Hmmm well that script does it job but it apparently doesn't lessen HFS' CPU load or help responsiveness at all nor does it remove the GET lines from the log, only the 503 reponses.
This might need to be coded into the software to work right.
BTW, based on my log and logging the user-agents, it seems as if IE8 and IE9 are the only ones that automatically try to open up a ton of download connections at once, every other browser does it normally. I think it might only be that it does it when loading files embedded into web-pages though (like JS, CSS, etc.)
Just a side note, today when I walked over to my computer I looked and saw that there were 55 download icons in the system tray and that HFS would not respond.... I went into Firefox and refreshed my page and it said it couldn't connect. Sound like an overload crash?
Perhaps I should limit the amount of hot-linking each user gets (my ratio of downloads to user accounts is well in excess of 300:1)
EDIT:
It seems as if the extreme connection acceleration in IE only happens when trying to download a certain SWF files...
EDIT 2:
I think the problem might just be that the user embedded the SWF file into their page multiple times... I'm gonna go investigate their code.
I guess my main problem here is not how annoying the log is but that since HFS takes the time to respond to the repetitive connections and execute the macros under [request] for them rather than just disconnecting without reply, it takes up a lot of CPU time and slows down the server, a lot.
What tpl are you using?
Not that it's relevant but I actually use several different templates based on section... the Global Template is >200KB in size
VFS, Events, all Templates, TXT Files in use by HFS, etc., amounts to over 700KB