Very Computer

I'm running a somewhat-upgraded version of Debian 2.1 with
kernel 2.2.6, and I've noticed this worrying message in my
startup sequence.

After some investigation, I've discovered it's occurring
during /etc/init.d/network, specifically, the "route"
command.

This is my /etc/init.d/network file:

        #! /bin/sh
        ifconfig lo 127.0.0.1
        route add -net 127.0.0.0
        ifconfig eth0 192.168.1.2
        route add -net 192.168.1.0 dev eth0

..and each (yes, I get it twice) of the "route add" lines
produces:

        SIOCADDRT: Invalid argument

I've tried issuing the commands manually as well, and
get the same thing.  However, the routing table
does appear to be being updated correctly.  I can
ping loopback, and my computer works fine on my local
mini-LAN via eth0.

This happens if just the loopback is used, as well -- i.e.,
if I remove the last two lines.  (Actually, this
started before I even added the eth0 device to my PC.)

I know this has been happening ever since I upgraded my
kernel to 2.2.6 -- I'm not sure if it happened before, but
I don't think it did...

It doesn't seem to be a serious error, but it worries me,
and I'd like to fix it.

If anyone could shed some light on this, I'd appreciate it.
Thanks...

-----------------------------------------------------------------
 Alex Taylor                  BA - CIS - University of Guelph

-----------------------------------------------------------------

Hi, I get this error every time I change IP addresses - not often -
but about 4 times now. There is a corruption in a config file
somewhere.

Delete your network bindings / configuration and rebuild it from
scratch. You can do this in X or linuxconf.

Fixes it for me every time. 5.1 was hell with this problem, 5.2 was
much better. 6.0 is too new to tell.

Charles

Charles Wilkins  CNE / MCP / A+
Network Design Consultant
Practical Computer Solutions
http://www.pcscs.com
609-321-1530
609-321-0840 - fax
--

[posted to cols only. It's a setup issue]

You need to add a netmask to each of the route commands.
Newer kernels no longer have a default.

Although it's not a problem you've encountered (yet?),
previous kernels didn't mind not being told which interface
each route applied to, so add a dev XXX to each.

Something like:

  route add -net 127.0.0.0 netmask 255.0.0.0 dev lo
  route add -net 192.168.1.0 netmask 255.255.0.0 dev eth0

should work. The netmask for 127.0.0.0 does work (here, at
least), but I'm not sure if it's *absolutely* correct.

I'm not sure if it's needed, but I've also got a netmask &
broadcast with 'ifconfig eth0 ...'

HTH
--
You're reading this in a newsgroup because that's where I posted it.
That's where I'd like to read your reply too.   DON'T MAIL AND POST!

But, if you really want to reply by email, my reply-to *IS* valid ;)