atom feed10 messages in org.freebsd.freebsd-questionsRe: Killing zombie processes
FromSent OnAttachments
Dean HollisterMay 26, 1998 5:06 am 
Dan NelsonMay 26, 1998 8:48 am 
Dean HollisterMay 26, 1998 9:00 am 
Dan NelsonMay 26, 1998 9:30 am 
Dean HollisterMay 26, 1998 9:37 am 
Dan NelsonMay 26, 1998 3:38 pm 
Doug WhiteMay 26, 1998 9:15 pm 
Dean HollisterMay 27, 1998 12:12 am 
Jason NordwickMay 27, 1998 9:09 am 
Doug WhiteMay 27, 1998 1:58 pm 
Subject:Re: Killing zombie processes
From:Dan Nelson (dnel@emsphone.com)
Date:May 26, 1998 9:30:04 am
List:org.freebsd.freebsd-questions

In the last episode (May 27), Dean Hollister said:

On Tue, 26 May 1998, Dan Nelson wrote:

When you say "killing the parent doesn't work", do you mean that killing the parent did NOT remove the child's pid, or you couldn't even kill the parent at all? What does a 'ps axl' print for the processes?

I could not kill the parent at all.

bash$ ps -t /dev/ttyv0 PID TT STAT TIME COMMAND 4803 v0- IEs+ 0:00.00 (sh) 4994 v0- Z+ 0:00.00 (tail)

See? Here, the sh process will not exit, because of the zombie. It is trying to exit, but can't.

It's definitely trying to exit, but the zombie probably isn't stopping it. Zombies are processes that have exited completely, but the parent hasn't checked the returnvalue for them yet. i.e. the tail command has finished, and the only remnant of that process is the slot in the process table (which will be removed once the parent exits or does a wait() ). The only resource a zombie takes up is one process table entry.

Do a "ps axl -t v0", and see what's under the "WCHAN" column for pid 4803. That's the kernel event the process is waiting for. If it says "ttywai", it's trying to write something to the console (make sure scroll-lock isn't on).

-Dan Nelson dnel@emsphone.com

To Unsubscribe: send mail to majo@FreeBSD.org with "unsubscribe freebsd-questions" in the body of the message