OSEC

Neohapsis is currently accepting applications for employment. For more information, please visit our website www.neohapsis.com or email hr@neohapsis.com
 
From: Kurt Seifried (bugtraqseifried.org)
Date: Thu Apr 04 2002 - 15:41:52 CST

  • Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]

    This is, to put it politely, incredibly old news. Let's face it, if you give
    a user a shell acount, with no restrictions on CPU time or memory usage,
    yes, they will be able to suck up as much resources as the computer can
    spare (this is, among other reasons why "nice" exists). I advise you place
    limitson the users, memory, cpu, stack size, file descriptors, etc, finding
    "good" limits can be tricky though, and you will also want to limit
    concurrent logins.

    I wrote an article on using PAM (pluggable Authenticaiton Modules) which
    covers these issues and a few others, available at:

    http://www.samag.com/documents/s=1161/sam0009a/0009a.htm

    Also you can view information on setting limits with various shells, and PAM
    as well at:

    http://seifried.org/security/os/linux/20020324-securing-linux-step-by-step.h
    tml
    goto "Limiting users overview".

    And the LASG, "Limiting and monitoring users"
    http://seifried.org/lasg/users/

    Better to use PAM to limit users then the shell because the various shells
    do not all support the limiting the same items, or soft/hard limits, and if
    you miss a shell and the user "chsh"'s they can avoid it, they can't really
    avoid pam. As for the "/*/../........." problem in general it was
    "discovered" many many years ago (more then two).

    Kurt Seifried, kurtseifried.org
    A15B BEE5 B391 B9AD B0EF
    AEB0 AD63 0B4E AD56 E574
    http://seifried.org/security/
    http://www.iDefense.com/