Red Hat

Hi there,

Anyone there who has a procedure on how to crash a linux a system? We need to do some testing on our machines. And one of the test would be to crash the os. Appreciate you help. Thanks.



Best regards,
Sodapop

What exactly you mean with "crash the OS" - how, which part of the system has to be crashed ? What is the OS ? What is the purpose, the final aim ? To see how the system will behave in extra ordinary situations ?

Thanks for the reply.

The OS is Linux. Distro is RedHat. Purpose of the crash is to see if analyze the errors that will come out after. The aim is to have the server pass an operational acceptance test. What we intend to do is simulate a kernel panic and see if the server will boot after the panic. Thanks again.

Again, this sounds like a high level requirement - if I have the design document and acceptance criteria, then I can write test case / plan. This task, however, will require coding skills against kernel, where I'm not good at, and I'm afraid I can't help, but in our forums there are lots of people who can help, I believe. So, if you post a particular requirement, someone else might be able to help. At first glance, I'd think of "fork bomb" search the Google for it.

Hi,

You may try the below two commands to crash the OS. I hope they will work.

sysctl -w kernel.sysrq=1
echo c > /proc/sysrq-trigger

Regards,
A.

Write a dumb kernel module, something like the following should kill it:

Code:

MODULE_LICENSE("GPL");
MODULE_AUTHOR("blah");
MODULE_DESCRIPTION("kernel killer");

static int __init killer_init(void)
{
	void *ptr = NULL;
	printk("time to die %d\n", (int)ptr);
} 
module_init(killer_init);