The Internet Information Server 4.0 Rapid Review Study Guide

Last Updated 2/3/2009 3:42:58 PM

Abstract

In this chapter, you learn how to solve Internet server problems and optimize performance. You will learn how to monitor performance of various functions, including HTTP and FTP sessions, and using Performance Monitor; resolve Web Service and FTP Service problems; analyze performance and identify bottlenecks; identify disk, network, and CPU performance issues; resolve IIS configuration, resource access, and security problems; and resolve setup issues when installing IIS on a Windows NT Server 4.0 computer. The chapter ends with a quiz and suggested activities.

Troubleshooting is a broad category that can include a wide variety of tasks, such as solving performance issues, fixing resource access issues, and monitoring the server to head off potential problems. In this module, we learn how to solve Internet server problems and optimize performance.

Upon completion of this module, you will be able to

Monitor performance of various functions, including HTTP and FTP sessions, using Performance Monitor
Resolve Web Service problems
Resolve FTP Service problems
Analyze performance and identify bottlenecks
Identify disk, network, and CPU performance issues
Resolve IIS configuration, resource access, and security problems
Resolve setup issues when installing IIS on a Windows NT Server 4.0 computer

TROUBLESHOOTING THEORY

Various problems can arise on a computer network. Solving these problems can be as simple as locating an incorrectly configured component, or as complex as monitoring a network for weeks to locate and isolate an intermittent problem.

In general, the steps for troubleshooting a technical problem are

Monitor the problem or gather symptoms
Define the problem
Generate solutions
Test solutions
Document the problem and the solution

In a well-organized network and under ideal circumstances, the first step, monitoring, takes place long before problems arise. Monitoring your network when there is no known problem is often called establishing a baseline. Baselines are important because they represent the “normal” state of your network. You will find it much easier to solve a problem if you know how your network should be behaving.

When you identify a problem (or a symptom of a problem), you should attempt to define it. You can define the problem as a simple statement initially, but that can grow quickly in scope. For example, if a user tells you that she cannot log on to the Web server, you might initially define the problem as “Susan cannot log on to the Web server.” If you attempt to solve the problem at this point, you might take a troubleshooting path that leads you to examine Susan’s computer and the connection between her computer and the Web server. However, if you later find that several other people cannot log on to the Web server, your troubleshooting approach may be quite different. It is important, therefore, to define the problem as thoroughly as possible in the beginning.

Look for the following details when troubleshooting an IIS problem:

What are the observed problems?
What is the scope of the problem? Does it affect a single user/computer, a group of users/computers, or everyone?
How does the situation vary from normal conditions?
Is there any other problem/event occurring?

Such questions help you define the problem, and once you have a problem statement, you can start investigating possible solutions. The idea is to try to locate the source of the problem, then test possible solutions. For instance, if several users cannot log on, you may first look at the server to ensure that there are no configuration errors and that its hardware is powered on and properly connected. However, if only a single user cannot log on, you may check the server to determine whether the user has been denied access or has been limited in some other way. You should also focus on the user’s hardware and his or her connection to the IIS computer.

Once you have found a solution for the problem, be sure to document the problem and its solution for reference, which could be useful if you have to troubleshoot a similar problem in the future.

TOOLS

You can use several tools native to Windows NT to monitor your network and Internet server, establish a baseline, and solve problems. These tools include Task Manager, Performance Monitor, and command-line utilities such as Netstat and Nbtstat. In addition, you can use many of the tools presented earlier in this book, such as Site Server Express, Internet Service Manager, and Index Server Manager to monitor and troubleshoot the Internet server.

Task Manager and Processes

You can use the Task Manager (Taskmgr.exe) to quickly assess your Internet server status, checking memory and CPU utilization on the Performance tab.

On the Task Manager’s Processes tab, you can find a list of processes currently running. As mentioned in previous modules, the Inetinfo.exe process represents the Internet services (FTP and Web). You can also see the Index Server (Cisvc.exe), Certificate Server (Certsrv.exe), and Microsoft Management Console (Mmc.exe) processes (see Figure 13.1).

When the System Idle Process is at or near 100, the computer is at rest. In the CPU column, you can see that the System Idle Process is consuming 90 percent of the processor time, which means that 10 percent of the CPU time is being utilized elsewhere. The figure also shows that 1 percent is going to the System process, but you have to scroll down through the Task Manager to determine which processes are consuming the remaining processor cycles. You can look at the memory used by various processes in the Mem Usage column. If you find that the Internet server has slowed down, it is a good idea to check the Task Manager to see which processes are consuming the resources.

You can also use the Task Manager to adjust the processor priorities of the running processes. There are four basic levels available: low, normal, high, and real-time. The real-time priority is available only to individuals with Administrative access to the server. You cannot adjust the priority of system processes or those related to IIS. However, you can adjust the base system priorities for many of the applications that run on the Internet server. You can even stop a process altogether by clicking the End Process button. If you find that an application is consuming all the processor cycles, you might right-click that application’s name in the Task Manager and check to determine whether its base priority is set too high. If so, reduce the priority of the application.

Note: Windows NT is a multitasking operating system, which means that many processes can run at the same time. However, only one thread per CPU can be handled at any given moment, so NT must prioritize the operation of these processes. The base priority setting for an application affects how soon and how often its threads will be executed. The priority ranges go from 0 - 31, and the highest number is always first to execute. Low priority starts out with a 4, high priority is 13, and normal is 8; real-time is 24. Processes waiting to execute get a priority boost of 1 every processor cycle, so every process eventually has an opportunity to run.

Most applications have a normal (8) base application priority. However, applications running in the foreground have a priority boost of 2 by default, meaning that whichever application you are actively working with has a base priority of 10 by default. Foreground applications receive a performance boost over background applications. To change this behavior,

Double-click the System icon in the Control Panel.
Choose the Performance tab of the System Properties dialog box.
Adjust the Boost slide bar under “Application Performance” and click OK.

Set the slide bar to the far left to give all applications the same base priority; set the slide bar to the middle to give the foreground application a base priority of 9 (see Figure 13.2).

Page: 1, 2, 3, 4, 5

<< Previous Page Next Page >>

Rate this: