Setting Up Failover Clustering for Hyper-V (Part 8)

by [Published on 22 Sept. 2011 / Last Updated on 22 Sept. 2011]

This article continues the discussion of Hyper-V clustering by demonstrating the procedure for creating a simple two node failover cluster.

If you would like to read the other parts of this article series please go to:

So far in this article series, I have spent a lot of time discussing iSCSI connectivity and the general architecture that is used by a Hyper-V cluster. Now it is time to take what you have learned and actually begin building a failover cluster for Hyper-V.

Before I Begin

Before I get started, I need to make sure that we are all on the same page. You should have two Windows 2008 or 2008 R2 servers that are each connected via iSCSI to a shared storage mechanism. Each of the two servers should also have Hyper-V installed and have multiple network adapters configured. Ideally, each cluster node should have three network adapters – one to connect the server to the storage LUN, another for cluster specific traffic, and a third adapter for general networking traffic. Of course in the real world host servers that participate in Hyper-V clusters typically use more than three network adapters, but three adapters are sufficient for demonstration purposes. For the purpose of this article series, I am also assuming that your cluster nodes are domain members.

Installing the Failover Clustering Feature

The next step in the configuration process is to deploy and configure the Failover Cluster feature. To do so, open the Server Manager on your first cluster node. Navigate through the console tree to Server Manager | Features and then click the Add Feature link. When you do, Windows will ask you which features you want to install. Select the Failover Clustering check box and then click Next, followed by Install and Close. Now, repeat the process to install the Failover Clustering Feature on your second cluster node.

Validating the Cluster

Before you can configure the cluster you need to run some validation tests to insure that your cluster nodes have the ability to participate in the cluster. Begin by going to your first cluster node and opening the Failover Cluster Manager Console, which is located on the server’s Administrative Tools menu. Now, click the Validate a Configuration link, which is located in the console’s Management pane, as shown in Figure A.


Figure A: Click the Validate a Configuration link.

At this point, Windows will launch the Validate a Configuration Wizard. Click Next to bypass the wizard’s Welcome screen and you will see a screen telling you that if you want to validate a set of servers then you must add the names of all of the servers. Enter the fully qualified domain name of your first cluster node and click Add. Next, enter the fully qualified domain of your second cluster node and click Add. Both nodes should be listed within the Selected Servers list, as shown in Figure B.


Figure B: You should validate all of your cluster nodes.

Click Next and you will see a screen asking you if you want to run all of the available tests or if you prefer to choose individual tests to run. Select the Run All Tests (Recommended) option and click Next. You should now see a summary screen outlining the validation tests which are about to be performed. Click Next one more time to begin the testing process.

The validation process only takes about three or four minutes to complete. When the validation finishes, Windows will display a Failover Cluster Validation Report like the one that is shown in Figure C. As you look at the figure, you will notice that there is a message above the actual report stating that the configuration is suitable for clustering. When you have finished reviewing the report, click Finish.


Figure C: The Failover Cluster Validation Report confirms that the configuration is suitable for clustering.

Creating a Cluster

Now that the cluster nodes have been validated, it is time to actually create the cluster. To do so, click the Create a Cluster link, which is shown in Figure A. When you do, Windows will launch the Create a Cluster Wizard.

Click Next to bypass the wizard’s Welcome screen and you will see a screen asking you which servers you want to include in the cluster (this screen looks almost identical to the one shown in Figure B). Enter the fully qualified domain name of the first cluster node and click Add. Next, enter the fully qualified domain name for the second cluster node and click Add, followed by Next.

You should now see a screen similar to the one shown in Figure D. As you can see in the figure, you must provide a name for the cluster. This name is similar to a computer name, except that it is used to reference the cluster as a whole. Just as a cluster name is required, so too is a cluster IP address. The IP address that you provide will be assigned to the cluster as a whole and is different from any of the IP addresses used by individual cluster nodes.


Figure D: You must assign a name and an IP address to the cluster.

Click Next and you will see a summary screen outlining the settings that you have provided. Assuming that the information shown on this screen is correct, click Next to create the cluster. When the process completes, click Finish.

Creating a File Share Witness

In the previous article, I explained that the only way that a two node cluster could provide failover capabilities was to use a file share witness in place of a third cluster node. Now it is time to create that file share witness.

You can create the file share witness on any server that is accessible to both cluster nodes, but as a best practice you should consider placing the file share witness either on a domain controller or on your storage server. Once you have decided which server to use, open an elevated Command Prompt window on that server and enter the following commands:

C:

CD\

MD MNS-FSW

Net Share FSM_Hypercluster=C:\MNS-FSW /Grant:Hypercluster$,FULL

Cacls c:\MNS-FSW /G BUILTIN\Administrators:F Hypercluster$:F

This block of code creates a folder named C:\MNS-FSM and then shares it using the name FSM_Hypercluster. I am using Hypercluster because that is what I named my cluster. You should change the share to reflect your cluster name.

The fourth line of code grants full rights to the share to Hypercluster, which is the name of the cluster. Notice that I placed a dollar sign after Hypercluster. The dollar sign indicates that this is a computer account rather than a user account.

The last line of code grants the necessary NTFS permissions to the Hypercluster computer account and to the built-in Administrators group. You can see what these commands look like in action in Figure E.


Figure E: You must enter these commands on the server that will host the file share witness.

Conclusion

Now that we have created the file share witness directory, we must configure our cluster to use it. I will show you how to do that in Part 9 of this series.

If you would like to read the other parts of this article series please go to:

Advertisement

Featured Links