Nice tip, I didn’t know about it.
Abetter way is to add the kernel parameter “panic=10″ (from lilo, grup, extlinux,…), so also earlier panics are rebooted.
I use it frequently, along with lilo -R / extlinux –once, for new kernels in remote machines.
It is good except if the problem lies in some piece of c^w hardware that doesn’t knows how to reset itslef. That happened to me with a E1 (telephony) card. Solution: bycicle the 2 km up hill up to the server, shutdown the machine and start it again.
Of course, I bought another card.
Comments are closed.