3. Creating Volumes
Windows SBS 2011 supports five volume types, as described in Table 1.
These volume types provide varying amounts of fault tolerance and
performance enhancement. The volumes you created during the Windows SBS
2011 installation are simple volumes, and because the volumes contain
the computer’s boot and system files, they must remain that way.
However, you can use any unallocated space on the system disk, plus the
space on your other disks, to create volumes of any type.
Table 1. Volume Types Supported by Windows SBS 2011
|
VOLUME TYPE |
NUMBER OF DISKS REQUIRED |
FAULT TOLERANCE |
DESCRIPTION |
|---|
|
Simple |
1 |
None. |
Consists of space from a single disk. |
|
Spanned |
2 to 32 |
None. The loss of one disk destroys the volume. |
Consists of space from multiple dynamic disks, combined to create a single volume. |
|
Striped |
2 to 32 |
None. The loss of one disk destroys the volume. |
Consists of space from multiple dynamic disks, combined to create a
single volume. The system writes data one stripe at a time to each
successive disk in the volume, increasing input/output (I/O) efficiency. |
|
Mirrored |
2 |
Yes. The volume survives the loss of one disk. |
Consists of an identical amount of space on two disks. The system
writes duplicate copies of all data to both of the disks, so that all
the data remains available if one disk fails. The volume size therefore
equals half of the total disk space. |
|
RAID-5 |
3 or more |
Yes. The volume survives the loss of one disk. |
Consists of space on at least three disks. The system stripes data,
along with parity information, across the disks. If one disk should
fail, the system can reconstruct the missing data using the parity
information on the other disks. |
In addition to the volume types, Windows SBS 2011 supports two disk types: basic disks and dynamic disks. When you first initialize a disk, it appears as a basic disk. A basic disk can have up to four partitions: three primary partitions and one extended
partition, with the extended partition hosting multiple logical drives.
As long as you create only simple volumes on a disk, it remains a basic
disk. The first three simple volumes you create on a basic disk are the
primary partitions.
When you create a fourth simple volume, the Disk Management snap-in
creates an extended partition using all the remaining unallocated space
on the disk, and a logical drive of the size you specified for the
simple volume, as shown in Figure 2.
Tip
It is possible to create a fourth primary partition on a basic disk,
but you cannot do it using the Disk Management snap-in. Instead, you
must use the Diskpart.exe tool from the command prompt.
To create volumes other than simple volumes, you must convert the disks into dynamic disks. Technically, a dynamic disk
has one partition that uses all its available space. You can then
create multiple volumes within that single partition. When you use
basic disks to create a spanned, striped, mirrored, or RAID-5 volume
using the Disk Management snap-in, the tool converts the disks from basic to
dynamic automatically. If you delete all the volumes on a dynamic disk
(erasing the data on them), Disk Management automatically converts it
back to a basic disk.
The Disk Management snap-in enables you to convert a basic disk into
a dynamic disk manually, by right-clicking a disk and using the
interface shown in Figure 3, although this usually is not necessary.
When you convert a basic disk with partitions on it to a dynamic disk, the snap-in converts each partition to a simple volume, as shown in Figure 4.
There is a corresponding command to convert a dynamic disk back to a
basic disk, but it only appears when there are no volumes on the
dynamic disk. Therefore, the only way to convert a dynamic disk to a
basic disk is to delete all its data.
Caution
The only drawback to using dynamic disks is that you cannot boot the computer from any volume on a dynamic disk other than the current boot volume. This prevents you from using a technique called dual-booting
to run two operating systems on the same computer. This is not a
serious problem for most users, as virtualization technology has
rendered dual-booting obsolete in most cases.
To create a simple volume with the Disk Management snap-in, perform the following procedure:
-
Click Start. Then click Administrative Tools > Computer Management. The Computer Management Console appears.
-
In the Scope (left) pane, select the Disk management node.
-
In the graphical view, right-click the unallocated space on any of your disks; from the context menu, select New simple volume. The New Simple Volume Wizard appears.
-
Click Next to bypass the Welcome page. The Specify Volume Size page appears.
-
In the Simple volume size in MB box, specify the size of the volume you want to create and click Next. The Assign Drive Letter Or Path page appears.
-
Select one of the following options and configure its properties:
-
Assign the following drive letter Enables you to select the available drive letter you want to use to access the volume.
-
Mount in the following empty NTFS folder
Enables you to access the new volume from within a folder on another
existing volume. This makes the new volume appear in the specified
folder, as though it is part of the existing volume.
-
Do not assign a drive letter or drive path
Enables you to create the new volume without any means of accessing it.
You can assign a drive letter or mount the volume to a folder at a
later time.
-
Click Next. The Format Partition page appears.
-
To format the volume, leave the Format the volume with the following settings option selected and configure the following properties:
-
File system
Specifies whether you want to format the volume using the NTFS, FAT,
FAT32, or exFAT file system. NTFS provides many more features than any
of the file allocation table (FAT) options, including access control,
compression, and encryption. The only compelling reason to use FAT or
FAT32 is if you must access the volume using an operating system that
does not support NTFS. The exFAT file system is a new, 64-bit version
designed for use on USB flash drives.
-
Allocation unit size
Specifies the size of the individual clusters the system uses when
allocating space on the volume. For a volume that stores a great many
small files, a smaller value provides more efficient use of disk space;
for a volume that stores extremely large files, a greater value
provides better performance. For most applications, however, the
default setting is preferable.
-
Volume label Specifies a name, up to 32 characters long, that the system uses to identify the volume.
-
Perform a quick format
Selecting this check box causes the wizard to format the volume without
checking for errors. Clearing the check box causes the wizard to check
the disk for errors as it formats the volume and mark bad sectors as
inaccessible. The time required for the error checking depends on the
size of the drive. For large drives, it can be a lengthy process.
-
Enable file and folder compression
Selecting this check box along with the NTFS file system causes the computer to store all the data it writes on the volume in a compressed form. This provides additional storage
space on the disk, but also places an additional burden on the system
processor, which is responsible for performing all of the real-time
compressions and decompressions.
Note
You can turn compression on and off for a volume at any time, from the General tab on the volume’s Properties sheet.
-
Click Next. The Completing The New Simple Volume Wizard page appears.
-
Click Finish. The wizard closes, and the new volume appears in the unallocated space.
Depending on the computer’s hardware, the size of the volume, and
the settings you selected, it might take several minutes for the
formatting to finish.
Creating Other Volume Types
The Disk Management snap-in also includes wizards for creating spanned, striped, mirrored, and RAID-5 volumes.
These wizards are identical to the New Simple Volume Wizard, except for
the addition of a Select Disks page. This page enables you to select
the disks you want to use to create the volume and specify the amount
of space you want to use on each disk.
When the wizard finishes creating a volume, the allocated space on
all of the disks involved changes color to reflect the properties of
the new volume, as shown in Figure 5.
When creating a spanned disk, as shown in Figure 6, you must select at least 2 and no more than 32 disks, and you can specify any amount of space on each disk.
The disk you selected when creating the volume appears already in the Selected
list, with all its available disk space allocated. You can modify the
amount of space you want to devote to the drive as needed. The Available list contains all the other disks in the system with available space. After selecting one or more disks and clicking the Add button, you can modify the amount of space allocated from each additional disk as well. The Total volume size in megabytes (MB) indicator specifies the total amount of space on all the disks you selected.
When creating a striped volume, as shown in Figure 7, you must select at least 2 and no more than 32 disks, and you must specify the same amount of space on each disk.
The process of selecting disks for a striped volume is essentially the same as that for a spanned
volume, except that the amounts of space allocated on each disk
automatically adjust themselves as needed. For example, if the disk you
selected when creating the volume has 40,000 MB free, and you add a
disk that has only 20,000 MB free, the wizard automatically adjusts the
configuration to use 20,000 MB from both disks because all the selected
disks must contribute the same amount of space. As with a spanned
drive, the Total volume size in megabytes (MB) indicator specifies the combined allocated space in all the selected disks.
Caution
As mentioned earlier, it is critical to understand that spanned and striped volumes do not add fault
tolerance, and are, in fact, less fault tolerant than the other volume
types. The loss of one disk in a spanned or striped volume causes the
entire volume to be lost, including the data on other undamaged disks.
Creating Mirrored Volumes
When creating a mirrored volume, as shown in Figure 8, you can select only two disks, and you must specify the same amount of space on each disk.
The Select Disks page in the New Mirrored Volume Wizard functions
just as when you created a striped volume, except that you can select
only two disks, and the Total volume size in megabytes (MB)
indicator specifies the amount of space allocated on each disk, not the
sum of both. A mirrored volume stores two copies of each file, thereby
realizing a volume size that is half of the total allocated space.
Note
A mirrored disk array enables the computer to survive the failure of
one disk with service interruption or data loss. Another variation on
this technology is to create a mirrored array using two disks that are
connected to different host adapters. This technique, called disk
duplexing, enables the computer to survive the failure of a disk or a
host adapter.
When creating a RAID-5 volume, as shown in Figure 9, you must select three or more disks, and you must specify the same amount of space on each disk.
The system provides fault tolerance by calculating a parity value for each bit in the array. A parity value
is a sum of the corresponding bits on each disk in the array. Every bit
has a binary value of 0 or 1, so the sum of all the bits in a
particular location on the array must be 0 or 1. (This is why all the
disks in the array must have the same amount of space.) If one of the
disks in the array is lost, the system can use the parity information
to recalculate the value of the missing bits.
For example, in a 4-disk array, the values of the first bit in each of the first 3 disks might be 0, 1, and 1, as shown in Figure 10.
Adding them together (in binary) gives you a parity value of 1,
which the system writes to the first bit in the fourth disk, as shown
in Figure 11.
If the second disk in the array fails, as shown in Figure 12, one of the four corresponding bits is lost.
When this occurs, the system can look at the values of the first bit
in the remaining disks (0 and 1) and the value of the parity bit (1),
and know that the value of the missing bit must be 1, as shown in Figure 13.
(If the parity bit were 0, the missing bit would have to be 0.) When
the malfunctioning drive is replaced, the system calculates the value
of each missing bit and reconstructs the data on the new disk.
In this way, a RAID-5 array can yield a greater amount of available storage space than a mirrored array. On a RAID-5 volume, the size of the volume specified by the Total volume size in megabytes (MB)
indicator is based on the size of the disks, multiplied by the number
of disks in the array, minus 1. Thus, a RAID-5 array that consists of
four 100 GB drives will have a volume size of 300 GB.
100 x (4 – 1) = 300
A RAID-5 volume can suffer the failure of any one disk without data
loss. When a disk fails, overall performance of the array suffers, more
so while the array is reconstructing the data on a newly installed disk.
The RAID-5 capability provided by Windows Server 2008 R2 is known as software-based RAID
because the system processor is responsible for performing all of the
bit parity calculations. Administrators who are serious about RAID data protection might want to consider a hardware-based RAID
solution, instead. Hardware-based RAID uses a separate host adapter
with its own processor to offload the bit parity calculations from the
system processor. This provides greater efficiency without
overburdening the computer’s main processor.
