This article details DHCP failover, a new feature in Windows Server 2012, which provides continuous address availability across DHCP servers.
With Server 2008 R2, many were disheartened not see a truly resilient rethinking of DHCP readiness. While Microsoft did expand DHCP clustering, the only real alternative was still Split Scope. The problems with Split Scope are well known.
DHCP Failover in Windows Server 2012
Following Microsoft’s guidelines of a 70/30 division between subnets, one could quickly run out of available IP addresses. Specific scope options and the reservation of important IPs make the proper maintenance nearly unmanageable in this setup as these have to be manually replicated. The DHCP role received some much needed attention with the release of Windows Server 2012. While Policy-based address assignment is getting a lot of the press lately, we will be focusing our attention on the addition of DHCP Failover.
In order for DHCP Failover to be completely effective, it has to accomplish two major goals. First, it must provide automatic address configuration at all times on a network. The second, more subtle, goal is the ability for a client to retain its previously assigned address in the event of an outage. If the failover option lacked this second goal, an unexpected DHCP failure would result in the manual renewal of addresses on clients or by having to restart machines. Neither option would be desirable. The DHCP server failover feature allows for a maximum of two DHCP servers to share the scopes and scope options through the replication of lease information.
DHCP failover replication can occur in one of two modes. Mode selection is largely dependent on your physical site layout with some emphasis on the reliability of inter-site connections. The first mode is Hot Standby mode. With Hot Standby, two DHCP servers are assigned to a subnet. The server marked as primary handles address allocation and replicates its data to the secondary server. The secondary server, though always active, only assigns addresses if the primary server becomes unavailable for your specified amount of time.
DHCP Failover Hot Standby
This configuration is ideal in Remote Site/Central Site (otherwise known as Hub and Spoke). The recommended setup places the primary server at the remote site and the secondary server in the central location. The key item to remember is that the primary/secondary setting is scope specific. A single server can be a primary server to one subnet and a secondary server to another subnet. Microsoft provides the following example of a Hot Standby situation.
DHCP Failover Hot Standby: The Central Site server provides secondary services to all remote sites.
DHCP Failover Load Sharing
The second and default DHCP failover mode is Load Sharing. With Load Sharing, the two DHCP servers assigned to a subnet can both distribute addresses. To ensure that multiple assignments are not attempted on a client, servers respond based on a predetermined load configuration. This load configuration can be altered at any time and acts like a live version of the Split Scope setup.
DHCP Failover Load Sharing: The two DHCP servers load share on a physical site that has two subnets.
DHCP Failover setup
The setup of DHCP Failover is really quite simple. In fact, it can be configured by right clicking on a subnet and selecting Configure Failover.
Windows Server 2012 DHCP Failover setup
With the addition of 103 new DHCP related PowerShell cmdlets, much of the initial work can be automated. Including the requirement that both servers run Windows Server 2012, the only other requirement are that the time difference between the devices be less than one minute.
As our businesses and organizations continue to demand 100% availability, it is awesome to finally have a complete solution to IP address assignment that includes a robust failover system. For those that can’t wait to get their hands on this feature, the links below will surely help!