Microsoft Windows Server 2008
Chapter 7
Configuring and Managing Data Storage
Objectives
Understand storage options for Windows
Server 2008
Use the Disk Management tool to configure and
manage storage
Explain and configure RAID disk storage fault
tolerance
Understand storage enhancements in Windows
Server 2008
Back up disk storage
Windows Server 2008 Storage Options
Basic disk
One that uses traditional disk management
techniques and contains primary partitions, extended
partitions, and logical drives
Dynamic disk
One that does not use traditional partitioning
Dynamic disk architecture provides more
flexibility than basic disks
So there is virtually no restriction on the
number of volumes that can be on one disk
Basic Disks
Partitioning
A process that blocks a group of tracks and
sectors to be used by a particular file system, such
as NTFS
Formatting
A process that creates a table containing file
and folder information for a specific file system in
a partition
Volume
A logical designation of disk storage that is
created out of one or more physical disks
Is partitioned and formatted with one file
system
Basic disks recognize primary and extended
partitions
Basic disks also can be configured for any
of three RAID levels:
Disk striping (RAID level 0)
Disk mirroring (RAID level 1)
Disk striping with parity (RAID level 5)
RAID stands for redundant array of
inexpensive (or independent) disks
A set of standards for lengthening disk life
and preventing data loss
MBR and GPT support
When a drive is partitioned, a Master Boot
Record (MBR) and a partition table are created
At the beginning track and sectors on the
disk
The MBR is located in the first sector and
track of the hard disk
Has startup information about partitions
and how to access the disk
The partition table contains information about
each partition created
MBR and GPT support (continued)
Globally Unique Identifier (GUID) Partition
Table or GPT
A newer way to partition disks, without
imposing the same type of limits on the number of
partitions as with MBR
GPT is one element of the Extensible Firmware
Interface (EFI) approach
Offered by the Unified EFI Forum
GPT disks store partition information in each
partition using main and backup tables
Primary and extended partitions on MBR
disks
A primary partition is one from which you can
boot an operating system
At least one primary partition must be marked
as active
Only one primary partition can be active at
a given time
The active partition is the partition where
your computer will look for the hardware-specific
files to start the operating system
Primary and extended partitions on MBR
disks (continued)
An extended partition is created from space
that is not yet partitioned
The purpose of an extended partition is to
enable you to exceed the four-partition limit of a
basic disk
Only one extended partition can exist on a
single basic disk
A computer with multiple partitions boots
from the partition that is designated as the active
partition
Must also be the system partition
Volume and Stripe Sets
Volume set
Consists of two or more partitions that are
combined to look like one volume with a single drive
letter
Stripe set
Two or more disks that are combined like a
volume set, but that are striped for RAID level 0 or
RAID level 5
Dynamic Disks
A dynamic disk does not use traditional
partitioning
Makes it possible to set up a large number of
volumes on one disk
Provides the ability to extend volumes onto
additional physical disks
The number of disks that can be
incorporated into one spanned volume is limited to 32
Plan to convert basic disks to dynamic disks
after you install Windows Server 2008
Simple volume
A portion of a disk or an entire disk that is
set up as a dynamic disk
Can be extended onto multiple sections of the
same disk
Spanned volume
Stored on 2 to 32 dynamic disks that are
treated as one volume
As you add new disks, the spanned volume can
be extended to include each disk
Striped volumes
Often referred to as RAID-0
Extend the life of hard disk drives by
spreading data equally over two or more drives
Another advantage: increases disk performance
In Windows Server 2008, striping requires at
least two disks and can be performed over as many as
32
Data can be lost when one or more disks in the
striped volume fail because the system has no
automated way to rebuild data
Shrinking a volume
Windows Server 2008 comes with the ability to
shrink a basic or dynamic disk volume
Shrinking a volume enables you to create a new
partition when one is needed and you dont have
extra disks
When you shrink a volume, Windows Server 2008
starts from the end of that volume
Works its way back through contiguous space
to create unallocated disk space
You can specify the amount of space to recover
Disk Management
Disk Management tool
Provides a central location for viewing disk
information and performing tasks such as creating and
deleting partitions and volumes
Creating a Partition and Simple Volume
Partitions operate as separate storage
units on a hard disk
The most basic way to create a partition is
to take unallocated disk space
Use the New Simple Volume Wizard to create a
simple volume
You can also delete a partition using the
Disk Management tool
Once a partition is formatted, it is called a
volume and can be assigned a drive letter
Mounting a Drive
Windows Server 2008 enables you to mount a
drive as an alternative to giving it a drive letter
Mounted drive
One that appears as a folder and is accessed
through a path like any other folder
You can mount a basic or dynamic disk
drive, a CD/DVD drive, or a removable drive
Home directory or home folder
A server folder that is associated with a
users account and that is a designated
workspace for the user to store files
Managing Disks
Using Disk Defragmenter
When you save a file to a disk, Windows Server
2008 saves the file to the first area of available
space
The file might not be saved to a contiguous
area of free space
The disk gradually becomes fragmented
The process of defragmenting
Locates fragmented folders and files and moves
them to a location on the physical disk so they are
in contiguous order
Using Disk Check
The Disk Check tool allows you to scan your
disk for bad sectors and file system errors
This tool is meant for use when no users
need to access the files on the disk you want to
check
Because the disk is made unavailable during
the scan for problems
Options:
Automatically fix file system errors
Scan for and attempt recovery of bad sectors
Using chkdsk
You can also check your disk for errors by
running the chkdsk utility from the Command Prompt
window
In NTFS, chkdsk checks files, folders,
indexes, security descriptors, user files, sectors,
and disk allocation units
Introduction to Fault Tolerance
Fault tolerance
The ability of a system to gracefully recover
from hardware or software failure
Windows Server 2008 provides a level of
fault tolerance through software-level RAID
With fault tolerance, data is written to
more than one drive
In the event one drive fails, data can still
be accessed from one of the remaining drives
RAID Volumes
RAID is a set of standards for lengthening
disk life, preventing data loss, and enabling
relatively uninterrupted access to data
RAID level 0
Striping with no other redundancy features is
RAID level 0
RAID level 1
Disk duplexing is the same as disk mirroring,
with the exception that it places the backup disk on
a different controller or adapter than is used by the
main disk
RAID level 2
Uses an array of disks whereby the data is
striped across all disks in the array
RAID level 3
Uses disk striping and stores error-correcting
information, but the information is only written to
one disk in the array
RAID level 4
Stripes data and stores error-correcting
information on all drives
RAID level 5
Combines the best features of RAID, including
striping, error correction, and checksum verification
Windows Server 2008 supports RAID levels 0,
1, and 5 for disk fault tolerance
Using a Striped Volume (RAID-0)
Reasons for using a RAID level 0 or a
striped volume in Windows Server 2008 are to:
Reduce the wear on multiple disk drives by
equally spreading the load
Increase disk performance compared with other
methods for configuring dynamic disk volumes
To create a striped volume, right-click the
unallocated space for the volume and click New
Striped Volume
Only dynamic disks can be striped volumes
Using a Mirrored Volume (RAID-1)
Disk mirroring involves creating a shadow
copy of data on a backup disk
Only dynamic disks can be set up as a mirrored
volume in Windows Server 2008
One of the most guaranteed forms of disk fault
tolerance
Disk read performance is the same as reading
data from any single disk drive
A mirrored volume is created through the Disk
Management tool
Using a RAID-5 Volume
Fault tolerance is better for a RAID-5
volume
A RAID-5 volume requires a minimum of three
disk drives
Parity information is distributed on each
disk
If one disk fails, the information on that
disk can be reconstructed
The parity used by Microsoft is Boolean
(true/false, one/zero) logic
The performance is not as fast as with a
striped volume
Takes longer to write the data and calculate
the parity block for each row
Accessing data through disk reads is as
fast as a striped volume
A RAID-5 volume is particularly useful in a
client/server system that uses a separate database
for queries and creating reports
Use the Disk Management tool to create a
RAID-5 volume
Software RAID vs. Hardware RAID
Software RAID implements fault tolerance
through the servers operating system
Hardware RAID is implemented through the
server hardware
Independent of the operating system
Advantages over software RAID:
Faster read and write response
The ability to place boot and system files on
different RAID levels
The ability to
hot-swap a failed disk
More setup options to retrieve damaged data
Windows Server 2008 Storage Enhancements
For medium to large networks, Windows
Server 2008 offers storage enhancements in two
important realms:
Management of Storage Area Networks
Using multiple paths to storage for fault
tolerance
Storage Manager for SANs
Storage Area Network (SAN)
A grouping of storage devices that forms a
subnet
The storage devices are available to any
server on the main network
Appear to the user as though they are attached
to the server they are accessing
Typically, the subnet containing the
storage devices uses Fibre Channel or iSCSI
technology
Storage Manager for SANs is used to manage the
logical unit numbers for Small Computer System
Interface drives
Small Computer System Interface (SCSI)
A 32- or 64-bit computer adapter that
transports data between one or more attached devices
and the computer
Logical unit number (LUN)
A number that identifies a physical SCSI drive
or logical SCSI targets
A SAN containing multiple drives configured
for RAID will have many LUNs to manage
Storage Manager for SANs is used for SANs that
employ Virtual Disk Service
Virtual Disk Service (VDS)
Used to enable management of disk volumes in
SANs through one interface at a server
Another enhancement to Windows Server 2008
is that the iSCSI initiator is now built into the
operating system
The iSCSI initiator is a driver that enables
Windows Server 2008 to communicate with an iSCSI SAN
Multipath Input/Output Enhancements
Multipath I/O
Provides a means to establish multiple paths
between a server and its disk storage
The first step in this process is to create
the multiple paths between the storage and the server
or servers
For a SAN, creating multiple paths might
involve establishing two or more network paths
through two or more network switches to the SAN
The next step is to install Multipath I/O
A feature installed through Server Manager
Device Specific Module (DSM)
Compatible with the following disk storage
array controller technologies:
Asymmetric logical unit access (ALUA)
Active/Active controller model
Configuration models:
Dynamic Least Queue Depth
Failback
Failover
Round Robin
Round Robin with a subset of paths
Weighted Path
Disk Backup
One of the best ways to make sure you do
not lose valuable information on a hard disk is to
fully back up information on a regular basis
These backups can be performed from the server
or from a workstation on the network
Performing backups from a backup device
installed on the server has several advantages:
No extra load is produced on the network
Perform backups on a multiple-server network
Provides more assurance that the Registry is
backed up
The advantages of performing a network
backup
Backup jobs can be stored on a single backup
media
One administrator can be responsible for
backing up multiple servers
The main disadvantages
The increase in network traffic
The Registry cannot be backed up from across
the network
Windows Server Backup
Windows Server Backup tool offers the
ability to back up all server files or files that
have changed
Enhancements in Windows Server 2008:
Is easier to recover from a backup
Has more backup options, including using the
Volume Shadow Copy Service (VSS)
Is more reliable in recovering applications
Provides information about disk use
Offers the wbadmin command-line tool
Has full support to back up to optical media
Considerations for using Windows Server
Backup
Tool only backs up NTFS volumes
Tool does not back up to tape
If you have backup media made from Windows
Server 2003 using Ntbackup.exe, you cannot restore
from that media using the Windows Server Backup tool
in Windows Server 2008
Backup Options
Full backup
A backup of an entire system, including all
system files, programs, and data files
Changes each files archive attribute to
show that it has been backed up
Incremental backup
Only backs up files that are new or that have
been updated
Backs up only files that have the archive
attribute marked
Custom backup
Enables you to configure backups differently
for each volume
Such as doing an incremental backup every
time you back up the C drive and a full backup each
time you back up the D drive
Scheduling Backups
Windows Server Backup includes a scheduling
capability
Can have the server automatically start
backups after regular work hours or at a specific
time of day
Configuring Backup Performance
Configuring the backup performance options
enables you to specify which types of backups to
perform: full, incremental, and custom
The default is to always perform full backups
Performing a Recovery
The Windows Server Backup tool enables you
to recover any of the following:
Files
Folders
Volumes
Applications and application data
The backup catalog (of information in the
backup)
The operating system (to the same computer or
to another computer using identical hardware)
Before you start, determine the following
information:
Date of the backup from which to recover
Type of recovery, such as files and folders or
applications
What to recover
Where to recover, such as in the original
location or another location
Summary
Windows Server 2008 uses basic and dynamic
disks
Dynamic disks can be configured as simple,
spanned, striped, mirrored, and RAID-5 volumes
If you need to recover space from a basic or
dynamic disk, you can shrink the disk
The Disk Management tool enables you to create
basic and dynamic disks
For optimum disk performance, plan to set up a
schedule to regularly defragment disks on a server
Use the Disk Check and chkdsk tools to find
and repair disk problems
RAID provides fault tolerance for hard disks
RAID level 0 is disk striping
With disk mirroring or duplexing (RAID level
1), the same data is written to a partition on each
of the two disks included in the mirror
With RAID level 5, data is written across a
minimum of three disks
Two important enhancements for Windows Server
2008 include new features for Storage Manager for
SANs and features for Multipath I/O
Windows Server Backup offers features to
schedule backups, perform full or incremental backups
(or a combination of both), and recover data from
backups
|