Read the Online.log

The online.log is the first place you'll visit. Be sure that you know how (if any) old logfile backups are stored. You'd like to be able to go back to the initial install in your logfiles if possible. Often, a client doesn't see the utility of keeping old logs. If this is the case, suggest they be saved and compressed or saved and archived on tape. Old logfile information can be very useful in tracing problems in the system.

There are several things that you look for in your first inspection of the online.log file. First of all, it's not always called online.log. This log is named in the $ONCONFIG file. For our purposes, though, it'll be online.log. A second thing you're looking for is "smoking guns" that indicate serious errors that have been encountered. These errors will be tagged with a "WARNING" label in the online.log. Any time you find a smoking gun, you need to know exactly what happened. You'll be more interested in recent errors in the logs. I prefer to read the logfile backward, starting at the end and working toward the beginning.

Get your users into the habit of logging errors and problems that they encounter, making sure that they log the exact date and time, along with any error messages they receive. If you have the dates and times, you can then check the online.log for error messages that appeared at that time.

Reading the logfile is only the first step of the tuneup process. Even if you find an obvious smoking gun, don't jump to the conclusion that this is either the only or even the primary problem. Once you have isolated and analyzed any warning messages in the logfile, don't just declare the tuneup project a success. Go through the entire process and at least hit on all the major areas, even if you think you've solved the problem early in the exercise.