I have the following network (map below). When the T1 is operational DHCP clients at the remote site in VLAN 1 (10.10.10.0/24) work as expected via the ip helper on the 'Site Router'. However, when there is a long term failure of the T1, clients can no longer get addresses.
After looking at some documentation (http://www.cisco.com/c/en/us/td/docs/ios-xml/ios/ipaddr_dhcp/configuration/12-4t/dhcp-12-4t-book/config-dhcp-relay-agent.html#GUID-154A7C7F-1126-474F-926D-7CCD8B73FF1C) I came across the following blurb:
"Relay agent forwarding is distinct from the normal forwarding of an IP router, where IP datagrams are switched between networks somewhat transparently. In contrast, when relay agents receive DHCP messages, the agents generate a new DHCP message to send out through another interface." -(emphasis added)
So the crux of my question is this: Does the Cisco implementation of the 'ip helper-address' for DHCP relay drop DHCP Discover packets when the destination DHCP server would be reachable via the same interface in which the DHCP Discover packet arrived?
Before everyone gets distracted by the routing aspects of this... Routing is fine. Clients that already have addresses can reach the Corporate Network just fine via the VPN provided by the ASA. Yes, they can even ping the DHCP server in question. It is *not* routing (unless you can come up with a compelling argument otherwise).
