In Windows Server 2012, we can manage disks with three different tools:
- Server Manager
- Disk Management (diskmgmt.msc)
- diskpart.exe (command line)
I'm going to use "Disk Management". When I open the tool, by entering diskmgmt.msc at the command line (or run box in other versions of Windows), this is what I see:
First, we have one physical disk, "Disk 0", divided into two partitions: System Reserved and (C:).
Second, we have a second physical disk that has not yet been configured.
Before configuring this second disk, let's examine the different choices we have for disk configuration in Windows Server 2012. Many of these options will apply to Windows 2008 as well (ReFS is one notable exception).
Disk Configuration options
In Windows Server 2012, physical disks and logical partitions can be configured in a number of different ways:
Partition Style - or Scheme
This can be either the traditional MBR (Master Boot Record) scheme, available since the 1980s, or the more recent GPT (GUID partition table), available as an option since the late 1990s. GPT was first incorporated in Windows Server systems with Windows Server 2003 SP1.
A detailed presentation of these partition schemes is beyond the scope of this post. I'll present below what I believe are the essential points for disk configuration in Windows 2012.
- MBR supports a maximum of four primary partitions. GPT allows for a maximum of 128 partitions. MBR does allow for the creation of more logical disks in an "extended partition" but this type of partition is apparently more prone to errors.
- MBR supports a partition size of 2 TB (terabytes). For decades, this was more than sufficient. As hard drive sizes now surpass 2 TB, MBR is becoming obsolete. GPT supports a partition size of 9.4 ZB (zetabytes) or... 9.4 billion TB. In reality, maximum size will be much lower because Windows Server 2012 supports a maximum volume size of "only" 18 EB (exabytes) and hard drives on the market are simply not that large anyway.
- GPT cannot be used for the boot partition unless the server (in this case) is a "UEFI based system" (as opposed to a traditional "BIOS").
- UEFI is not a requirement if GPT is used on a simple storage partition (a partition from which the operating system does not boot).
If the terminology used above is not clear (UEFI versus BIOS), please refer to online sources for clarification. Here is one source that I consulted:
Microsoft TechNet or Wikipedia articles could also be used to clarify in greater details "MBR", "GPT", "BIOS", "UEFI" and even "boot partition".
In summary, however, if a Windows server system uses "UEFI", GPT can be used for all partitions, including the boot partition. If not, and unless the server in question has an operating system that precedes Windows 2003 SP1, GPT can be used for storage partitions. GPT for the boot partition also assumes a 64 bit operating system which is a "given" for Windows Server 2012 (as for Windows 2008 R2, there is no 32 bit version).
Disk Type (basic versus dynamic)
The "basic" disk is the default. The "dynamic" disk allows the configuration of different types of volumes such as spanned or striped, or two types of "software RAID", RAID 1 (mirroring) or RAID 5 (striping with parity).
Personally, I see limited advantages in dynamic disks:
- Spanning volumes may resolve disk space issues but also (like striped volumes) increases the risk of data loss. The failure of any disk comprising the volume results in the loss of data. The more disks comprising the volume, the greater the risk of data loss.
- If the data is worth protecting, it would be preferable, by far, to use hardware RAID with a high quality controller on the server itself.
Volume Types
We can have the following volume types in Windows Server 2012:
- Simple
- Spanned
- Striped
- RAID 1 (mirrored)
- RAID 5 (striped with parity)
I've already commented on what I perceive as the limited usefulness of dynamic disks in a production environment where protection of data is paramount.
File Systems
In Windows Server 2012, we have three options:
- exFAT
- NTFS
- ReFS
Only two - NTFS and ReFS - are serious choices.
ReFS means "Resilient File System" and is designed to be even more robust than NTFS. It is an excellent choice for data storage. However, it does (currently) have some limitations.
- No EFS encryption (it is compatible with Bitlocker however).
- No compression
- No quotas
More importantly, it is (currently) incompatible with Active Directory to the extent that the Active Directory database, log files and SYSVOL folder should not be stored on a ReFS volume.
Furthermore, it has limited compatibility with Hyper-V since CSV must be disabled.
Windows Server 2012: Does ReFS replace NTFS? When should I use it?
Configuration of Disk 1 (the second hard drive)
Now that we are aware of the different options for disk configuration, let's configure the second hard drive of our server.
The second physical disk is offline so the very first step is to bring it online. We can simply right-click on the "Disk 1" icon and select "Online":
Now the disk is "Not initialized". Moreover, disk type is unknown and of course, there is no file system. Let's configure the disk. Right-click and select: "Initialize Disk". Here we must chose between two partition styles: MBR and GPT:
I'll select "GPT".
The status of the disk changes to "Online" but it is still "unallocated".
We'll right-click and select "New Simple Volume".
Spanned and stripped volumes increase risk of data loss and should be used with caution (or not at all). RAID 1 is an option and RAID 5 would be if there was a 3rd disk. Generally, however, "hardware RAID" is preferred to Microsoft's software RAID and the latter would typically be used as a last option.
The New Simple Volume Wizard opens. Click on next.
We specify the volume size. For this example, I'll use the entire physical disk.
Click on Next. This brings us to the "Format Partition" page.
We'll assign drive letter E:
We have to select a file system. The choice is currently between NTFS and ReFS. It would be extremely rare that one would opt for exFAT. I'll select NTFS for this example. We will keep the default "Allocation unit size". This could be changed if required or recommended for a particular application. For an Exchange 2010 DAG (Database Availability Group), the recommended allocation unit size is 64 KB. A quick format is fine for our practice environment (a "long" format would check for bad sectors which might be recommended if the drive will hold important data).
We can click on next and the following summary displays:
Here is the result in Disk Management...
And in Windows Explorer:
Thank you for reading the article about Windows Server 2012 - Disk Management on the blog NEW TECH If you want to disseminate this article on please list the link as the source, and if this article was helpful please bookmark this page in your web browser by pressing Ctrl + D on your keyboard keys.