Cisco: A Beginners Guide, Fourth Edition

OVERVIEW

Keeping an internetwork going is a full-time job. As you saw earlier, problems emerge with such frequency that the industry invented routing protocols to deal with them automatically, not even waiting for a network administrator to intervene. They do a pretty good job, at least with problems that can be ameliorated by detouring to a new route. But changing routes is only a temporary solution. In order for a network to run effectively, all its components must be running properly and constantly. After all, delivering available bandwidth to users under normal circumstances is hard enough without having a LAN segment out of commission or a router functioning at partial capacity.

For this reason, a big part of a network team's time is spent troubleshooting. Problems range from a single user unable to access a service to an entire LAN segment crashing. Troubleshooting isn't just a matter of finding and fixing broken parts; much of it is dedicated to fixing performance bottlenecks. When a problem emerges, the network administrator often has no idea which device is causing the trouble. And once the problem device is identified, the cause of the problem must be diagnosed. Then decisions must be made on how to fix the situation.

A methodical approach should be taken to troubleshooting; otherwise, a lot of time can be wasted trying to figure out what's causing the problem. Like a doctor, the troubleshooter must recognize the symptoms, associate them with a set of probable causes, and then progressively narrow down the list until the culprit is finally identified. From there, a proper action plan must be devised and implemented. That's troubleshooting.

In this chapter, we'll review how to troubleshoot problems in a variety of Cisco configurations by running through some troubleshooting scenarios. For simplicity's sake, we'll assume IP as the network protocol and the Microsoft Windows platform as the host. Although the terminology can vary, networking problems are largely the same, regardless of the protocol or host environment. We'll also restrict the examples to troubleshooting routers, which is where most of the action is.