strange msg lines..

Ian Smith smithi at nimnet.asn.au
Thu Nov 17 18:02:19 GMT 2005


Re: freebsd-questions Digest, Vol 113, Issue 12
 > Message: 28
 > Date: Wed, 16 Nov 2005 23:56:06 +0100
 > From: dick hoogendijk <dick at nagual.st>

 > I get a lot of these rules in my log file lately. Don't know why they
 > are not logged in the error.log file. And if they are harmful or not.
 > 
 > ==--==
 > 83.30.48.99 - - [16/Nov/2005:23:44:18 +0100] "GET / HTTP/1.1" 200
 > 1860 "http://puttane-grandi-tette.com" "Mozilla/4.0 (compatible; MSIE
 > 6.0b; Windows NT 5.0; .NET CLR 1.0.2914)" 

 > 85.106.229.37 - - [16/Nov/2005:23:44:24 +0100] "GET / HTTP/1.1" 200
 > 1860 "http://hosting-siti-adulti.com" "Mozilla/4.0 (compatible; MSIE
 > 6.0b; Windows NT 5.0; .NET CLR 1.0.2914)" 

[.. etc ..]

I've seen Dinesh's reply, and your later response, but I've seen batches
of these at various times too, and think it's something other than DNS
(though it did look like maybe a test of a distributed fetch, many IPs)

 > These are not normal requests to my apache server. But it seems to
 > "listen" to them. Am I 'in danger?'

No, they're not errors, they're just requests for your home page (GET /) 
which is presumably 1860 bytes .. the Referer (sic) in each case is one
of these apparent porn sites, but could easily be forged - it's unlikely
that the pages at the URLs given do in fact have any link to your site;
more than likely they want you go check out their stuff looking for one! 

So there's no danger involved, unless there are enough of them to DoS
your server.  I tend to deal with such as these by blocking them in
apache|httpd.conf so they just get a 403 access denied response, eg:

 # 1/3/5 multiple browsers, multiple IPs, all the same referrer:
 SetEnvIfNoCase Referer buy-vicodin-online\.us go_away

or in this other case, various different GETs attempting to access
various porn URLs as wannabe proxy requests, all from the one IP:

 # 6/10/5 porn link referers regularly, different browsers ..
 SetEnvIf Remote_Addr 209\.172\.35\.44 go_away

In your case, the browser identification, most likely bogus, is a common
factor in each, and could be blocked with such as: 

BrowserMatch "Mozilla/4\.0 \(compatible; MSIE 6\.0b; Windows NT 5\.0; \.NET CLR 1\.0\.2914\)" go_away

or some unique part of that string.  whereas others as above will cycle
through different browser strings - there's usually some common thread
to such bot-made requests.  I only hit on them when they become annoying
(but sometimes I'm easily annoyed :)

Then of course you'd need something along the lines of:

<Directory "/usr/local/www/data">
  [.. other stuff ..]
  # 18Mar02 - allow only this file to otherwise denied bots
  <Files "robots.txt">
    order allow,deny
    allow from all
  </Files>
  order allow,deny
  allow from all
  deny from env=go_away
</Directory>

Cheers, Ian



More information about the freebsd-questions mailing list