[Sussex] Anyone into bondage?

[Brendan Whelan]

Jon,

Thanks for the info.

On the connectivity, I mean that if the cable is moved from eth0 (Onboard
NIC) to eth1 (Offboard NIC) then I cannot connect.

I am using Fedora core 3.

Using "ifconfig eth0 up" and "ifconfig eth1 up"made no difference.

When the cable is connected to the onboard NIC:
ethtool eth0|grep "Link" gave Link detected yes
ethtool eth1|grep "Link" gave Link detected no

When the cable is connected to the offboard NIC:
ethtool eth0|grep "Link" gave Link detected no
ethtool eth1|grep "Link" gave Link detected no - the cable is obviously
physically connected as one of the leds is on and another is flashing -
presumably because the other end of the link is transmitting.

Brendan
----- Original Message -----
From: "Jon Fautley" <jfautley at redhat.com>
To: "LUG email list for the Sussex Counties" <sussex at mailman.lug.org.uk>
Sent: Thursday, September 29, 2005 3:29 PM
Subject: Re: [Sussex] Anyone into bondage?


> Brendan Whelan wrote:
> > Jon,
> >
> > Thanks for the response..............
> > The problem is that, when I try to use bonding, I can connect to the
> > Ethernet socket on the motherboard but not to the one on a separate
card.
> > (When not bonded, if I configure the two Ethernet ports with different
IP
> > addresses then they both work.) Below are the various files and system
> > information. I have executed:
> > service network restart
> > ifup eth0
> > ifup eth1
> >
> [SNIP]
>
> I'm not quite sure what you mean by this? When using bonding, the
> seperate physical interfaces 'go away' and can't be used individually.
> All network configuration or usage needs to go through the bondX device
> (bond0 in your case).
>
> Do you mean that when you remove one of the network cables, it's not
> working? I.e. you pull the cable from the onboard NIC and it doesn't
> continue to use the offboard NIC?
>
> Try running 'ifconfig eth0 up; ifconfig eth1 up' and see if this makes a
> difference. I don't think it will do, but it's worth a try ;)
>
> What mode are you running the bonding in? Are you using the default? (if
> so, what Distro).
>
> it's worth checking that the system realises when you remove the cable
> too, run ethtool eth0 | grep "Link" and see what it reports, both before
> and after the cable pull. Try this for eth1 too.
>
> > I don't know where the 169.254.0.0 is coming from as it isn't one of my
> > addresses?
>
> That's the standard 'autoconfigure' address block as defined by RFC1918.
> All systems that can't obtain a DHCP lease are entitled to grab a random
> address in that range (169.254/16). It's most commonly used by Microsoft
> Windows to enable NetBIOS over TCP/IP to function in a broken network
> environment.
>
> Regards,
>
> Jon
> --
> Jon Fautley <jfautley at redhat.com>     direct: +44 1483 739615
>   Presales Technical Consultant        office: +44 1483 300169
>   Red Hat UK                           mobile: +44 7841 558683
>   10 Alan Turing Road, Surrey Research Park, Guildford GU2 7YF
>
>
> _______________________________________________
> Sussex mailing list
> Sussex at mailman.lug.org.uk
> Sussex LUG Website: http://www.sussex.lug.org.uk/
> http://mailman.lug.org.uk/mailman/listinfo/sussex
>
>