On Sunday 18 April 2004 17:39, Hans du Plooy wrote:
I'd still like to know how to kill a process that normally resists kill -9
You can't. Sending SIGKILL to a process is the most violent method at your disposal. If the kernel ignores that, usually because the process is inside a kernel syscall, usually waiting for some sort of I/O, there's nothing you can do about it. Having said that, sometimes it is possible to get lucky. For example, once I had a problem with a scanner, the scanning program hung hard while waiting for the scanner, and nothing happened. The scanner didn't work, and the process just hung in state D forever. Cycling the power on the scanner however triggered some initialisation or something inside the scanner so the driver reset it self and the scanning process died. If you don't have a way to trigger some interrupt to reset things, the process is usually there until you power down the machine