Binding error/bug ?
dhcp1 at thehobsons.co.uk
Fri Apr 20 12:25:26 UTC 2007
Bruce Hudson wrote:
> 2) > Bootstrap Protocol
> > Client MAC address: 00:06:0d:4d:c0:be (HP_4d:c0:be)
> 3) > Option 61: Client identifier
> > Hardware type: Ethernet
> > Client MAC address: 00:50:b9:80:0e:56 (XitronTe_80:0e:56)
> The first is presumably the address of your relay or the last router in
>the path between it and your server.
No, that is MUCH higher up in the packet trace, and is the layer 2
source address of the packet. What we are looking at here is the
CONTENT of the DHCP packet. The Client MAC address is a field filled
in by the client and which SHOULD be left alone by the relay agent to
allow the server to identify the client.
> The second is the MAC address that is
>associated with the lease. The client identifier, which is a MAC address by
>co-incidence (or client choice) is the unique key that will determine what
>address you get. It will match client identifiers in your configuration, not
REPEAT AFTER ME
Option 61 is NOT the MAC address of the client, it is an optional
field that the client MAY supply and which MAY contain the MAC
address. On many clients this can be configured, but in Windoze it is
not available for user configuration.
You absolutely CANNOT rely on option 61 being the client hardware address.
We CANNOT comment further on this query until we've seen packets
traces, for the same packet, but before and after it goes through the
relay agent. All we have seen is one packet dump without any further
information that would allow us to consider whether any particular
element is or isn't correct. It is perfectly plausible for the stated
client mac address to be correct, and the client id to be something
manually configured by someone or left over from a previous config.
For example, it's perfectly possible that an install script could
copy the hardware address into a config file and configure the client
id as shown, if that config is then copied to another machine (or the
NIC replaced) then the two would no longer match. WE JUST DON'T KNOW
because NOTHING has been stated other than the one packet dump and an
opinion that something is wrong !
More information about the dhcp-users