Re: Possible DoS attack to NT boxes running OpenNT 2.1

From: n3m0 (n3m0at_private)
Date: Sat Aug 15 1998 - 09:35:13 PDT

  • Next message: Alan Brown: "Re: MySQL DoS ?"

    First of all I must say that OpenNT it's a wonderful product. It works
    really fine and it really surprise me everyday. My posting here was ONLY
    intended to aware people for something I found and test, not to shoot a
    product.
    
    
    > There's two things wrong with this. First, it's hardly a DoS
    > attack when you
    > had to authenticate yourself to the system to make the attack. If an admin
    > saw several dozen instances of a Win32 app belonging to user Nemo, said
    > admin could simply call up Nemo and yell at him for sucking up memory.
    > There's no anonymous attack here; no username/password, no access.
    
    
    That's true. This is not a DoS attack on a traditional way. I mean, it's not
    like 'teardrop', 'nestea' or whatever. But it could be a problem for those
    systems offering anonymous or guest telnet access: a guest user could log
    into the system and hang it.
    
    You are also right when you say that I, the sysadmin, can face a registered
    user who is trying to kill my system. But, anyway there's a lack of inner
    security and it's also possible for a user to hang the computer before being
    caught.
    
    
    > Second, the Win32 GUI app is running just fine, in a non-displayed Windows
    > Station. It is consuming some resources, but mostly swap space;
    > no CPU time,
    > once the app has started up and is waiting for user input. A user with
    > appropriate privileges (say, Administrator) should be able to use
    > TKILL.EXE
    > or the Task Manager or any other appropriate utility to shoot the
    > non-visible GUI app. Certainly, Nemo could log back on via telnet
    > and shoot
    > his own non-visible GUI app via tkill.
    
    
    I'm sorry but I can't agree with this. I am the system administrator and I
    have tested it thoroughly before I send my first post and I have tested
    again before sending this new one. I have tried the experiment from accounts
    with different access rights, even administrative ones, and NO ONE on the
    system (Administrators included) could kill the process. They seem to be
    "protected" system tasks. They may inherit this property from its parent
    POSIX processes.
    
    I couldn't find any file called TKILL.EXE, so I tryed to kill them trough
    the Task Manager and the kill command, but none of them were able to free
    the resources.
    
    You say there's no CPU use... I must say this is not what I have suffer.
    Sorry, but there IS CPU hogging. Its use rises to 100% and kernel activity
    rises to 50% forever. Finally the foreground work turns horrible and the
    operation turns impossible.
    
    {Nemo}
    
    ---------------------------------------
    Nemo - n3m0at_private
    
    BlackBrains Security Team member
    http://www.thepentagon.com/blackbrains/
    http://blackbrains.onlinet.com
    ---------------------------------------
    



    This archive was generated by hypermail 2b30 : Fri Apr 13 2001 - 14:12:45 PDT