status update

Simon Hobson dhcp1 at
Wed Dec 13 07:52:00 UTC 2006


>In the short term, we have thrown more addresses at the problem -
>creating a second subnet on the same VLAN (Cisco environment - using the
>"secondary" keyword").
>However, we're having an issue where the machines in the new/secondary
>subnet can't complete their PXE process.  The vendor is telling us that
>they boot a couple of different times in the PXE process before the OS
>is laid down - and that during this dual boot process - they're getting
>different IP addresses from the ISC daemon - even though they're sending
>the same GUID each time (they acknowledge that the GUID is changing
>after Windows is laid down).   From what I've read - this is not
>expected behavior.  Can anyone validate please ?

Correct, if the ClientID is the same and not empty, or the ClientID 
is empty AND the MAC address is the same, then the client should get 
the same address.

I find it very worthwhile to sniff the network packets so you can see 
exactly what options are being requested/passed etc in the data 
packets - just in case there is a 'difference of opinion' either in 
the creation or interpretation. For instance, is the ClientID 
actually identical (eg could it be NULL in one request but an empty 
string in another ?)

More information about the dhcp-users mailing list