Windows Server 2008 R2 Disks

Windows disks can be defined as basic or dynamic disks. Furthermore, these same disks can be defined as Master Boot Record (MBR) or GUID Partition Table (GPT) disks. A simple way to clearly differentiate how to choose between these disk types is to consider that basic disks only support simple volumes, whereas dynamic disks allow logical volumes to be created across multiple physical disks. Choosing between MBR and GPT disks depends on the size of the disk, as well as understanding how many partitions you will need to create on the disk.

Windows Server 2008 R2 also supports VHD or virtual hard disks, for Hyper-V virtual machines. VHD disks can now also be created and mounted directly within a Windows host operating system, regardless of whether the Windows Server 2008 R2 system is hosting the Hyper-V role.

Master Boot Record Disks

Master Boot Record (MBR) disks utilize the traditional disk configuration. The configuration of the disk, including partition configuration and disk layout, is stored on the first sector of the disk in the MBR. Traditionally, if the MBR became corrupted or moved to a different part of the disk, the data became inaccessible. MBR disks have a limitation of three primary partitions and a single extended partition that can contain several logical drives. Choosing to create an MBR disk should provide administrators with a more compatible disk that can easily be mounted and/or managed between different operating system platforms and third-party disk management tools.

GUID Partition Table (GPT) Disks

GPT disks were first introduced in Windows with Windows Server 2003 Service Pack 1. GPT disks are recommended for disks that exceed 2TB in size. GPT disks can support an unlimited number of primary partitions and this can be very useful when administrators are leveraging large external disk arrays and need to segment data for security, hosting, or distributed management and access. GPT disks are only recognized by Windows Server 2003 SP1 and later Windows operating systems. Attempting to manage a GPT disk using a previous operating system or third-party MBR disk management tool will be blocked and virtually inaccessible.

Basic Disk

A Windows disk is defined as a basic or a dynamic disk regardless of whether the disk is an MBR or a GPT disk. A basic disk supports only simple volumes or volumes that exist on a single disk and partition within Windows. Basic disks contain no fault tolerance managed by the Windows operating system, but can be fault tolerant if the disk presented to Windows is managed by an external disk controller and is configured in a fault-tolerant array of disks.

Basic disks are easier to move across different operating systems and usually are more compatible with Windows and third-party disk and file system services and management tools. Basic disks also support booting to different operating systems stored in separate partitions. Furthermore, and most important, if the disk presented to Windows is from a SAN that includes multiple paths to the disk, using a basic disk will provide the most reliable operation as a different path to the disk might not be recognized if the disk is defined within Windows as a dynamic disk.

Dynamic Disk

Dynamic disks extend Windows disk functionality when managing multiple disks using Windows Server 2008 R2 is required. Windows administrators can configure dynamic disks to host volumes that span multiple partitions and disks within a single system. This allows administrators to build fault-tolerant and better performing volumes when RAID controllers are not available or when a number of smaller disks need to be grouped together to form a larger disk.

In some server deployments, dynamic disks are required as the disk controllers do not support the necessary performance, fault-tolerance, or volume size requirements to meet the recommended system specifications. In these cases, dynamic disks can be used to create larger volumes, fault-tolerant volumes, or volumes that can read and write data across multiple physical disks to achieve higher performance and higher reliability. Dynamic disks are managed by the operating system using the Virtual Disk Service (VDS).

Virtual Hard Disks

Virtual hard disks or VHDs are used by virtual machines to emulate Windows disks. Virtual hard disks can be created on an existing Windows Server 2008 R2 system using the Hyper-V Management console or they can be created directly using the Disk Management console. VHDs are primarily created on the Windows host system as a file on an existing Windows volume that has a .vhd extension. VHD disks can be created to be fixed size or dynamically expanding. A fixed-sized VHD that is 10GB in size will equate to a 10GB file on the Windows host server volume. A dynamically expanding VHD file will expand as files are stored on it, only as necessary. VHD files can easily be moved across servers and between virtual machines, and also can be expanded quite easily, granted that the VHD is not in use and there is ample free space on the host volume. VHD files can be attached directly to a Windows Server 2008 R2 host using the Disk Management console, unlike in previous releases, which required scripts to mount the file. This added functionality is a needed improvement to the integrated VSS Hyper-V backup functionality, included with Windows Server Backup and available to third-party backup software vendors. 

Partition or Volume

When referring to Windows disks, administrators might consider partitions and volumes interchangeable. In fact, even though the graphical user interface makes no clear distinction and might refer to everything as a volume, volumes only exist on dynamic disks and partitions only exist on basic disks. This is especially important when managing disks using the diskpart.exe command-line utility, which defines a clear delineation between partitions and volumes.

Mount Point

When a new volume is created in Windows, it can be assigned a drive letter or mounted into an existing empty folder on an existing volume. When a volume is mounted into a folder, this is known as a mount point or junction point. Mount points can be very useful in situations where administrators want to simplify disk access for end users, but must also make use of a number of small disks versus a single large disk. For example, on a database server with three disks, an administrator might assign disk1 the D drive, disk2 would be mounted in d:\data, and disk3 would be mounted in d:\logfiles. Any administrator would only need to connect to the D drive to access the databases or log files. One thing that administrators must test before using mount points is to see that all clients, applications, and backup agents support the use of mount or junction points and can successfully access and back up data stored within them. With many backup applications, enabling a backup job to back up data stored on a mounted volume is not the default and can cause major problems if the correct backup configuration is not selected before a failure occurs.

Simple Volumes

A simple volume is a single partition created on a single basic or dynamic disk. On a basic disk, simple volumes can be extended to include free, unallocated space that exists in a sequential section of the disk. To extend a simple volume to a noncontiguous, unallocated space on the same disk or a different disk, the disk will need to be upgraded to a dynamic disk.

Spanned Volumes

A spanned volume is treated as a single drive, but the volume spans two or more disks or different noncontiguous areas of the same disk. Spanned volumes provide no disk fault tolerance but can be used to meet disk storage needs that exceed the capacity of a single disk or volume. Spanned volumes are slowest when it comes to reading and writing data and are recommended only when the space of more than a single disk is necessary or an existing simple volume needs to be extended to add disk space and there is no available, unallocated space located next to the volume. For instance, if an application, file share, or service is dependent on the drive letter, does not support the moving of data or system files to another drive, and the current drive is nearly full, a simple volume can be upgraded to a spanned volume and extended with unallocated space on the same or another disk to add additional disk space. A simple volume that has been extended with unallocated space on the same disk is still considered a simple volume. If the simple volume is extended to a different disk, it is automatically converted to a spanned volume. The allocated space on each of the disks can be different sizes, and there is no space lost when creating a spanned volume. One thing to keep in mind, though, is that a spanned volume can never be reverted to a simple volume.

Striped Volumes

A striped volume or RAID-0 compatible volume requires two or more Windows dynamic disks and provides the fastest of all disk configurations. Striped volumes read and write data from each of the disks simultaneously, which improves disk access time. Striped volumes utilize all the space allocated for data storage but provide no disk fault tolerance. If one of the disks should fail, the entire data set would become inaccessible. Stripe sets require the exact amount of disk space on each of the allocated disks. For example, to create a 15GB stripe set array with three disks, 5GB of unallocated space would be required on each disk.

Fault-Tolerant Volumes

When fault-tolerant disk arrays managed by hardware controllers are not available, fault-tolerant volumes can be created using multiple Windows dynamic disks. Fault-tolerant volumes in Windows are able to maintain data availability in the event of a single disk failure. Windows Server 2008 R2 supports two types of fault-tolerant volumes, including mirrored volumes and RAID-5 volumes.

Mirrored Volumes

Mirrored or RAID-1 compatible volumes require two separate disks to create. Furthermore, the size of the volume must be equal and available in one contiguous, unallocated section of each of the disks. Mirrored volumes duplicate data across each disk and can withstand the failure of a single disk. Because the mirrored volume is an exact replica of the first disk, the total space capacity is the capacity of one disk.

RAID-5 Volumes

Software-based RAID-5 volumes require three or more Windows dynamic disks and can provide faster disk read access than a single disk because all disks in the set can be read at the same time. Write performance can be slower than a single disk because of the parity stripe that must be generated and written. The space allocated to the RAID-5 volume on each disk in the volume must be equal and contiguous unallocated space. For example, to create a RAID-5 volume that requires 100GB on each disk, a disk with two separate areas of 50GB of unallocated space cannot be used to participate in the volume.

RAID-5 sets can withstand the failure of a single disk in the volume. During a disk failure, the remaining disks in the volume will continue to provide access to data but at a slower or degraded rate. This capability is achieved by reserving a small portion of each disk’s allocated space to store data parity information that can be used to rebuild a failed disk and to continue to provide data access. This is called a parity stripe. RAID-5 parity information requires the total space of a single disk in the array. For example, if five 10GB dynamic disks are used to create a single RAID-5 volume, 40GB would be available for data storage. The reserved 10GB would be spread evenly across all five disks. The formula for usable capacity of a RAID-5 array is (N - 1) * S, where N is the total number of drives in the array and S is the capacity of the smallest drive in the array.