed, backups must be performed and any other required maintenance typically demands that the server be down for a time. Planned downtime is a part of maximizing uptime, and is typically scheduled on weekends in order to minimize user impact. While this "down" and "up" of the server is technically a bounce, the term is more typically applied to emergency downtime.
Video of the Day
On occasion a server will experience "emergency downtime." This can happen when a disk drive fails or a power supply burns out or when any of a number of unexpected problems occurs. This can happen regardless of the level of attention to operational procedures. Typically, this type of downtime occurs at the worst possible time, such as during payroll processing. Occasionally, a server will just "lock up," giving no indication of what the problem is, but not entirely going down.
Bouncing the Server
Bouncing the server typically requires a hard power off. This means that the administrator or technician must remove power to the device in a "non-controlled shutdown." This is the "down" part of the bounce. Once the server is completely off, and all activity has ceased, the administrator restarts the server. This is the "up" part of the bounce. If all goes well, the server will come back up to normal operating status. In the best of circumstances the administrator will be able to find log records indicating the reason for the problem. However, most of the time there is no indication of what triggered the need for the server bounce.