لطفا منتظر باشید ...
Introducing Fault Trees
Troubleshooting is more of an art than a science. Sometimes, you can easily see what the problem is and how to fix it. At other times, that’s not so easy. The degree of difficulty you have in fixing a problem depends on how complex the problem is and how well you know your stuff. Obviously, the better acquainted you are with computers and Linux, the better you are at troubleshooting.Every problem has a solution. Computers are cause-and-effect-based machines. When something breaks or doesn’t work, there’s always a reason. The reason may not be easy to find, but it exists.How do you find the cause? That’s a million-dollar question. Getting a million bucks isn’t easy unless you’re willing to grind your teeth, plot against your fellow contestants for weeks on a remote island, purchase 10 million PowerBall tickets, or — believe it or not — work hard and work smart. Some people are willing to eat rats for the chance or are lucky enough to win the lottery, but most just have to work hard. Oh, well.Working hard is conceptually easy, but how do you work smart? This concept is where the idea of the fault tree comes into play. The fault tree is a conceptual aid that helps you to eliminate all but the real cause of your problem. The fault tree looks like an upside-down tree, where the trunk of the tree represents the fault, or problem. The ends, or leaves, of the branches represent all the possible causes. After that’s done, solving the problem is virtually guaranteed.For example, Figure 18-1 shows part of a fault tree that points out which major subsystems you should examine. To find the solution to a problem, you have to systematically identify what’s working. You work your way to what’s not working and then when you find it, you usually solve your problem. The fault tree simply helps to formalize the process of problem solving.
Figure 18-1: The fault tree.
Here are some possible faults:
The first branch on the left involves problems with the physical connection. Do you have a network adapter? Is the cable connected properly to the adapter? Do you have a break in the cable? If so, you have to fix or replace the cable.
The second branch deals with the network interface configuration. Have you configured the IP address for your Ethernet adapter correctly? If so, is the netmask correct?
The third branch helps you to decide whether the problem exists with the network routing. Can your network packets be directed toward the correct network?
The fault tree helps you to break down any big problem into several simpler ones. By eliminating each simple problem one by one, you should eventually locate the root cause.
Remember The blind leading the blind
Paul’s colleague Ken Hatfield once said, “One of the side benefits from lots of troubleshooting comes from what I call ‘the value of blind alleys.’ Most often in troubleshooting, you go down blind alleys or, in your tree example, the wrong branches of the solution tree. But in doing so, you learn something. In the future, when you encounter a different problem, that previous blind alley may be the road to the solution.” Well said.Here’s an example: Paul recently had a server that was having lots of problems. The /var file system had filled up, which caused some programs to fail. When space on /var was freed up, most of the programs started to do their jobs again. But one program didn’t work. Paul spent a long time trying to figure out why it didn’t work even after the problem was fixed. As it turned out, this particular program’s real problem was that its license had expired. He had not only walked down a blind alley but also bumped into a wall and kept trying to go forward. D’oh!
