[SunRescue] Killing Processes [Was: Rescue digest, Vol 1 #773 - 16 msgs]

[Mike Hebel]

Actually no flames are necessary from me.  You've just explained to me why I
can't remove a passal of Zombie processes from our server here at work.  Now
I just have to find out what's not responding that keeps the Zombies from
shutting down.  Thanks!

Mike Hebel

-----Original Message-----
From: rescue-admin at sunhelp.org [mailto:rescue-admin at sunhelp.org]On
Behalf Of Ed Pate
Sent: Wednesday, December 06, 2000 10:26 AM
To: rescue at sunhelp.org
Subject: [SunRescue] Killing Processes [Was: Rescue digest, Vol 1 #773 -
16 msgs]


> --__--__--
>
> Message: 7
> From: "Eric Ozrelic" <erico at bendcable.com>
> To: <rescue at sunhelp.org>
> Date: Tue, 5 Dec 2000 01:23:42 -0800
> Organization: BendCable Communications Inc.
> Subject: [SunRescue] Killing Processes
> Reply-To: rescue at sunhelp.org
>
> I know I can use "kill <pid>" to kill a process, but how do I absolutly
> destroy a process, when kill doesn't work?
>
> Regards,
>
> Eric Ozrelic
>
> --__--__--

You can use kill to send any signal to a process.  "kill <pid>" sends signal
15 which
tells the process to "settle its affairs" and then kill itself.  As many
have replied,
you can use "kill -9 <pid>" to send a signal 9 which tells the process to
kill itself
immediately, don't close any files, don't flush buffers, don't do anything
except die.
The only time this will not work is if the process is "wait"ing on hardware
or some
kernel routine that hasn't (or will never) respond, such as a dead harddrive
or an NFS
request to a remote server that is down.  The process never wakes back up to
process the
signal.  There are only two ways to get rid of it then, 1) reboot, and 2)
twiddle with
kernel data structures.  I whole heartedly recommend the first option.

Alright, alright, I know I over-explain.  Flames are welcome OFF-LIST.

Ed