Alex - stock.adobe.com

How to choose and configure Windows Server RAID levels

Evaluate your storage on Windows Server to see if it can benefit from RAID's fault tolerance or performance options. These three RAID levels have different uses and setups.

Storage disk failures and capacity and performance issues are a fact of life for server administrators. Grouping disks together through RAID, though, helps to minimize the impact.

RAID is one of the oldest and most reliable approaches for aggregating space. RAID -- which stands for redundant array of independent disks or redundant array of inexpensive disks, depending on who you ask -- associates two or more drives to a single storage area. Depending on the configuration choice, the resulting RAID may provide fault tolerance and better performance.

Modern OSes, like Windows Server, can manage software-based RAID configurations. This article examines the Windows Server RAID levels available in modern versions of the platform -- 0, 1 and 5.

RAID 0 (disk striping)

RAID 0, or disk striping, distributes data across two or more drives. The data is distributed in stripes across the allocated spaces. From the user's perspective, the storage space is a single volume, such as the M: drive. Each disk is only responsible for a small portion of the overall read/write effort, greatly increasing the storage system's performance. However, one failed drive in the array results in the loss of all stored data.

When to use RAID 0

Use RAID 0 when your goal is increased performance and you don't care about fault tolerance. Disk striping does not implement any data redundancy or use parity to duplicate data stored on the other disks.

RAID 0 uses include the following:

  • Cached or other temporary information.
  • Noncritical data backed up elsewhere.
  • Scratch or working space that does not permanently store data.

Think carefully about the number of devices in the array. More devices increase performance. However, the additional drives also increase the chances of failure, resulting in the loss of all data.

Configure RAID 0

Windows Server makes configuring RAID easy. The settings are wizard-based. Determine which disks will be part of the array, what drive letter you'll assign and which file system to use -- usually New Technology File System (NTFS). For best results, the disks should be the same size and type.

Here is the process:

  1. Open the Disk Management utility.
  2. Confirm all disks are online and initialized.
  3. Convert any Basic Disks to Dynamic Disks.
  4. Right-click unallocated space on one of the disks, and select New Striped Volume (RAID 0).
  5. Select the remaining disks you want to add to the array in the wizard. You must select a minimum of two disks.
  6. Assign a drive letter -- or mount point -- and select NTFS.
  7. Select Finish in the wizard.

Be sure to define a backup plan. Remember, the failure of any one disk results in all the data being gone.

Screenshot of how to select a New Striped Volume
Right-click a volume, and select New Striped Volume to build a RAID 0 array.
Screenshot of disk selection
Select at least one more disk to include in the array.
Screenshot of drive letter and mount point assignments
As with other storage space, assign a drive letter or mount point.
Screenshot of storage space formatting
Format the storage space, and provide a volume label.
Screenshot of completing New Striped Volume Wizard
Select Finish to begin building the array.
Screenshot of RAID 0 array
The RAID 0 array is shown above as drive letter E: in green.

Recover from a failed drive in RAID 0

The clearest indicator of a problem is unavailable data. If you suspect a problem, look for Event Viewer logs indicating a failed drive. You may also see indicators of a failed, unavailable or offline disk in the Disk Management utility.

You cannot recover or rebuild a RAID 0 array in Windows. You'll simply create a new one by replacing the failed drive and building a new Windows Server RAID.

RAID 1 (disk mirroring)

RAID 1 mirrors the data to two identical storage devices. If one device fails, the other device contains the same data, so nothing is lost.

When to use RAID 1

Use RAID 1 when data redundancy is critical, but there's not a lot of it. The problem with RAID 1 is its cost relative to its storage capacity. Suppose you want to mirror a 75 GB partition containing an OS. You've already paid for the original 75 GB of capacity and are now allocating another 75 GB for the same information. You've paid for 150 GB of space, but it only stores 75 GB of content. The idea is that the data is critical enough to justify the design.

Configure RAID 1

The configuration process for a RAID 1 array in Disk Management is nearly the same as with RAID 0:

  1. Make sure both disks are online and initialized.
  2. Convert any Basic Disks to Dynamic Disks.
  3. Right-click the original source disk, and select the Add Mirror option. Note that this disk can have existing content.
  4. Select the target drive in the configuration wizard, and choose Add Mirror to designate it as the destination.

Data begins to sync immediately, though the process can take some time depending on the speed of the storage devices and the quantity of data. New data written to the storage area goes to both disks. You can expect an improvement in read performance for the RAID 1 array.

Screenshot of existing volume selection for RAID 1
Select an existing volume, and choose Add Mirror.
Screenshot of disk mirroring selection.
Select a disk to mirror the existing volume to.
Screenshot of RAID 1 mirror building
The RAID 1 mirror is building (resyncing) in this image. Notice it is drive letter C:.

Recover from a failed drive in RAID 1

Start the recovery process by replacing the physical drive and then constructing a new mirror. You don't recover a RAID 1 array; you just create a new one.

RAID 5 (disk striping with parity)

RAID 5, or disk striping with parity, is a common storage platform for Windows file servers containing user data. Like RAID 0, it divides storage capacity into stripes and distributes data across them, resulting in a significant performance boost. However, RAID 5 introduces parity information for fault tolerance.

Data parity enables the system to reconstruct lost data from any one drive in the array. If a disk fails, the data on the other drives is combined with the parity information to regenerate the lost content. The approach provides excellent fault tolerance.

RAID 5 cannot reconstruct the data if two drives fail simultaneously. It only handles one drive failure at a time.

When to use RAID 5

RAID 5 balances performance and fault tolerance. It requires at least three disks, but the more devices you add, the better the storage ratio and performance. RAID 5 is particularly useful for standard user data, like home directories, project folders or other file storage repositories.

Configure RAID 5

With Disk Management open, do the following:

  1. Confirm all disks are online and initialized.
  2. Convert any Basic Disks to Dynamic Disks.
  3. Right-click unallocated space on one of the disks.
  4. Select New RAID 5 Volume. The RAID configuration wizard opens.
  5. Select at least three disks for the new RAID 5 array.
  6. Select a drive letter for the storage space.
  7. Select the file system -- probably NTFS.
  8. Select Finish to begin building the array.

Creating a RAID 5 volume uses the same wizard and process as the other RAID types. As with other RAID deployments, use disks of the same size and type for best results. The wizard prompts you for a drive letter and file system format. It then allows you to review the process before building the array.

Recover from a failed drive in RAID 5

You may not notice a Windows Server RAID 5 disk failure right away. The data remains available -- that's the point. However, your users may experience a performance hit. Investigate these concerns using Event Viewer or Performance Monitor, and then begin the recovery process:

  1. Use the Disk Management utility to identify the failed device. It will probably show as Offline or Failed.
  2. Replace the physical disk. Bring it online, and initialize it using Disk Management.
  3. Right-click the RAID 5 volume, and select Reactivate Volume or Repair Volume to recreate the lost data.

Rebuilding the array can be intensive and time-consuming. However, the data is available to users during the rebuild process.

Additional RAID tips

RAID does not replace a solid backup strategy. Instead, it's a key part of a layered approach to protecting and maintaining data and OS availability.

The traditional RAID deployment through Disk Management is an effective option, especially in smaller environments. The newer Storage Spaces approach to managing Windows storage extends RAID's original flexibility by adding new options and greater efficiency.

If you prefer working at the command line or scripting your deployment solutions, use the diskpart command to manage the server's storage settings.

Damon Garn owns Cogspinner Coaction and provides freelance IT writing and editing services. He has written multiple CompTIA study guides, including the Linux+, Cloud Essentials+ and Server+ guides, and contributes extensively to Informa TechTarget, The New Stack and CompTIA Blogs.

Dig Deeper on Storage architecture and strategy