Chapter 21. Server Testing and Maintenance

To get the most out of your servers, you need to know how well they are operating, what is "normal" behavior, and what to change when your server(s) don't operate normally. Part of good management practice is the collection of performance data because with that data you can better tune your servers, plan for upgrades, and know when to upgrade your systems.

This chapter describes some of the tools and techniques you can use to test and monitor your servers. Modern operating systems and server applications provide some powerful tools that you can use to help in this area. Performance monitoring is accomplished by using small code snippets called counters that trap and measure when system events occur. Windows, UNIX, NetWare, and other server operating systems come with a large array of counters for measuring the resource consumption of key system components, such as processors, memory, disk space, and more. Beyond simple consumption, you can also get information about the efficiency of your system, speeds, throughputs, and more. When you install a server application, it may install additional counters for you to use. You can track the results of your performance measurements by using performance monitors or by executing a number of commands from the command line.

The second area of server testing is troubleshooting, and one of the primary tools for troubleshooting is the collection and analysis of system events. In an event-driven programming environment, a system is constantly reacting to events, and a typical system may generate hundreds, if not thousands, of events each second. Many events, such as screen redraws, may not be particularly significant, but an event such as a user logon is a valuable event to record. Systems store a number of different logs you can use, including application, security, and system logs. This chapter looks at events and event logs and discusses how to get more information about logs as well as how to determine which events are significant. You can build different redundancies into your servers to get different levels of fault tolerance. By paying attention to good backup strategies and building redundancy into your servers, you can give your systems a fault tolerance that will help you survive any disaster that could conceivably happen.

Finally, this chapter presents the topic of disaster recovery. Disaster recovery involves planning and preparation, coupled with having an understanding of the kinds of things that can go wrong.