On Tue, 13 Apr 1999, Matthew Hixson wrote:
> On Tue, 13 Apr 1999, Andy Poling wrote:
> >
> > You guys are *scaring* me. You don't kill the kernel because of
> > mis-behaving processes. Zombie processes are waiting for the parent process
> > to reap them (using wait())... or init will when the parent exits.
> >
> > Kill the problem processes... not the kernel.
>
> Yes, init usually cleans things up, but I know that I have seen a zombie process
> that would not die. Unfortunately I don't remember what program that was.
There've been lots of them over the years and operating systems. SunOS
used to make terminal zombies fairly regularly. With linux, right now I
often end up with
bin 10484 0.0 0.0 0 0 ? Z Apr 1 0:00 (rpc.portmap <zombie>)
bin 10485 0.0 0.0 0 0 ? Z Apr 1 0:00 (rpc.portmap <zombie>)
bin 16591 0.0 0.0 0 0 ? Z Apr 6 0:00 (rpc.portmap <zombie>)
that seem to accumulate gradually if my system has been up a long time.
My original portmapper is still running, so clearly it has somehow lost
track of these. These zombies will of course ignore attempts to kill
them since they're already dead. Now, if I killed and restarted the
portmapper I imagine that init would reap them -- yup, it does -- but
this is something one usually doesn't do. Some zombies that "won't die"
simply belong to a parent that won't reap them and that one doesn't
really want to kill.
The processes on the old Suns were a bit different. IIRC they already
belonged to init and would never go away as killing init is a synonym
for rebooting. They'd accumulate up to a reboot, eating a tiny bit of
resources I expect (a PID if nothing else).
rgb
Robert G. Brown http://www.phy.duke.edu/~rgb/
Duke University Dept. of Physics, Box 90305
Durham, N.C. 27708-0305
Phone: 1-919-660-2567 Fax: 919-660-2525 email:[EMAIL PROTECTED]
-
Linux SMP list: FIRST see FAQ at http://www.irisa.fr/prive/mentre/smp-faq/
To Unsubscribe: send "unsubscribe linux-smp" to [EMAIL PROTECTED]
