Recommendations for Using Cluster Shared Volumes in a Failover Cluster in Windows Server 2008 R2

Updated: January 21, 2010

Applies To: Microsoft Hyper-V Server 2008 R2, Windows Server 2008 R2

This topic lists recommendations for using Cluster Shared Volumes (CSV) in a failover cluster running Windows Server 2008 R2 or Hyper-V Server 2008 R2. For important information about requirements for CSV, see Requirements for Using Cluster Shared Volumes in a Failover Cluster in Windows Server 2008 R2, in this guide.

In this topic

• Recommendations for planning the arrangement of LUNs or disks, volumes, and VHD files

• Recommendations for planning the number and size of LUNs and volumes for CSV

Recommendations for planning the arrangement of LUNs or disks, volumes, and VHD files

To make the best use of CSV, it is helpful to review the way you would have arranged the LUNs or disks if you were configuring physical servers. When you configure the corresponding virtual machines, try to arrange the virtual hard disks (VHD files) in a similar way.

Consider a physical server for which you would organize the disks and files as follows:

• System files, including a page file, on one physical disk

• Data files on another physical disk

For an equivalent clustered virtual machine, you should organize the volumes and files in a similar way:

• System files, including a page file, in a VHD file on one volume in CSV

• Data files in a VHD file on another volume in CSV

If you add another similar virtual machine, where possible, you should keep the same arrangement for that virtual machine's VHDs: place system and page files on the CSV volume that is used for system files, and data on the CSV volume that is used for data.

The following figures illustrate this arrangement. Figure 1 shows a configuration that contains physical servers only.

Figure 1   Physical servers with system and data disks

Figure 2 shows virtual machines, similar to the physical servers in Figure 1, running in a cluster that uses CSV. The VHD files are arranged as they would have been for the equivalent physical servers. Because the cluster uses CSV, the cluster nodes can readily access each file in this configuration as needed.

Figure 2   Virtual machines with system and data VHDs

Recommendations for planning the number and size of LUNs and volumes for CSV

When planning the storage configuration for a failover cluster that runs Windows Server 2008 R2 or Hyper-V Server 2008 R2 and uses CSV, consider the following recommendations:

• When you are deciding how many LUNs to configure, consult with your storage vendor. For example, your storage vendor may recommend that you configure each LUN with one partition and place one Cluster Shared Volume on it.

• When planning the storage configuration for a particular virtual machine, consider the disk requirements of the service, application, or role that the virtual machine will support. Understanding these requirements helps you avoid disk contention that can result in poor performance. The storage configuration for the virtual machine should closely resemble the storage configuration that you would use for a physical server running the same service, application, or role. For more information, see Recommendations for planning the arrangement of LUNs or disks, volumes, and VHD files, earlier in this topic.

  For example, consider a physical server for which you place the operating system files, the log files, and the application data on separate physical disks. For the equivalent virtual machine, you should place the operating system files, the log files, and the application data in separate VHD files on separate volumes in CSV. If you have multiple virtual machines with the same requirements, you can designate one volume in CSV for VHDs for system files, one for VHDs for log files, and one for VHDs for application data.

  You can also mitigate disk contention by having storage with a large number of spindles (physical disks that operate independently of one another). Choose your storage hardware accordingly, and consult with your vendor so that you understand how to optimize the performance of your storage.

• To decide how many Cluster Shared Volumes to have, consider the number of virtual machines that you plan to have in the cluster, and the workload for each virtual machine. Consider the following examples:

  • One organization is deploying virtual machines that will support virtual desktop infrastructure (VDI), which is a relatively light workload. The cluster uses high-performance storage. The cluster administrator, after consulting with the storage vendor, decides to place a relatively large number of virtual machines per Cluster Shared Volume.

  • Another organization is deploying virtual machines that will support a heavily-used database application, which is a heavier workload. The cluster uses lower-performing storage. The cluster administrator, after consulting with the storage vendor, decides to place a relatively small number of virtual machines per Cluster Shared Volume.