unkillable zombies from using gdb and kernel 2.4.15 and higher?

[Gabriel Paubert]

On 10 Dec 2001, Michel [ISO-8859-1] Dänzer wrote:

> On Mon, 2001-12-10 at 21:20, Gabriel Paubert wrote:
> >
> >
> > On 10 Dec 2001, Michel [ISO-8859-1] Dänzer wrote:
> >
> > >
> > > On Mon, 2001-12-10 at 15:11, Kevin Hendricks wrote:
> > >
> > > > So these zombies are ignoring kill -9  signals which I thought was a big
> > > > no-no?
> > >
> > > Zombies are dead, they don't get more dead if you kill them. ;)
> > >
> > > AFAIR zombies have died but are waiting for their return value to be
> > > read out of the process table. I'd like to know myself how that can
> > > happen when their parents are gone...
> >
> > They are reparented to init. One of init tasks is to reap all
> > processes who lose their parents.
>
> So how often does it do that? Can it be beaten into submission? :)

I believe init is more or less looping constantly reaping orphaned
processes, respawning the tty controlling ones (per inittab), and waiting
for signals for runlevel changes (once system startup is over). But check
the source...

It sounds more like a debugging related (ptrace ?) kernel bug.

	Gabriel.


** Sent via the linuxppc-dev mail list. See http://lists.linuxppc.org/