I have a tool that parses through the log files of my server. The output shows many errors and especially broken links(404 Errors). Unfortunatelly, this tool does not show where these broken links come from, ie where the anchor text of these broken links is. So i used Xenu's tool (very helpful) which returned the broken links and where they are linked from. I cleared all the broken links, still the first tool that processes the log files shows loads of 404 errors. How can I find where these 404s come from?
You can turn referer tracking in most logging applications and that will tell you something. However, just because a url was requested, that doesn't mean that it came from a link somewhere on your site. It could be a type-in with a typo, for example.
There is a handy, free application called Xenu Link Sleuth - it will spider your site and give you a report on where bad links occur on your site. This is better than using logfiles, because you don't have the ability to fix every 404 that you see in the server logs.
Tedster is correct (as always!) I use Xenu for a huge site... It works great, and it's free. I'm considering sending them $300 for the registered edition 1) just to support their continued efforts and 2) so that I can use command line parameters and call Xenu from a scheduled job. The free version has to be run interactively, and it runs for several hours on our site.