Re: [kosuke@rustybear.com] Re: can't kill a PID

[Russell <rjshaw@iprimus.com.au>]

Oliver Elphick wrote:
> 
> On Sun, 2002-11-03 at 12:23, Rupert wrote:
> > > > Kevin Coyner told:
> > > > > I always thought that with 'kill -9 PID' you could clean up just about
> > > > > any process, but I've run into one that just won't go ...
> > > > >
> > > > > sakura:~$ ps aux |grep xmms
> > > > > kosuke    9026  0.0  0.9 14460 4932 ? D  00:16   0:00 xmms
> > > > > kosuke    9027  0.0  0.0     0    0 ? Z  00:16   0:00 [xmms <defunct>]
> > > > >
> > > > > I've tried 'kill -9 9026 9027', but every time I go back and ps/grep it,
> > > > > it's still there.  And in the meantime, if I try to start a new xmms, it
> > > > > will start a new PID in addition to 9026, but the program itself won't
> > > > > show up.
> > > > >
> > > > > Brainwashed from too many early years in the MS world, I'm tempted to
> > > > > reboot.  But hoping there's a better, Linux way to clean this up.
> > > >
> > > > killall -9 xmms
> > > >
> > >
> > > Just tried that, and the monster continues to live .....
> > >
> > > Kevin
> >
> > I've done something along the lines of
> >
> >   while killall -9 xmms; echo -n .; sleep 1; done
> >
> > in the past, with varying success.
> 
> If kill won't work, killall and the like won't; they are merely
> convenient front-ends to kill a collection of process all at once.
> 
> Sometimes a process gets stuck waiting for an event that is never going
> to happen.  xmms (9026), in that ps report, is in state D -
> uninterruptible sleep.  It's probably waiting for some kind of IO
> event.  Because it is uninterruptible, it never wakes up to find that it
> has been killed, so it won't go away, even for a SIGKILL (kill -9).
> Process 9027 is dead, but it won't go away until its parent (9026) wakes
> up and cleans up; until then it is in state Z - zombie.
> 
> I think you will have to reboot the machine.

Try kill -3 or -6 or -11.