Hi,
I am learning about linux-programming and wrote this prog:

int main(){ while(1) {fork()} }

I thought it would be fun. At first it was: I ran it as a normal user, tryed to start a program, and it said "out of pids", then somehow the gterm which ran it crashed, and I wasn't able to do a killall or a shutdown or a "su" because I was out of pids Basically I did a ctrl-alt-del and it didn't work either so I had to reset:

!
The ext3 system partition was broken and unreparable => Reinstall. Luckyly I have a separate partition for /home so I am up and running now (didn't loose data).

So my question: a simple user can render a computer unusable? with such a simple prog? Why is so. Shouldn't users have a limited number of pids?

I admit I was a bit disappointed by this "security hole"

SUSE 9.0 btw.

Have you tried ReiserFS or XFS as file system instead?

And this doesn't belong in this forum.

XFS sucks, that is even less tolerant to abrupt shutdowns.

Switch to a different vterm, login as root and kill the user processes?
At least I'd hope the system keeps some PIDs reserved for root...

ctrl+alt+del killed it. It can happen any place. It is journaled, a bad jorunal or corruption caused by an abrupt shutdown, or shutdown with data still in the HD cache can cause that stuff.

enric: I agree it's odd that there is no default PID limits. As said you can put one in, and you should. It's listed in the Gentoo security guide, BTW. Which is a semi-good read even if you don't use Gentoo.
Why there is no default limit is beyond me though. What normal user would be running over, say, 1000 processes?

I've user ReiserFS and had some crashes(couple tests with a bit too unstable machine and some driver problems) but I've never destroyed or harmed the file system.

Is ext3 that unstable? I've killed mine using the power switch a number of times without a problem (of course it has to run it's nice check at the beginning). This is both on gentoo and mandrake.

Sometimes I experienced lost files on ext3 when my system crashed, actually I once lost some init script, so I had to reinstall... With ReiserFS things were much better in that sense, altough performance goes down... I am trying XFS now to see how it is, altough I had no crashes yet...

I've been running linux on ext2 through reiserfs since '97 and never had the filesystem mess up (although that's partly because it never crashes unless I do something truly stupid like a fork bomb "just to see what happens". OTOH, Windows messes up so often (dual boot) that I have the procedure to save & restore windows partitions to a linux dir down cold.

BSOD as per usual, hardware reset necessary
On booting windows "To avoid seeing this error always shut down from start menu" implies it's MY fault, stick more pins in the bill troll.
windows scandisk turns everything into a CHK file
reboot, tell grub to boot linux.
dd bs=512 count=20000 if=/dev/zero of=/dev/hda1 (then hda2 if D: munged)
run the windows install cd
use linux install disk rescue option to dd bs=512 count=1 if=grubimage of=/dev/hda to repair windows rewrite of MBR.
cd /winbak2384/C
cp -r --preserve=timestamps allegro /windows/C
rinse lather repeat...

to me, "doll" implies something attractive... (Hey doll!)

I like real live DOS! I still have a "I (heart) DOS" bumper sticker which nobody understands anymore.:D

No, that's DoS.