DHCP Failover patch to address issues caused with a redundant router configuration
Are you using or plan to use a redundant IP helper configuration on routers with a Windows Server 2012 DHCP failover? Some earlier versions of router redundancy implementations (HSRP-Hot Standby Router Protocol/VRRP-Virtual Router Redundancy Protocol) by Cisco and other vendors introduce redundancy in the DHCP messages while relaying DHCP requests. This means a duplication of relayed DHCP client messages. The DHCP failover servers receive the lease requests twice from the same client.
The issue that you might face if you use such a redundant router configuration with a DHCP failover pair is the DHCP servers responding differently to the two identical requests from the two routers depending on which request arrives first. Some symptoms of this issue are:
- Clients getting leases with inconsistent lease duration (sometimes Maximum Client Lead Time and sometimes the lease duration configured on the scope).
- Clients being assigned duplicate leases (IP addresses which have already been allocated to others).
The fix for this is now available here. Applying this fix on both the DHCP servers participating in the failover relationship will enable them to handle the duplicate lease requests in a consistent manner and not run into issues like duplicate leases being issued.
Cisco had removed the redundancy in DHCP requests in later releases of IOS. If you are using one of these later releases then you would not need to apply this fix. For Cisco routers and their proprietary Hot Standby Router Protocol the information is available here.