Overview

Most Web servers, including the ones under discussion here, log accesses in a standard format. (You can generally customize how much information gets logged, and further on I'll discuss the commands for how to do that. If you're considering a customized log-file format, though, you'd better have a plan for what you're going to do with it; many log-file analysis tools understand only the standard formats.)

A typical log entry will have the IP address of the originator, the URL requested, and the status returned. Counting the 200 status lines gives you a rough indication of how many hits you had. However, each page you serve will probably result in multiple entries in the log file one for the actual page and one each for any images linked in the page.

Apache and WASD offer cookie-based user-tracking features that generate a unique user ID, which you can use to track "visits" rather than hits (this won't work if cookies aren't enabled on the browser). (Some Perl modules exist for session tracking that fall back to URL-encoded IDs if cookies aren't enabled, but they're not core browser functionality so I won't document them here.) In addition to access logs, your server can generate error logs, which are useful in identifying and solving problems.

12.1 Customizing your logging

You can locate your log files where you want them. (In a high-performance configuration you'll want to put them on a different disk and ideally on a controller different from either your Web content...