How to Configure DHCP Faileover Cluster on Windows Server 2012

All of you know that we didn’t have any Failover option for DHCP in the past. So must of us created different DHCP scopes for one IP range on different servers. This was needed to get a partly redundant DHCP option. This was working but any changes for reservations, scope changes or configuration changes must be done manually or with scripts. This took time or wasn’t really successful.

Now with Windows Server 2012, we get a real Failovercluster including configuration replication. Please notice, the only available options are load balanced and hot standby. I will explain you later when you should use which option.

So let us start to configure our cluster.

 

1. You need to install the first DHCP Server and configure the DHCP Scope. This DHCP Server has to be Windows Server 2012 Standard or Datacenter.

http://cloud.klafox.rocks/?p=350

In this scenario I configured the first DHCP Server on Flo-SVR-DC01.

 

2. Next install a new server with Windows Server 2012 or take one other free server from your existing Windows Server 2012 Systems as Windows Server 2012 DHCP Failover Clusterpartner.

How to install a Windows Server 2012 http://cloud.klafox.rocks/?p=203

First Configuration of a Windows Server 2012 http://cloud.klafox.rocks/?p=222

In my case I installed a fresh Windows Server 2012 VM as Failover Partner.

 

3.  Now your could add the new node to your Server Manager, if you want to manage the Server remote. You can also configure the Failover Setup without this, but it helps to manage both Servers later.

http://cloud.klafox.rocks/?p=496

 

4. When the DHCP role was installed correct on the second host and you added the server for management to your management host, than you should see both systems under DHCP.

5. In the next step open the DHCP MMC.

6. In the DHCP MMC please add the second DHCP Server first. You can do this via right click on “DHCP” and than “Add Server”.

http://cloud.klafox.rocks/?p=689

7. Now you should see both DHCP Servers in the list.

12. In the next step we authorize the DHCP server to our Domain.

http://cloud.klafox.rocks/?p=688

13.  Now click right on the scope that you want to cluster and select “Configure Failover”.

14. When the Wizzard starts, you see the Scope that can be clustered.

If you see no available Scope, you have to check if DHCP Service is up, DHCP Server is complet configured and there are no issues with DNS and ADDS.

 

15. Next step is to selecte the failover partner.

16. When you authorized the second DHCP server before, you see it in the second list. Otherwise you have to select “This Server:” and “Browse”.

17.  Now you type in the name of the server.

18. When you entered the name click “Check Names”. When the wizzard found the server, click “ok”.

19. Click “ok” and the server will attached to DHCP MMC.

20. Now you see the selected Server with complet FQDN in the Patern Server field.

Click “Next” to go on.

 

21. Now you have to set the clusterconfiguration.

Load Blanced:

Relationship Name: Name of your Failover Cluster

Maximum Client Lead Time: Defines the amount of time the surviving server will wait before assuming control of the entire scope.

Mode: Load Balanced – When the cluster is configured in Load Balance mode, this results in an active-active setup of the two DHCP Servers.

You should use when you have big networks with many clients or you want to deploy the cluster in different branch offices.

Load Balance Percentage: Means how the work is splitted up between both hosts. The percentages together can only be 100%. The node with the highest percentage gets the highest workload.

State Switchover Interval: automatically change state to partner down after <time>.

Enable Message Authentification: enables authentification from clusternodes.

Shared Secret:Validation Passwort that identifies the node as partners against each other.

 

Standby:

Relationship Name: Name of your Failover Cluster

Maximum Client Lead Time: Defines the amount of time the surviving server will wait before assuming control of the entire scope.

Mode: Hot Standvy – When the cluster is configured in Hot Standby mode, one node is active and the second is standby and will only take over when the primary DHCP Server failes.

You should use when you need the partner as fault tolerance.

Adresses reserved for standby server: Means how many adresse the standby can lease before he takes over the entire scope and becomes active.

State Switchover Interval: automatically change state to partner down after <time>.

Enable Message Authentification: enables authentification from clusternodes.

Shared Secret: Validation Passwort that identifies the node as partners against each other.

 

22. After klicking “Next” you see a short summary of your configuration.

23.  Klick “Finish” and the cluster configuration starts.

24. In the DHCP MMC click right on the Scope and force a replication by clicking on “Replicate Failover Scope” and than on the refresh  button or press F5.

25. On the failover node check the config. If the config is right your finished.

 

 

 

 

How to authorize a Windows DHCP Server to Active Directory (AD, ADDS)

If you want to authorize please perform following steps.

1. In the DHCP Server MMC, click right on the Server Scope (IPv4 or IPv6) that you want to Authorize and click “Authorize”.

2. Now you should see a green hook on the Scope Symbol and that’s all. You authorized the DHCP Server Scope to your Domain.

 

 

 

How to install a Server Role and/or Feature remote via Server Manager

1. Add the Server you want to install the role and/or feature to Server Manager

http://cloud.klafox.rocks/?p=496

2. Go to “All Servers” in Server Manager and right click on the System you want to install. Than select “Add Roles and Features”.

 

3. The wizzards starts.

 

4. Select the server where you want to install the role or feature and klick “Next”

 

5.  Select the role you want to install. In my case the DHCP Role (When you install Roles like DHCP, DNS or AD roles, please set a static IP for the NIC of the targetserver).

 

6. Select the feature you want to install (in my case I don’t want to install one so I click directly on “Next”.

 

7. A short infoscreen appears. Click “Next”.

 

9. Now you see a summary which roles and features will be installed on the system. Click “Install” to install this components.

 

10. Installation begins, you can close this window. The installation will take up a few minutes.

 

11. To check if your role is installed on the system, click on the role in Server Manager and check if you see the server.

Some thoughts on NIC Partitioning (NPAR)

Hello community,

today some thoughts about NIC Partitioning.

What means NPAR?

NPAR is a technology used by Broadcom and Qlogic, which enables you to split up traffic on a Networkcard (NIC) in 4 partitions. NPAR is only working with 10GbE technology. NPAR is similar to QoS on Networklayer.

How can I split up the traffic?

You can split up the traffic in different ways. Like with QoS with UDP or TCP packet filter but you have also the option to split up in VLANs, Virtual Machine or application client. For application client your software must support multitenant.

When I split up the bandwidth is this dynamic or static?

You can have both, it is possible to set up static partitions or you can set it dynamic.

As you see in the picture, there are static bandwidths configured for iSCSI, FCoE, TOE and other protocols and dynamic for VMs. It is also possible that more than one Protocol or VM uses one Partition.

Has my switch to support NPAR?

No. Your switch has not to support, for your switch there is no different between an NPAR NIC than a normal NIC.

What about the OS, can every OS handle NPAR?

No. There is a list of supported OS,  Windows Server 2008 and 2008 R2, Windows Server 2012, Windows Hyper-V Server, RHEL, 5.5 and 6, SLES 11 SP1, VMware ESX, ESXi 4.1 and ESXi 5.0.

Why should I use NPAR?

From datacenter perspective it would reduce your cabling and maximize bandwidth.

How does this work?

Let me give you an example. For an Hyper-V cluster you need one NIC for Management, one NIC for Cluster Heartbeat, one for Livemigration, one for cluster shared volume, two iSCSI NICs and two or more NICs to attach you Company and VM Network.  So in the past you had many NICs or a few NICs with many ports. With NPAR it is possible to use at least two 10GbE NICs  Ports to setup the complete network environment. That reduces cabling and you don’t need big 2U sever to hold all the cards.

How to configure NPAR?

Check out the links and the video below.

 

 

Qlogic

http://www.qlogic.com/OEMPartnerships/Dell/Documents/Converged/QME8242-k_converged_IO_Deployment_Guide.pdf

http://www.youtube.com/watch?v=rK1OXNKynNw

Broadcom

http://i.dell.com/sites/doccontent/shared-content/data-sheets/en/Documents/dell-57712-k-cndc-npar-setup-guide.pdf

For more questions and information take a look in this whitepapers and blogs.

http://www.dell.com/downloads/global/products/pedge/en/Dell-Broadcom-NPAR-White-Paper.pdf

http://social.technet.microsoft.com/wiki/contents/articles/9711.hyper-v-network-design-configuration-and-prioritization-guidance.aspx

http://www.dell.com/downloads/global/products/pwcnt/en/broadcom-57712-k-faq.pdf

http://bladesmadesimple.com/2011/10/dell-network-daughter-card-ndc-and-network-partitioning-npar-explained/

Thank you for reading.

Regards

Flo