[clug] Linux routing problem - two interfaces on same media segment, one seems isolated

[Robert Edwards]

Hi Steve,

You don't happen to have NetworkManager running on this box, do you?
Cause of all sorts of network ills around here.

Also, check your iptables - some recent distros "helpfully" add
(dynamically on interface up/down) all sorts of bizarre rules.

Cheers,

Bob Edwards.

steve jenkin wrote:
> steve jenkin wrote on 24/8/08 11:14 AM:
>> I've inherited some machines to admin and run into a knotty
>> routing/network problem - and was hoping somebody on the list might have
>> had this problem or know how to fix it.
> 
> Todays' update.
> 
> eth0 on the .15 machine can now see the 172.16.x.x network.
> Huh?! Since when did networks 'self-heal'?
> 
> This is working like I'd expect - but does say 'don't touch'.
> 
> 
> I did a canonical test on the .80 machine.
> (Interfaces on separate segments)
>  - eth0 connected to router/firewall, only device. DHCP 192.168.x.x
>  - eth1 connected to dumb 100M switch. one other device, a laptop (.67)
> 
> Initially, ping yielded 'Host unreachable' messages.
> Adding 'arp' entry resolved this - but no port activity.
> 
> Huh?! This should Just Work - and has always worked for me in the past.
> 
> I'm starting to think Daniel/Alex are right - this could be dodgy
> hardware (network interfaces).
> 
> Or some strangeness in the '/proc/sys/net/ipv4/conf/' :-)
> 
>