Perl for System Administration

Walking or Traversing the Filesystem

By now you are probably itching to get to some practical applications of Perl. We’ll begin by examining the process of “walking the filesystem,” one of the most common system administration tasks associated with filesystems. Typically this entails searching an entire set of directory trees and taking action based on the files or directories we find. Each OS provides a tool for this task. Under Unix it is the find command, under NT and Windows 2000 it is Find Files or Folders or Search For Files or Folders, and in MacOS it is Find File or Sherlock. All of these commands are useful for searching, but they lack the power by themselves to perform arbitrary and complex operations as they encounter their desired search targets. We’re going to see how Perl allows us to write more sophisticated file walking code beginning with the very basics and ratcheting up the complexity as we go on.

To get started, let’s take a common scenario that provides a clear problem for us to solve. In this scenario, we’re a Unix system administrator with overflowing user filesystems and an empty budget. (Unix is being picked on first, but the other operating systems will get their turns in a moment.)

We can’t add more disk space without any money, so we’ve got to make better use of our existing resources. Our first step is to remove all the files on our filesystems that can be eliminated. Under Unix, good candidates for elimination are the core files left around by programs ...