Network Warrior

Chapter 36. Failures

Outright failures can often be detected easily; when a router fails completely, there are usually some pretty obvious symptoms. When a WAN Interface Card (WIC) starts mangling packets, the problem can be a little harder to diagnose. In this chapter, I'll cover some examples of what can cause failures, and discuss how to troubleshoot them effectively.

Human Error

Human error can be one of the hardest problems to track, and, once discovered, may be almost impossible to prove. Getting people to own up to mistakes they've made can be a troublesome task—especially if the person responsible is you!

Once, I was working on a global network, administering changes to the access lists that allowed or denied SNMP traffic to the routers themselves. We'd just added a new network management server, and we needed to add its address to all the routers so they could be monitored and managed from the central office.

Instead of properly writing the configurations ahead of time, testing them in a lab, and then deploying the proven changes during a change-control window, I decided that because I was so smart, I would apply the changes on the fly. The changes were minuscule—just one line—what could go wrong?

Naturally, I bungled something in one of the routers, and ended up removing an active ACL on the inbound interface. The router was the sole means of entry into the entire continent of Australia for this company. Because nothing simple ever happens to me, the disaster struck ...