Re: Hearing the truth??

From: Daniel J Shauver (shauver_at_rush.edu)
Date: 02/19/04

  • Next message: OBrien, Brennan: "RE: Hearing the truth??"
    To: focus-sun@securityfocus.com
    Date: Thu, 19 Feb 2004 14:05:11 -0600
    
    

    > All:
    >
    > I've got an interesting situation at the office I could use some advice
    > on. I'm being asked from a security perspective whether the following
    > statement (made by our Unix admins) could be considered true:
    >
    > "The only way you can delete a user account on a unix environment is to
    > write a series of scripts to eliminate file associations."
    <SNIPPED>
    > We're in a Sun environment. Some thoughts would be appreciated.

    Brennan,

            There have been a number of comments, useful and not, about
    deleting users, and removing files. Dealing specifically with file
    associations, while local unix filesystems could be handled with the
    simple find/rm trick mentioned in another email, that may not work/be
    appropriate if remote filesystems are involved. There may also be
    complicating factors if you're dealing with encrypted filesystems. The
    task could also become more complicated if you don't have a centralized
    authentication scheme, and/or are dealing with inconsisten UIDs across
    the various Unix servers (user a has uid 9 on box a, uid 10 on box b,
    etc).

            While it does seem like a single script could be written to
    handle most of the issues, it's hard to make a definitive statement
    without knowing more about your environment than you're likely to be
    comfortable divulging, even on such a highly reputable security-focused
    list. ;)

            If you extend the question beyond simply file associations, it's
    not uncommon for a user to be defined both locally on a unix system
    (with or without login priveleges) as well as within an application
    resident on a unix system. If your Unix admins are responsible for
    both, and haven't implemented a single-sign-on solution (far easier to
    type than to do), they may also have to map varying loginids/usernames
    on separate servers to a number of different applications. That could,
    in the end, require more than one script, depending on the complexity of
    the environment, and the state of user account management and
    centralized authentication.

    Dan Shauver
    Unix Geek


  • Next message: OBrien, Brennan: "RE: Hearing the truth??"

    Relevant Pages

    • Re: Linux - SAMBA/OpenLDAP/Plone/OpenExchange
      ... There is a drastic difference between Windows environment ... adminstration and Unix (Linux mimics Unix, but its close enough to be ... appropriate configuration file to the new server, ... but what I have done in MS world is make my clients ...
      (alt.os.linux)
    • Re: Absolute path in argv[0]
      ... >>the way things are usually organized in Unix and Unixish systems. ... application for Linux. ... twist that environment into an MSWindowsish environment will just cause ... Lew Pitcher, IT Consultant, Enterprise Data Systems ...
      (comp.os.linux.misc)
    • Re: Absolute path in argv[0]
      ... >>the way things are usually organized in Unix and Unixish systems. ... application for Linux. ... twist that environment into an MSWindowsish environment will just cause ... Lew Pitcher, IT Consultant, Enterprise Data Systems ...
      (comp.unix.programmer)
    • Re: Porting logical name applications to Unix
      ... quite a while since I did any serious work with any Unix so take this ... I looked at the execution environment of a Linux program ... OpenVMS application programmer for Digital). ... Here are the environment variables for my Mac OS X root account: ...
      (comp.os.vms)
    • Re: Learning C with Older books ?.
      ... >> environment is at least vaguely similar to UNIX. ... >> environment had lacked them, the design would probably be different. ... > the Pascal edition of this book led Kernighan to write his famous essay on ... > why Pascal was not his favourite programming language. ...
      (alt.comp.lang.learn.c-cpp)