Re: [logs] nimda web server logs

From: wolfgangat_private
Date: Sat Jun 15 2002 - 03:01:02 PDT

  • Next message: Sweth Chandramouli: "Re: [logs] Generic Log Message Parsing Tool"

    Sweth Chandramouli wrote:
    > On Fri, Jun 14, 2002 at 12:27:19PM +0200, wolfgangat_private wrote:
    >> No requests for *.cif where seen in this case, and no requests for repair/sam
    >> either.
    
    > 	In which case, what is it about the requests that makes
    > you think they are related to the ones Tina was seeing?  Hmm... I feel
    > like this problem needs to be better defined...
    > 	In my world, a Nimda scan is one with a fairly constant
    > and short time delta between requests to machines on the same network,
    > and whose GET request string matches the following pseudo-grammar:
    
    > nimda:               base_dir traversal_string [ desired_file | command ]
    > base_dir:            /^(scripts|msadc|_mem_bin|_vti_bin|[cd])$/i
    > desired_file:        'Admin.dll'
    >                    | 'httpodbc.dll'
    > command:             command_interpreter command_string
    > command_interpreter: '/winnt/system32/cmd.exe?/c'
    >                    | 'root.exe?/c'
    > command_string:      tftp_command | local_exec
    > local_exec:          'dir'
    
    > 	, where I'm leaving traversal_string and tftp_command
    > undefined for now (both because I'm lazy and because I don't think it
    > will affect the problem definition significantly), and assuming the
    > parser uses forward slashes to separate tokens.
    > 	I think Tina's scans aren't Nimda, then, because they have
    > a different time signature (relatively long irregular pauses between
    > requests, as though someone were thinking about what to do next), and
    > because they have either a different desired_file ('winnt/repair/sam')
    > or a different local_exec ('dir+c:\\*.cif/s/b').
    > 	So, for these other scans that people are reporting: do
    > they match my definition of Nimda above, or not?  If not, how
    > specifically are they different?
    
    What gave me the idea that the requests I am seeing got something to do
    with the requests seen by others was the "galaxy_*"-requests. I agree
    that Tina was probably seeing a manual attack disguised as Nimda, where
    I am seeing probably an automated scan, even if its not an "original"
    Nimda attack. Key differences are apparently:
    - First request is always a GET for /galaxy_XXXXX.XXXX
    - the nimda-ish requests that follow are actually not GET but HEAD
    - there are a few (GET-)requests for /NULL.printer, NULL.ida and NULL.idq
      thrown in as well.
    I put a log of the scan against one of my servers up at
    http://www.jpaves.com/nimdalog.txt if anyone wants to see for himself.
    
    Wolfgang
    
    ---------------------------------------------------------------------
    To unsubscribe, e-mail: loganalysis-unsubscribeat_private
    For additional commands, e-mail: loganalysis-helpat_private
    



    This archive was generated by hypermail 2b30 : Sat Jun 15 2002 - 12:27:32 PDT