Narvi <[EMAIL PROTECTED]> writes:
> [i think this might just as well belong in -questions]
>
> On 10 xxx -1 [EMAIL PROTECTED] wrote:
>
> > I'm 18-year-old newbie UNIX programmer that currently use
> > FreeBSD and is really thankfull of it.I run it on DUAL PII/333.
> >
> > Some days ago my friend tell me that with simple user rights
> > and whit only 1 line of code he could crash my machine. I laught
> > but he did it :(.
> >
> > What he wrote was ' int main(void) {while(1) fork(); }' compiled it
> > and run it. Within a second /kernel said "proc: table is full" and
> > died. I tried this on some other BSD unixes and the result was
> > same. (BTW Minix 2.0 seem unaffected and probably other SVR4
> > variants, because you can limit the number of system processes
> > and system still have resources to work fine(although slow))
> >
>
> And you can do the same with BSD. See limits(1), csh(1), sh(1),
> login.conf(5)
some time ago, I had a similar problem. too many processes forked, power off...
reboot impossible. the cause of this problem was to define nisdomainname w/
activating nis services. so portmap give up thoses processes to log errors
messages because it was trying to contact nis services which was not there.
I take some time to find /etc/login.conf. the question is, why all default
limits are so permissives (unlimited) by default ? as I remember, it took me
some days w/ many boots to find the reason of portmap failure. an idea would
be to add some limit to limit the number of processes forked by a process (at
one time in addition to the number of processes by user which may be relative
to the system wide limit (maxprocperproc=nproc-10). which is something like the
openfiles limit (w/o the system wide reference but which is possible as well,
like maxfilesperproc=nfiles-10).
Cyrille.
--
home:mailto:[EMAIL PROTECTED] Supprimer "%no-spam" pour me repondre.
work:mailto:[EMAIL PROTECTED] Remove "%no-spam" to answer me back.
To Unsubscribe: send mail to [EMAIL PROTECTED]
with "unsubscribe freebsd-hackers" in the body of the message
