networking stop hangs trying to contact portmapper
Hi,
I am observing problems shutting down networking
on two recently installed Debian/Woody systems:
a reboot command first stops the portmapper,
then tries to stop networking and just hangs
indefinitely.
When trying to debug the problem I tried
"/etc/init.d/networking stop",
"ifdown eth0" and finally
"strace -f ifdown eth0" and found that it
repeatedly sends data to port 111 on 127.0.0.1
via sendto(). Aborting with CTRL-C is possible.
Is this the correct behaviour? The portmapper
has been stopped already, which (as far as I
can tell) is correct.
However, starting portmapper and networking
seems to be done in different orders, depending
on the initial run level. Or am I reading /etc/rc?.d
incorrectly:
$ ls /etc/rc?.d/*networking* /etc/rc?.d/*portmap*
/etc/rc0.d/S10portmap
/etc/rc0.d/S35networking
/etc/rcS.d/S40networking
/etc/rcS.d/S41portmap
/etc/rc6.d/S10portmap
/etc/rc6.d/S35networking
Both machines have a pretty ordinary
/etc/networking/interfaces:
auto lo
iface lo inet loopback
auto eth0
iface eth0 inet static
address <..>
netmask <..>
gateway <..>
One machine runs kernel 2.4.16 (as it doesn't work
with 2.2.x), the other can use both, but I think
the problem was independent of the kernel version.
Any idea? I don't know if this is a configuration
problem, a bug, and (if it is a bug) in which package.
--
Freundliche Gruesse / Best Regards
Patrick Ohly
Software Engineer
--------------------------------------------------------------------
//// pallas
Pallas GmbH / Hermuelheimer Str. 10 / 50321 Bruehl / Germany
Patrick.Ohly@pallas.com / www.pallas.com
Tel +49-2232-1896-30 / Fax +49-2232-1896-29
--------------------------------------------------------------------
Reply to: