mac address problem

Simon Hobson dhcp1 at
Sun Sep 17 06:54:32 UTC 2006

garry saddington wrote:

>We bought a bunch of Wyse Winterms cheaply to use in an LTSP setup.
>About one third of them work perfectly by flashing them with a kernel
>and an initial ram disk.
>However, the others do not work and the problem is that they all report
>their mac address as 00:00:00:00:00:00.

Now you know why they were cheap ;-)

>Consequently they all get the
>same ip address from dhcp and everything gets a little confusing!

Not only that, but even if they did get different addresses (which 
you could do by setting different client-ids BTW), networking would 
still not work because the MAC address is used to identify the 
intended recipient of each packet.

>Is there any way to solve this? We thought about a kernel parameter
>specifying mac address on boot but all our searches have not found
>anything. Can someone here suggest an avenue to follow?

There will almost certainly be a utility somewhere to allow you to 
set the mac address of the device and store it in non volatile memory 
- the manufacturer probably has such a utility.

It might be more fruitful to look at the network port hardware rather 
than the terminal as a whole, not having looked into it, I would 
imagine that most ethernet chips include a small memory for such 
configuration values.


PS - This brings to mind a story told to me by someone running a 
network in a University. A couple of years ago they renewed the PCs 
in a number of areas and so bought them (Dells) in bulk. They had a 
curious network problem which they eventually tracked down to ... a 
duplicate MAC address. It turned out that due to a bug in Dells 
systems, in every block of 256 machines, there were two with the same 
MAC address. This could have gone on for years, because unless they 
sold a large number of the same model in one order, the machines 
would have been on different networks owned by different people - and 
so the duplicates would not matter.

More information about the dhcp-users mailing list