Kernel memory disable overcommitment - java and other things dysfunctional

uname -a
Linux Laptop 2.6.24-19-generic #1 SMP Wed Aug 20 22:56:21 UTC 2008 i686 GNU/Linux

java -version
java version "1.6.0_06"
Java(TM) SE Runtime Environment (build 1.6.0_06-b02)
Java HotSpot(TM) Server VM (build 10.0-b22, mixed mode)

sysctl vm.overcommit_memory=2
(if you're unlucky fork will fail right from there on! so don't try this when you got importand files open)

then start java (or anything else..)

java
Error occurred during initialization of VM
Could not reserve enough space for object heap
#
# An unexpected error has been detected by Java Runtime Environment:
#
# SIGSEGV (0xb) at pc=0xb7dee4da, pid=7477, tid=3084225424
#
# Java VM: Java HotSpot(TM) Server VM (10.0-b22 mixed mode linux-x86)
# Problematic frame:
# C [libc.so.6+0x6f4da] cfree+0x7a
#
# An error report file with more information is saved as:
# /home/ubuntu/hs_err_pid7477.log
#
# If you would like to submit a bug report, please visit:
# http://java.sun.com/webapps/bugreport/crash.jsp
#
Aborted

This is just a java example. Other applications segfault. The problem itself is the kernel memory overcommitment and the resulting oom. To disable it or to trying out a workaround.. just makes things worse. Maybe the vm.overcommit_memory was a bad advice i read somewhere.

There is a lot of memory free. So this error is just.. i see no reason for this. Is the kernel memory management so dependent on the bad behavior of overcommiting memory?

To return to the default:

sysctl vm.overcommit_memory=0

and java works again.

Further reading to show that this seems to be a well known problem for years - still not fixed. There they just ask for the 'kernel version' but it's there since all those versions.
http://www.redhat.com/archives/rhl-devel-list/2005-February/msg00671.html

I got enabled swap. I got enough RAM. So all those factors are allright. Still the kernel faults with crappy memory management from time to time.

Some more information rich reading about kernel memory:
http://www.win.tue.nl/~aeb/linux/lk/lk-9.html

Another thought, why this should be a concern for the maintainers. This results in random - not by the application caused - crashes. I guess some of the error reports are the result of the kernel when it gets to memory limits in one library or in any application. I have seen many errors that only seem to happen, if the memory gets 'closer' to the limits. An ubuntu with only one application running is more likely to work than a ubuntu with lots of stuff running.