Specifications

Best Practices for Virtualizing and Managing Exchange 2013

Best Practices and Recommendations

All hardware should be certified for Windows Server 2012, and the failover cluster solution should

pass all tests in the Validate a Configuration Wizard. For more information about validating a

failover cluster, see Validate Hardware for a Windows Server 2012 Failover Cluster.

Windows Server 2012 Hyper-V supports scaling clusters up to 64 nodes and 8,000 virtual machines per

cluster. Windows Server 2012 also provides Windows PowerShell cmdlets and a snap-in for Failover

Cluster Manager, which allows administrators to manage multiple clustered nodes.

The aim of failover clustering is to provide a resilient solution for workloads that are running on top of the

cluster. These could be clustered roles such as a file server, a DHCP server, or (as in the case of this paper)

a virtual machine. The Windows Server 2012 Failover Clustering capability provides resiliency for a number

of other roles or services. Additional information about this capability can be found here.

With virtual machines running on top of a Hyper-V cluster, should a Hyper-V cluster node fail, the virtual

machines also will experience downtime. However, the remaining cluster nodes immediately start working

to bring those virtual machines up again on an alternative node within the cluster—helping to ensure that

downtime is minimal and that administrator intervention is not required. This means that the workload—

in this case, Exchange 2013—has an extra level of resiliency when compared to a physical implementation.

The physical implementation requires the administrator to rely on only application-level resiliency, as

opposed to the dual layers of resiliency provided by a clustered Hyper-V implementation. This topic is

explored in more detail later in this paper.

Cluster Shared Volumes

With Cluster Shared Volumes (CSV), Windows Server 2012 provides shared storage resources for clustered

virtual machines. The CSV technology makes it easy to manage large numbers of LUNs in a failover

cluster. CSV allows simultaneous read-write permissions to the same NTFS-provisioned LUN disk from

multiple cluster nodes (Figure 23). Due to significant improvements in the CSV design in Windows Server

2012, Cluster Shared Volumes can perform an increased number of operations in direct I/O mode as

compared to previous versions of Windows Server. This feature also can be helpful while operating

workloads for failover, conducting live migration, or moving workloads. It allows workloads to share the

same volume without impacting the other workloads.

 CSV provides better security for mission-critical data by supporting BitLocker Drive Encryption.

 For decrypting data, CSV uses cluster name objects as identity for data decryption.

 CSV provides continuous availability and scalability of file-based server storage for applications or

databases using the Scale-Out File Server feature.

 CSV supports two types of snapshots, including application-consistent and crash-consistent

Volume Shadow Copy Service (VSS) snapshots.

 CSV supports clustered VHD files for clustered workloads.