home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!enterpoop.mit.edu!usc!nic.csu.net!nic.csu.net!nntp
- Newsgroups: comp.sys.sun.admin
- Subject: Re: Bringing a Sun to it's knees
- Message-ID: <1993Jan26.164557.3940@nic.csu.net>
- From: oleg@gd.cs.csufresno.edu
- Date: 26 Jan 93 16:45:55 PST
- Sender: oleg@gd.cs.csufresno.edu
- References: <T6AXBYAV@cc.swarthmore.edu> <1993Jan25.221022.28758@ra.msstate.edu>
- Organization: Computer Science Departement of California State University in Fresno
- Nntp-Posting-Host: gd.cs.csufresno.edu
- Lines: 24
-
- In article <1993Jan25.221022.28758@ra.msstate.edu> fwp@CC.MsState.Edu (Frank Peters) writes:
- >In article <T6AXBYAV@cc.swarthmore.edu> eoliver@ralph.cs.haverford.edu (Erik Oliver) says:
- >: During our Operating Systems class, we were being asked to see for
- >: ourselves that "modern" operating systems place limits on the number of
- >: processes an individual user can run thus preventing over load.
- >
- >Interesting. Who came up with that theory? It isn't true and in many
- >cases it isn't desirable.
- >
-
- SunOS limits a normal user to 25 processes or so. This usually prevents a runaway program
- from hanging the system, but won't stop one from crashing the system if that's what you
- want. First, this doesn't apply to setuid processes, like many system commands. Also, process
- table is not the only finute resource. Try allocate ~32M array in your program and write
- bytes in random positions.
-
- I have yet to see the system that survives this:
-
- $ while :
- do
- /usr/etc/ping -s localhost >/dev/null 2>&1 &
- done
-
- Oleg
-