Processes that won't die

[Eric Eisenhart]

On Mon, Nov 12, 2001 at 11:59:28AM -0800, E Frank Ball wrote:
> Yesterday I had hung processes on two machines that refused to die.
> Both are using RedHat 6.2.  Machine one had a couple of netscape
> processes left after netscape died.  Restarting netscape resulted in
> another hang and two more processes that wouldn't die.  The other was an
> application we use at work.  machine two (output of 'ps -ef | grep rmb'):
> 
> yellowst  3556     1  2 Nov09 ?        01:07:24 rmb -title YELLOWSTONE_SYSTEM_CO
> yellowst  3581  3556  0 Nov09 ?        00:00:01 [rmbkbdx <defunct>]
> yellowst  3582  3556  0 Nov09 ?        00:00:00 [rmbhil <defunct>]
> yellowst  3583  3556  0 Nov09 ?        00:00:00 [rmbtmr <defunct>]
> yellowst  3741  3556  0 Nov09 ?        00:00:10 [rmb_pipe.3556 <defunct>]
> 
> The nescape processes looked similar.  The top level process had a PPID
> of 1 (init), the child processes were zombies.  kill -9 by either the
> process owner or root had no effect.  Both machines had to be rebooted.
> 
> There must be a better way.  Any advice?

The PPID becoming one just means that the real parent either died or
"disowned" that process.  The only way to kill a zombie is to kill its
parent or wake up its parent.  (Zombies need to be "reaped" by the parent;
but all they are is a process table entry and a return status.)

ps -ef doesn't show the "STAT" field, though; next time run 'ps ax | grep
rmb'.  I have no idea what rmb is, but the most frequent reason I've seen
for kill -9 not working is that a process is in the "D" state
(uninterruptible sleep) which usually means that the process is blocking on
I/O.
-- 
    Eric Eisenhart   Freedom is slavery.      http://eric.eisenhart.com/
 ^  ICQ#: 48217244   Ignorance is strength.   eric-dot-sig at eisenhart.com
/e\ Perl&SQL Coder   War is peace.            IRC Nicks: Falsch Freiheit
---                        -- George Orwell