Home > 70-410 Installing and Configuring Windows Server 2012 > 70-410 Windows Server Local Storage
Welcome to the ITFreeTraining video on MBR and GPT partition tables. A partition table forms the low level data structure of a drive. By the end of this video, you will understand what partition tables are available and what features each offer.
Download the PDF
0:20 Before a drive can be used by an operating system, whether it be physical or virtual, a partition table needs to be created on the drive. The partition table forms the low level data structure of the drive. This structure allows the drive to be divided up into individual partitions. If I were to have a physical drive that has a partition table created on it, the drive will contain only unused space. Before the unused space can be used, the drive needs to be divided up into one or more partitions. In this case, the drive will be divided up into 3 partitions. In many cases, it is common for a hard drive to be divided up into one large, main partition for the operating system and its data, and some smaller partitions for booting, recovery, and tools. After creation, a partition table can be formatted for the operating system to store data on. There are two partition tables that are available. These are MBR and GPT. MBR is older and this has more compatibility with older and obsolete operating systems and hardware. GPT is a newer standard and has more features but requires newer operating systems and, depending on what you are trying to achieve with it, it may also require newer hardware. To start with, I will look at MBR.
Master Boot Record (MBR)
1:38 The MBR or Master Boot Record, was first introduced on PC’s back in 1983. Since then it has become the de facto standard for PC’s working on DOS based, Windows based, and alternative operating systems like Linux. If the MBR partition table is used, this allows the drive to be divided up into a maximum for 4 primary partitions. The operating system allows each partition to be formatted to use a file system. So with MBR, the administrator could create 4 primary partitions and format them so the operating system can see 4 drives. In most cases, this should be sufficient for most requirements. However, in some rare cases it may be necessary for more than 4 partitions. If this is the case, one of the primary partitions can be changed into an extended partition. In the DOS days, this extended partition could hold 23 partitions. This was a limitation of DOS since it could only support 26 drive letters. On modern Windows operating systems, the administrators can create as many extended partitions as they wish. However, extended partitions are linked to each other so if you use a lot of them they could have an impact on performance. Also, the BIOS in the computer may not support booting from an extended partition; however, a workaround would be to place the boot loader on a primary partition that would boot the operating system stored in the extended partition. The MBR partition table offers the fundamental features required for an operating system and offers compatibility with older operating systems and hardware. The only limitation is that it is limited to 2 Terabytes of usable space. If the administrator uses MBR on a drive larger than 2 Terabytes, all the space after Terabytes will be unavailable to the operating system and thus unusable. There is some new technology that has been developed to extend the life of MBR which I will look at later on in this video. Before this, I will have a look at MBR’s replacement, GPT.
GUID Partition Table (GPT)
3:50 The replacement for MBR is GPT or GUID Partition table. Using Windows, GPT support 128 partitions. To be more specific, Windows uses the minimum default settings using GPT. The standard for GPT allows an unlimited number of partitions. So using an alternative operating system, the administrator could potentially have access to more than 128 partitions. Why you would need more than that, I can’t really say. The biggest difference with GPT is that it supports drives up to a Zettabyte. It is not limited to 2 terabytes like MBR. The actual amount may vary a little depending on how large the drive manufacturer makes the sectors. Currently, as there are only terabyte drives on the market, it’s not possible to reach a Zettabyte. First a Petabyte and Exabyte drive would need to have been created before we could even get close to a Zettabyte hard disk.
In order to boot an operating system using GPT, a few requirements need to be met. First the computer needs to support Unified Extensible Firmware Interface or UEFI. UEFI was designed to be a replacement to BIOS; however, UEFI generally does require 64-bit (x64) hardware. It is possible to run UEFI on 32-bit (x86) hardware but manufacturers generally don’t use UEFI on 32-bit systems. To boot Windows using a drive with a GPT partition table, you will need Windows XP professional 64-bit (x64) or Windows Server 2003 64-bit (x64) or above. If you are using Linux, most modern 32-bit and 64-bit Linux systems will support booting from GPT hard disks or the GPT hard drive being used as a data drive. You will need to check the details of your particular distribution to see it if is supported. If your computer does not meet these requirements, it may still be possible to use a GPT drive as a data drive. For Windows, this requires Windows Vista or Windows Server 2003 with Service Pack 1 or above. Data drives using GPT do not require a 64-bit version of Windows or 64-bit hardware. Many Linux distributions support GPT drives as a data drive, however you need to check the individual distribution to confirm it is supported. You can see GPT addresses the issue of MBR only being able to be used on drives up to 2 Terabytes in size. However, in order to do this, it requires hardware and software support. So what do you do if you need to use more than 2 Terabytes on MBR?
6:44 To improve support for larger drives with older hardware, Advanced Format was developed and is available for use with all hard drives manufactured since 2011. This potentially allows hard disks larger than 2 Terabytes to be used on older hardware and software that does not support GPT.
To better understand how this advanced format works, I will first look at how it works without advanced format. When an MBR partition table is used, it uses 32 bits to hold its data. The largest sector that can be used is 1512 bytes. This gives us a total limit of 2 Terabytes in size. You can see this is where the 2 Terabyte limit comes from. A limit imposed by MBR of 32 bits and a sector size of 512 bytes.
So how does the advanced format bypass this limit while still working with old hardware? When using a drive with advanced format, you still have a limit of 32 bits and 512 bytes, this is what the older hardware expects and supports. However, the firmware on the drive allocates each block of data as eight 512 byte sectors. Because of this the hardware is accessing a 512 byte sector but in reality it is accessing eight 512 byte sectors making the sector 4 kilobytes in size. The end result is that the old hardware can be fully utilized up to 16 Terabytes of data on the drive. This allows larger drives to be used with old hardware but it is not without some problems.
8:32 On the older operating systems, the drive will need to be partitioned with Hybrid MBR in order to use more than 2 Terabytes. Unfortunately, Windows does not have the ability to format a drive in this way, you will need to format the drive using an alternative operating system. It may be possible, for example, to format the drive in Linux and use it in Windows as a data drive. To use the drive for booting is difficult, but not impossible to implement. Other operating systems may support this format, but you would need to perform your own checks to see if it does and how well it is supported. Many disk utilities will expect a 512 byte sector and will not understand what is happening when they encounter the larger 4 kilobyte sector. Even attempting to *use* a disk utility that doesn’t support this format may damage the data on the drive as the disk utility tries to correct a problem that is not there. This applies to Linux and Windows utilities.
Converting Between MBR and GPT
9:36 In the real world, if you have a drive less than 2 Terabytes, MBR should meet all your needs. There is no need to change to GPT. If, however, you do find that you need to convert between the two, it is possible in Windows; however, this will require all the existing partitions to be deleted which will erase all the data on the drive. If you are using Linux, it is possible to convert the partition table without losing data; however, you should always backup your data beforehand just in case. Lastly, there is 3rd party software available that will perform the conversion if you are willing to pay for it. Paragon offers free software that will do the conversion between MBR and GPT for home use. You may also look at some of the free Linux distributions to achieve the same result.
In a next video, I will look at how disk management with MBR and GPT works in Windows. I hope you enjoyed the video and found it useful and hope to see you in next video. Bye for now.
“Installing and Configuring Windows Server 2012 Exam Ref 70-410” pg 42 - 43
"GUID Partition Table" http://en.wikipedia.org/wiki/GUID_Partition_Table
"Partition table" http://en.wikipedia.org/wiki/Partition_table
"Extended boot record" http://en.wikipedia.org/wiki/Extended_boot_record
"Windows support for hard disks that are larger than 2 TB" http://support.microsoft.com/kb/2581408
"GUID Partition Table" http://en.wikipedia.org/wiki/GUID_Partition_Table
"Advanced Format" http://en.wikipedia.org/wiki/Advanced_Format
"Windows and GPT FAQ" https://msdn.microsoft.com/en-us/library/windows/hardware/dn640535(v=vs.85).aspx#gpt_faq_how_many_partitions
Script: Phil Guld http://philguld.com
Voice Talent: HP Lewis http://hplewis.com
Companion Document, Website and Social Media Manager: Phil Guld http://philguld.com
Video Production: Kevin Luttman http://www.KevinLuttman.com
Quality Assurance and Web Hosting: Edward M http://digitalmaru.com
This video will look at how to configure drives to use the GPT and MBR partition tables on Windows 8 and Windows 2012. This will be done using the GUI and the command line.
Download the PDF handout
Demonstration Windows 8
1. The disk management tool can be run by itself or as part of computer management. To open computer manager, open Windows Explorer, right click on computer and select the option manage.
2. Disk Management can be found under storage in Computer management. This tool will show you all the drives that are currently connected to that computer and allow you to perform administration on them like creating partitions and formatting.
3. When a new drive is added to Windows, this will appear in Disk Management as not initialized. The process of initializing the drive writes a signature on the drive. This means that if the drive is moved to a different controller inside the computer, Windows will recognize this as the same drive even though it is on a different controller. If the drive is moved to a different computer, Windows will be able to recognize this as a foreign drive to that computer and is up to administrator to decide if they want to use it.
4. To initialize a drive, right click the drive and select the option Initialize Disk. This will show you all the drives on the system that are currently not initialized. The process of initializing the drive also creates a partition table. On this screen you will need to decide if you want to use the MBR or GPT partition table. If you have multiple drives in the system that have not been initialized, it is just a matter of selecting the drives that you want initialized and they will all be initialized at once.
5. If the drive data does not have any partitions created on it, you can right click the drive and you will get the option to convert it to GPT or MBR - the opposite of the current partition table used on that drive at the time. If partitions are on the drive, the option will be grayed out and you will need to delete these partitions before you can use this option.
6. If you want to see which partition table that is currently being used on that drive, right click the drive and select properties. In the properties, select the volume tab. On the volume tab the partition table will be listed next to partition style.
Demonstration Windows Server 2012
1. To access disk management in Windows Server 2012, open Server Manager and then select Computer Management from the tools menu. Once in computer management select Disk Management under Storage.
2. Even though Disk Management will essentially work the same in any editions of Windows, Windows Server 2012 has a policy that is set differently than on the desktop versions of Windows. When a new drive is added to Windows Server 2012 you will see a message stating “Offline (The disk is offline because of policy set by an administrator”. In order to use the drive, the administrator will have to right click the drive and select online. This is to prevent a drive accidently being added to the server and causing unexpected results. Essentially it adds an extra step and adds a little bit more security to the server. The biggest reason for this is the SAN (Storage Area Network) could share the same drive to multiple servers. If two servers access the same drive at the same time this can cause problems and is not recommended when it can be avoided.
3. To change the disk policy in Windows Server 2012, this needs to done using DiskPart. This can be run by opening a command prompt and running DiskPart.
4. To see the current SAN policy in DiskPark run the command SAN. To change the policy enter in SAN POLICY=OnlineAll or SAN POLICY=OfflineShared.
5. Disk Management can also be run as a separate application rather than as part of Computer Management. To do this, run diskmgmt.msc from run.
6. If you right click a drive on Disk Management you have the option to online or offline the drive. If the drive is online, you also have the option “Initialize Disk” if the disk has not been initialized before. You will then be asked if you want to use an MBR or GPT partition table.
1. “List Disk”. This will show all the disks that are currently connected to that system.
2. “Select Disk n”. This command will select a disk in the system where n is the drive. To obtain the drive number, use the list disk command.
3. “convert gpt”. This will convert the currently selected drive to the GPT partition table.
4. “convert mbr”. This will convert the currently selected drive to the MBR partition table.
5. “list disk”. This will show all the disks in the system. If you have a disk selected using the select disk command, the currently selected disk will have an asterisk to the left of it.
6. “clean”. This command will remove all data, partitions and signatures on the drive. Effectively restoring the drive back to factory defaults. This command does not give any warning messages so be careful you have the right drive selected before you run it.
“Installing and Configuring Windows Server 2012 Exam Ref 70-410” pg 42-43
This video looks at the four file systems supported by Windows. These are ReFS, NTFS, FAT and exFAT. The video looks at what each file system is capable of and its limitations.
Download the PDF handout
Resilient File System (ReFS)
The Resilient File System is a new file system built from scratch by Microsoft. Since it is a new file system it requires Windows 8 or Windows Server 2012 in order to operate. The main design difference between it and previous operating systems is that it is designed to fix problems while the operating system is online. For this reason the check disk feature that is found in previous operating systems that can be run to fix problems no longer exists. Given a new approach has been taken in the operating system, it is better at ensuring data integrity and corruption than previous operating systems.
ReFS was designed to replace NTFS, but at the present time there are some limitations which may mean that you will need to stay with NTFS.
Disk quotas: Disk quotes are not supported. Microsoft states in a blog post that this is a feature that can be supported outside the file system so it is possible for this feature to be supported in software. Possibly Microsoft will add this feature later on or some 3rd party software is available that will add this feature.
NTFS compression and EFS: File compression and encryption (Encrypting File System) are not supported.
Hard links: Hard links are not supported which is required by data duplication. Data duplication allows a file in the drive to be stored in one location but referenced in many different locations. It is used in the Windows Side by Side store (used to store software and drivers in the Windows directory) and may be the reason why ReFS cannot be used for the operating system at this time.
Booting the OS: The biggest limitation is that ReFS does not support being used to boot the operating system. For the present, the drive containing the operating system will need to use a different operating system.
Named streams/Object ID’s: Named stream allows a name to be given to a file stream by the software. This is generally a feature that is used on alternative operating systems not Windows systems. Object ID’s allow a file to be given a number. This is generally used for indexing software.
Extended attributes/8.3 Filenames: Extended attributes are also not supported. Extended attributes is a feature that allows software to add additional attributes to a file in addition to the standard read only, hidden, archive and system. This feature is mostly used on alternative operating systems. It is rare for Windows software to use these features. 8.3 filenames are not supported. This dates back to the MS-Dos days. Windows 95 introduced long file names.
Software support: Because ReFS also does not support named streams and object ID’s. Some software does make use of these features so it is important to check all software before using it. For example SQL does not fully support ReFS.
Convert from NTFS: Currently there is no software available which will convert a drive for NTFS to ReFS or ReFS to NTFS. If you want to change file systems you will need to back up your data, reformat the drive and then restore the data to the drive.
1) To format a drive to use ReFS, right click the drive in Windows Explorer and select the option ReFS from the file systems menu.
2) To view the properties for a file, right click the file and select the properties option. In ReFS the only file attributes that are available are Read-only, Hidden and Archive. If you compare this with the properties of a file stored on a NTFS drive, the properties will have the following which are not found on ReFS. “Allow this file to have contents indexed in addition to file properties”, “Compress contents to save disk space” and “Encrypt contents to secure data”. The option “File is ready for archiving” is effectively the archive attribute found in ReFS.
Enabling on Windows 8
On Windows 8 you will not by default be able to format a drive to use ReFS, however Windows 8 will be able to read a drive formatted with ReFS. If you want to enable Windows 8 to be able to format a drive with the ReFS file system you need to add the following to the registry.
Create DWORD with the value of 1 in the following location to enable ReFS for Windows 8.
New Technology File System (NTFS)
This file system was first introduced in 1983 with Windows NT. It differed from the previous operating system in that it supported security. It also has additional features like compression and encryption. File systems like NTFS, ReFS and Linux system Ext3 use a transaction log. When changes are made these changes are stored in a log file before being written to the drive. This does slow the drive down a little, however it makes the drive less prone to data corruption and data loss, in particular during power outages. Since NTFS has been introduced there have been many different versions of NTFS offering additional features. Features like self-healing have been added, however these are limited to what they can do due to their limitations by the original design.
File Allocation Table (FAT)
The file system FAT was first introduced in Dos. The FAT32 system that is found in modern Windows is an improved FAT system of the original FAT file system. Since FAT has been around for such a long time it is compatible with almost all operating systems including alternative operating systems. It does not offer additional features like security, compression or encryption. Since it is a basic operating system with a lot of compatibility it is often used in removable media. For example USB thumb drives.
Extended File Allocation Table (exFAT)
This file system was created by Microsoft. For this reason it is proprietary and patented. In order for a devices like a TV to use this file system they need to pay Microsoft a fee. Support for it is available in Windows XP and above, however you may need to apply some updates in Windows XP to use the file system. The file system itself was not designed to be used as a replacement to NTFS or other file systems as it does not support features like file compression and encryption. The file system itself is optimized for flash drives. When comparing it to FAT32, it supports big drives and offers better performance, however is not supported on as many devices or operating systems.
File System Comparisons
When deciding which operating system to use, you need to decide which features you want and which operating system you want the operating system to be able to be used in. If you require encryption or compression you will need the NTFS file system. If you do not require these, the FAT32 file system offers the most compatibility being supported since Dos version 7. Fat32 is limited to 4 Gigabytes max file size and 2 Terabytes max drive size. For this reason and the reasons of compatibility, it is often used for removable media like flash drives. In some references they may state that in ReFS the max file size is 32k (32768 bytes) however when released this was reduced to 255. This was most likely done for compatibility reasons.
To Format a drive in Windows, right click it and select the format option. Window will choose the file system that it thinks is most appropriate. You are free to change this file system to what you want. In Windows 8 client operating system and above, the file system ReFS is not available. To make it available you need to configure the following DWORD in the registry.
Create DWORD with the value of 1 in the following location to enable ReFS for Windows 8.
“Installing and Configuring Windows Server 2012 Exam Ref 70-410” pg 45
“Resilient File System Overview” http://technet.microsoft.com/en-us/library/hh831724.aspx
“Building the next generation file system for Windows: ReFS” http://blogs.msdn.com/b/b8/archive/2012/01/16/building-the-next-generation-file-system-for-windows-refs.aspx
"Comparison of file systems" http://en.wikipedia.org/wiki/Comparison_of_file_systems
“Maximum Volume Sizes” http://technet.microsoft.com/en-us/library/cc938432.aspx
“Windows Server 2012: Does ReFS replace NTFS? When should I use it?” http://blogs.technet.com/b/askpfeplat/archive/2013/01/02/windows-server-2012-does-refs-replace-ntfs-when-should-i-use-it.aspx
“File Allocation Table” http://en.wikipedia.org/wiki/File_Allocation_Table
“NTFS vs FAT vs exFAT” http://www.ntfs.com/ntfs_vs_fat.htm
“File System Functionality Comparison” http://msdn.microsoft.com/en-us/library/windows/desktop/ee681827(v=vs.85).aspx
This video will look at how multiple disks can be combined together to form the one volume and how basic and dynamic disks work. Dynamic disks are required in most cases when combining disks together.
Download the PDF handout
Basic and Dynamic Disks
Basic disks have been around since the MS-DOS days and thus have the most compatibility. Dynamic Disks were introduced in Windows 2000 and thus have good compatibility in Windows. Compatibility may become a problem if you use a dynamic disk in Linux. Dynamic disks are supported in Linux however support is dependent on the distribution. Additional packages may need to be installed and the distribution may not support booting from a dynamic disk. Dynamic disks also offer additional features over basic disks. For example, they add a signature to the disk so that if the disk is moved from one computer to another Windows will detect the disk has been moved.
Basic disks cannot be combined with other disks and thus features are limited to one disk. For example, an existing volume can be resized on a basic disk. In order to do this there must be contiguous space free at the end of the volume. If this is not there, the disk will need to be defragmented. Basic disks also support booting where dynamic disks only support booting when configured in certain ways. If you want to combine multiple disks together you require dynamic disks. Windows supports a one way conversion between basic and dynamic disks.
Dynamic Disk Support
Dynamic disks are supported in all Windows Server editions after Windows Server 2000. Client operating systems between XP and 7 require the high business level operating system but after that, all operating systems support dynamic disks.
Combining Multiple Disks
Dynamic disks allow multiple drives or the space from multiple drives to be combined together in order to create one volume. There are a number of different configurations supported.
Spanned volume: A spanned volume combines the space from multiple drives together to form the one volume. This allows you to combine unused space into one volume, however it does not offer redundancy if a drive fails nor does it offer speed improvements. It also does not support booting.
Mirrored Volume: This is when two disks have the same data on them. This allows the operating system to continue running if one disk fails. Unlike other dynamic disk configurations, mirrored volumes support booting of the operating system.
Striped Volume: This volume takes disks of the same size and combines them together to form the one disk. The data is evenly distributed across the disks. It offers excellent read and write performance however does not offer any redundancy. It also does not support booting.
RAID-5: This configuration requires 3 or more drives with the same amount of free space. The drives are combined into one volume. One drive is lost to parity, which is extra data, that means that volume will still keep running if one drive is lost. This volume gives fast read performance but slow write performance. It is available only in Server operating systems and does not support booting the operating system.
“Installing and Configuring Windows Server 2012 Exam Ref 70-410” pg 44
“Basic and Dynamic Disks” http://msdn.microsoft.com/en-us/library/windows/desktop/aa363785(v=vs.85).aspx
“Logical Disk Manager” http://en.wikipedia.org/wiki/Logical_Disk_Manager
“What are basic and dynamic disks?” http://windows.microsoft.com/en-au/windows-vista/what-are-basic-and-dynamic-disks
This video will look at a number of different storage solutions. These include software and hardware based systems. A storage solution is a system that allows more drives to be combined together for performance or redundancy reasons.
Download the PDF handout
What’s in this video
The following storage solutions will be looked at in this video.
Software vs Hardware: The advantages to using hardware over a software solution.
JBOD: Just a bunch of disks. Allows different sized drives to be combined together to form the one drive.
Spanning: Allows multiple drives to be combined of different sizes.
RAID: Redundant Array of Inexpensive Disks is a system that allows multiple drives to be combined to form the one drive.
Windows Storage Spaces: This is a new system implemented in Windows Server 2012 that allows multiple drives to be combined together.
Logical Volume Manager: Is an alternative storage system used by operating systems like Linux.
Software vs Hardware
Hardware based systems typically cost more than a software solution as software solutions usually come with the operating system free of charge. A lot of motherboards now come with free hardware based solutions. You will find that if you purchase a server this may come with some hardware based solutions. Some servers may require additional hardware in the server or a higher model may need to be purchased to gain access to some hardware based solutions. The biggest advantage of a hardware solutions is that the operating system sees the drive as a single physical drive. This means the operating system can be booted from this drive. Some software based solutions do not support booting of the operating system. Software solutions may also support some additional features not supported by hardware. For example a software based solutions may allow for multiple files containing the same data to use the same physical space on the drive. Enterprise hardware solutions will often offer additional features as well but do cost more. For example, enterprise hardware solutions will have a web interface allowing access to additional features.
Just a Bunch of Disks allows multiple drives to be combined together. This includes different sized drives and different types. For example you could combine solid state drives and mechanical drives together. JBOD does not offer any performance increase and if one of the drives was to fail you would lose all the data on all the drives.
Spanning is similar to JBOD however it combines free space on multiple drives together into the one drive. The advantage of spanning is that it allows space that may have otherwise been lost to be used. Spanning does not provide any speed advantages and also does not offer any redundancy. If a drive that is used in spanning was lost, then all the data in the spanned set would be lost.
Redundant Array of Inexpensive Disks is a system which allows multiple drives to be combined together to form the one drive. The drives need to be the same size in order to be used. If one drive is larger than the others, typically it still can be used, however the extra space will be left unused. Depending on which RAID solution is used will determine if there are any redundancy or speed advantages. The more expensive RAID solutions may allow drives to be added to the RAID, increasing the amount of space in the RAID. A lot of RAID solutions do not offer this feature and thus if you want to change the size of the RAID you need to destroy the RAID and recreate it.
RAID 0 (Striping)
RAID 0, otherwise known as striping, requires 2 or more disks in order to operate. Data is divided between the drives in blocks. This offers fast read and write performance but no redundancy. If you have a file that is smaller than a block it will be stored on the one drive. Files that are larger than a block will be divided up and stored on different drives. If one drive is lost in the set, all data in the set is lost. If you have difficultly remembering which RAID level is striping, remember that RAID 0 offers zero redundancy.
RAID 1 (Mirroring)
RAID 1 requires two drives. It works by storing two copies of the data. One copy on each drive. This means that if one drive was to fail no data would be lost. This gives redundancy; however, it does not offer any performance increases. It also effectively doubles the cost of storage. RAID 1 is often used for the operating system drive.
RAID 5 (Striped with Parity)
RAID 5 divides the data up onto multiple drives using one drive as parity. The parity information allows for a single drive in the system to fail while allowing the RAID to still operate. Let’s consider the following 1 + 2 = 3. If you did not know the value 2, that is 1 + x = 3, you could work out that the missing value is 2. This is the same principle of parity, adding additional information so that if a drive is lost, the additional information can be used to work out what this information was. RAID 5 requires 3 or more drives to operate. It offers excellent read performance as multiple drives can be read at once. Write performance is slow however when data is written to the drive as the parity information must also be updated. Depending on how much data is being written to the drive, updating the parity may involve a read before the write can be performed.
This is sometimes referred to as RAID 10. RAID levels like this combine two RAID levels together. In this case drives are stored using RAID 1. All the RAID 1 drives are then striped, thus RAID 1+0 is also known as a stripe of mirrors. This solutions offers excellent read and write performance as there is no parity drive that needs to be updated. The disadvantage is the cost of storage doubles. The advantage of RAID1+0 is that is can still operate after multiple failures.
RAID 1+0 uses multiple drives to store the same data on and thus can still operate after multiple failures. Since the system does use RAID 1 sets, it is possible for the system to fail after losing 2 drives. This will occur if two drives were to fail that were holding the same data. If multiple drives were to fail that are not holding the same data the RAID 1+0 will still work. Business solutions will often have a large number of drives in them so the chance of two drives holding the same data to fail is very low. Thus RAID 1+0 offers a performance increase over RAID 5 but does not have the disadvantages of RAID 5.
There is also RAID 0+1 which is essentially a mirror of a striped set. Like RAID 1+0 it offers excellent read and write performance and can still operate after multiple drive failures. However, if 2 drives were to fail with the same information on them then all data would be lost.
A storage system is a system that manages the physical drives in a system for the administrator. The administrator will generally create drives using the free space. For example they could create a mirrored drive or a drive with parity. The storage system is responsible for keeping the required level of redundancy. If a physical drive was to fail and later was replaced, the storage system would take care of making sure the redundancy level is still keep. For example, if a large drive was to fail and was replaced by two smaller drives, the storage system would reorganize the data so the required redundancy level is still kept. Software systems can be hardware or software based. A storage system in hardware is usually found in high level solutions like SAN solutions. Software based storage solutions are often included as part of the operating system.
Windows Storage Spaces
Windows Storage Spaces was added in Windows Server 2012 and Windows 8. It is designed as a replacement for the existing system, Logical Disk Manager, however it lacks some features of Logical Disk Manager, most noticeable is that the operating system cannot be booted from a Windows Storage Spaces drive. Windows Storage Spaces support the creation of multiple drives from a storage pool. The drives can have mirroring, parity or no redundancy. If you are using Windows Server 2012 R2, there is the option for 3 way mirror and extra redundancy. Windows Storage Spaces adds the option for thin provisioning. This allows a virtual drive to be created that only uses space as required. This is the same system that is used for virtual machines. Windows Storage Spaces was also designed to work with ReFS. ReFS is a new file system that was introduced in Windows Server 2012 and Windows 8.
Logical Volume Manager (LVM)
This is a system used in a lot of Linux based systems. It can create drives with or without redundancy and allows for features like the volume to be resized as required. The advantage of LVM is that the operating system can be booted from a LVM drive.
“Installing and Configuring Windows Server 2012 Exam Ref 70-410” pg 49-55
“Nested RAID levels” http://en.wikipedia.org/wiki/Nested_RAID_levels
“Step-by-step for Storage Spaces Tiering in Windows Server 2012 R2” http://blogs.technet.com/b/josebda/archive/2013/08/28/step-by-step-for-storage-spaces-tiering-in-windows-server-2012-r2.aspx
“Logical Volume Manager (Linux)” http://en.wikipedia.org/wiki/Logical_Volume_Manager_(Linux)
This video will look at how to use the logical disk management (LDM) utility included with Windows. This tool allows physical disks to be used as well as combining multiple physical disks together.
Download the PDF handout
1) To start disk management, run “diskmgmt.msc”. This can be done from the run box from the start menu. On Windows 8.1 this can be done by right clicking the start menu. This can also be accessed from computer management, however running it from computer management uses less screen real estate.
2) If you wish to make a volume smaller, right click the volume and select the option “Shrink Volume”. The volume will be able to shrink if there is contiguous free space at the end of the volume. If there is not, the drive will need to be defragged first. This can be done by opening the properties for the drive, selecting the tools tab and then press the button “Optimize”.
3) To create a new simple volume right click some free space and select the option “New Simple Volume” to start the wizard. In the wizard, you can select the size of the drive and a drive letter. If you do not want to assign a drive letter you can mount the volume to a folder or not assign anything and do this later. You can also decide how you want the drive formatted.
4) To extend a volume, right click the volume and select the option “Extend volume”. If this option is not available there is most likely no free space after the volume.
Demonstration Dynamic Disks
In order to access advanced features like combining multiple physical drives together, the drive must be converted to a dynamic disk. This can be done by right clicking the physical drive and selecting “Convert to dynamic disk” or this will be done automatically as required.
1) To mirror an existing drive, right click the drive and select the option “Add mirror”. It is just a matter of then selecting a physical drive with enough free space for the mirror. The data on the drive must be synced with the other physical drive which may take some time to complete.
2) If you wish to convert a physical drive to a basic disk, this will not be possible unless all the data for the physical drive is removed first.
3) To create a striped volume right click on some free space and select the option “New Striped Volume”. A striped volume spreads out the data on multiple physical drives eventually. This offers no redundancy; however, it gives better performance. This also means that if one of the physical drives was to fail, all data that was stored in the striped volume would be lost. Since data is spread eventually, if you add multiple physical drives together, the total space will be a multiple of the drive with the least amount of free space. Once the striped volume is created, you will not be able to add additional drives to the striped volume or change the size of the striped volume.
4) To create a new RAID-5 volume, right click on some free space and select the option “New RAID-5 Volume”. As RAID-5 volumes use one physical drive to store parity, they require 3 or more physical drives in order to operate. This effectively means that 1/3 of the total drive space is lost to parity, but it also means the volume will still work after the loss of one of the physical drives. Once the RAID-5 volume has been created, the data between the physical drives will not to be re-synced. This does take a while to complete. RAID-5 is only available in Windows Server.
5) If you have multiple physical drives with free space at the end, the space from these physical drives can be combined together to form the one volume. The space on each drive does not have to be the same amount. A spanned volume offers no redundancy or speed increases and if one physical drive is lost in the volume, all data on the volume is lost. Spanned volumes however can be shrunk or extended, unlike striped or RAID-5.
6) If you no longer require a mirrored drive, you can right click the volume and select the option “Remove mirror”. This will create two physical drives with the same data on it. If the volume is still responsive, this is the clean way to remove a mirror.
7) If a physical drive in a mirror volume was to fail, you can use the option “Break Mirrored Volume”. This will remove the mirror, however if both physical drives are still accessible, now or later, they may not have the same data on them so this option is generally only used when a physical drive is no longer responsive and is going to be removed.
8) If you need to replace a failed drive in a mirror, use the “Break Mirrored Volume” and then create a new mirror using the new physical drive.
9) If you place a physical drive in the system that has been configured as a dynamic disk in another system this physical drive will appear in disk management as a foreign disk. In order to use it, you need to right click the disk and select the option “Import Foreign Disks”. If it is part of a set of volumes, for example a striped volume, Windows will attempt to import the whole volume, however the volume will not be usable if physical drives are missing. To use the physical disk, the volume must first be imported with the other physical disks missing and then the volume needs to be deleted.
10) If you have a physical disk in a RAID-5 volume that stopped working and is now working, you can refresh the volume by selecting the option “Reactivate Volume”. This will perform a check of the volume and make corrections where needed.
11) If you have a physical disk that is a RAID-5 volume that has failed, you can replace the physical disk by right clicking the volume and selecting the option “Repair Volume”.
This video will look at the new features added to Windows Server 2012 R2 with Windows Storage Spaces. Window storage Spaces is a system introduced in Windows 8 and Windows Server 2012 that allows multiple physical drives to be combined together to form the one drive.
Download the PDF handout
Storage Spaces New Features
This video will look at the 5 new features included in Windows Storage Spaces in Windows Server 2012 R2. Three of the features are transparent and are available automatically and do not need to be configured.
Solid state drives differ from Traditional Hard disks in that the access time is quite fast. Storage Tiers allow the administrator to divide storage into a solid tier and a physical hard disk tier. When an administrator creates a simple or mirror drive using Windows Storage Spaces, the administrator can choose a percentage of how much space will be used on each tier. Windows will automatically place files that are used more often on the solid state tier and less used files on the other tier. At this stage, Windows Storage Spaces does not support parity for storage tier. Parity allows data to be divided up between drives for performance and redundancy.
Traditional on a RAID system, the overall system will be slowed down when a lot of small random writes are performed. If you have solid state drives in the Windows Storage pool, Windows Server 2012 R2 will automatically reserve part of the solid state drive to use as a write-back cache. When a lot of small random writes are received, these are written to the solid state cache. Since solid states have fast access time, this is very quick. Later on, the data in the cache is transferred to the other physical drive in the system as required. This feature is enabled automatically, however in order for it to work it requires a certain number of solid state drives depending on the type of redundancy you are using in the system. For example, a 2-way mirror requires two copies of the data to be available at all times in order for it to be redundant. For this reason, the write back cache needs to hold 2 copies of the data in case a solid state drive failed before the data could be transferred. If you do not have the required number of solid state drives, the feature will not be available.
Parity space support for failover clusters
Support has been added for Windows clusters to use drives created in Windows Storage Spaces that use parity. Previously in Windows Server 2012 parity support was not available.
Windows Server 2012 R2 Storage Spaces adds the option for dual parity. Dual parity means that the virtual drive can still keep working after 2 disk failures. In order to use this feature, you need to have 7 drives in the storage pool.
Auto rebuild storage spaces from free space
Let’s consider a standard RAID of 5 drives with one being used as parity. If a drive was to be replaced, the 4 other drives would need to read to perform a read and then the data written on the new drive. This is a slow process as writes are only being performed on the one drive. With Windows Storage Spaces for Windows Server 2012 R2, rather than writing all the data to a single drive, any free space on the other drives in the Windows Storage Space can be used to write to. Assuming redundancy requirements are still met, this means multiple drives are being used rather than just the one.
“What's New in Storage Spaces in Windows Server 2012 R2” http://technet.microsoft.com/en-us/library/dn387076.aspx
“Step-by-step for Storage Spaces Tiering in Windows Server 2012 R2” http://blogs.technet.com/b/josebda/archive/2013/08/28/step-by-step-for-storage-spaces-tiering-in-windows-server-2012-r2.aspx
This video will look at using Windows Storage Spaces in Windows Server 2012 R2. Windows Storage Spaces is a new system introduced in Windows Server 2012 that allows multiple physical drives to be combined together to form the one virtual drive. This can be done for redundancy or capacity reasons.
Download the PDF handout
What is looked at in this video
This video will look at seven different topic areas in storage pools and how to configure them in Windows Server 2012 R2. These seven areas are
1) Creating a storage pool
2) Automatic vs manual allocation
3) Creating simple, mirror and parity virtual disks
4) Removing a disk from the storage pool
5) Extending a virtual disk
6) Replacing a failed drive
7) Storage Tiering
Demonstration creating a storage pool
1) If you right click on the start icon this will display a menu. Select the run option and then run “Diskmgmt.msc”. This will run disk management in full screen mode. Windows Storage Spaces is not managed through Disk Management; however Disk Management will tell you what physical disks are present in the system.
2) To configure Storage Spaces, run Server Manager and then select the option on the left hand side “File and Storage Services”.
3) In the File and Storage Services admin tool, if you select Disks this will show all the physical disks installed in the system. This will show all the physical disks in the system including those that are not useable by Windows Storage Spaces. For example, the physical drive that contains the operating system will not be able to be used with Windows Storage Spaces.
4) If you select the option “Storage Pools” this will show any storage pools that have been created. If a physical drive is available but has not been put into a storage pool it will be listed under Primordial.
5) If you right click on Primordial this will give you the option for “New Storage Pool” which will start the new storage pool wizard.
6) The wizard will allow you to enter in a name and description and also select which physical drives you want to use in the storage pool.
7) When selecting the type, it is best to choose all physical drives as automatic or manual but not a mixture of both as this will limit what kinds of virtual disks you can create if the storage pool does not have enough physical drives in it. If you select the option for hotspare, this physical drive will automatic replace another drive in the system if it fails.
8) Once you have selected all the required options, at the end of the wizard you have the create button which will create a storage pool. Once the storage pool is created, virtual disks can be created in the storage pool.
Automatic vs manual allocation
When creating a Storage Pool, Windows will ask if the physical drives used in the storage pool should be automatic or manual. It is recommended that one or the other is used. In this example 7 physical drives exist. Four have been put in an automatic allocation and 3 in a manual allocation. When this occurs, when creating a new virtual drive, Windows will ask which allocation you want to use. In this case, options like 3 way mirror and dual parity are not available as both pools do not have the minimum of 5 physical disks. If all 7 physical drives were put in the same allocation then all options for virtual drives would be available.
Demonstration Creating simple, mirror and parity virtual disks
1) Once a storage pool has been created it will appear in Storage Pools in File Storage Services. If there is no storage pool available you will need to create one.
2) To create a virtual disk inside a storage pool, right click on the storage pool and then select the option “New Virtual Disk” to start the new virtual disk wizard. Windows can support 160 virtual disks in the one storage pool.
3) The wizard will ask which storage pool that you want to use, a name for the virtual disk and a description for the virtual disk.
4) For the layout of the virtual disk you can choose simple, mirror and parity. The option will change depending on which operating system you are using and how many physical drives are in the storage pool. It important to remember that if you choose simple there is no redundancy. If any physical drive in the storage pool was to fail that was being used with that simple disk, all the data from the simple disk is lost.
5) The provisioning option will determine if the virtual disk will allocate all the space when the virtual disk is created or as needed. If you want the virtual disks to expand in size as required use the thin option. If you want to allocate all the space at once use the fixed option. The fixed option gives better performance and offers some additional options. For example, Dual Parity requires the fixed option.
6) The wizard requires the size of the virtual disk to be entered. Windows will automatically check to ensure that there is enough space in the Storage Pool to create a virtual disk of that size.
7) When you reach the last screen of the wizard, once you press create the virtual disk will be created and available in the operating system as an empty unused drive. In other words, it will still need to be partitioned and formatted like any physical drive would need to be. After the wizard has completed, the new volume wizard can be run which will allow the virtual disk to be formatted. If you do not do this now, you can always perform this step using disk management later on.
8) The “New Volume Wizard” is a simple wizard which allows you to prepare the virtual disk for use in the server. It asks a number of simple questions like the size of the drive you want to create, the drive letter you want to use and also what sort of file system that you want to use.
9) Now running the virtual disk wizard, options will be available based on the operating system used, the number of physical disks in the pool and which type of provisioning that you have selected. For a list of what is available refer to the PDF.
10) If you want to get an indication of how much space is left in the storage pool, you can right click the storage pool and select the option properties. Also there are options for health and additional details about the storage pool.
Demonstration removing a disk from the storage pool
1) On the right hand side of File and Storage Services is listed the physical drives.
2) To remove a physical drive from the system, right click the physical drive and select the option “Remove disk”.
3) When this is attempted, Windows will attempt to move all the data off the physical disk to other disks. If a simple disk is on that physical drive, the simple virtual disk will need to be removed first before you can remove the physical disk.
4) To remove a simple disk from the storage pool, right click in the disk section and select the option “Delete Virtual Disks”.
5) If you attempt to delete a virtual disk that has a volume on it, the operation will fail. In order to delete the virtual disk you need to first go into Disk Management and delete any volumes of the virtual disk first.
6) If you attempt to remove a physical disk from the system, Windows will attempt to move any data from that physical disk to other physical disks. In this example there is a dual parity disk which requires 7 physical disks and there are only 7 physical disks in the system. Before the disk can be removed, another physical disk needs to be added to the system and adding it to the storage pool.
Extending a virtual disk
1) To make a virtual disk larger, right click the virtual disk and select the option “Extend Virtual Disk”.
2) Extending a virtual disk will only make the virtual disk larger. If you have any volumes on the virtual disk you will still need to go into Disk Management and make the volume larger to use the extra space added to the virtual disk.
Replacing a failed drive
When a physical drive fails in a Storage Pool, an exclamation mark will appear to indicate that there is a problem. This will appear in the storage pool, the physical disk, and the virtual disks that are effected.
1) To repair a virtual disk, right click the virtual disk and select the option “Repair Virtual Disk”.
2) If there are not enough available physical disks in the system to repair the virtual disk the operation will fail.
3) The simplest way to replace a physical disk that has failed is to put in an additional physical disk and set the type to hot spare. When you select the option for “Repair Virtual disk”, Windows will copy any data that was on the failed physical disk to the hot spare. Once the data has been copied to the new physical disk, the old physical disk should be reported as retired and can be removed.
This is a feature that is available in Windows Server 2012 R2. This allows data that is used more often to be placed on Solid State drives and data that is used less often to be placed in physical hard disks.
1) Under physical disks there is a media type that should show the type of hard disk as being mechanical hard disks or solid state hard disks. If this is not the case, open PowerShell and run the following.
2) “Get-PhysicalDisk” This command will show all the current physical disks in the system.
3) “Get-PhysicalDisk | where FriendlyName –Like PhyiscalDisk[2-4]” This will display physical disks 1 through to 4.
4) “Get-PhysicalDisk | where FriendlyName –Like PhyiscalDisk[2-4] | Set-PhysicalDisk –MediaType HHD” This command will configure disks 2 through 4 as mechanical hard disks.
5) “Get-PhysicalDisk | where FriendlyName –Like PhyiscalDisk[5-8] | Set-PhysicalDisk –MediaType SSD” This command will change physical disks 5 through 8 to solid state drives.
6) To create a new virtual disk that uses storage tiers, right click the storage pool and select the option “New Virtual Disk”.
7) Creating a virtual disk that uses Tiers is much the same as creating a standard virtual disk. When you get the option for virtual disk name, there is a tick box at the bottom of the screen called “Create storage tiers on this virtual disk”. This option needs to be ticked in order for the virtual machine to use tiers.
8) Storage Tiers require fixed provisioning so when you get to that part of the wizard you will not be able to select the thin option.
9) On the size screen you have to enter in the size of the virtual disk. The big difference here over non tier storage is that you need to enter in how much space that you want to use on the SSD space and how much you want to use on HHD space. Adding both together will give the total amount of space available for the virtual disk.
“Resilient file system” http://msdn.microsoft.com/en-us/library/windows/desktop/hh848060(v=vs.85).aspx
“Storage Spaces: How to configure Storage Tiers with Windows Server 2012 R2” http://blogs.technet.com/b/askpfeplat/archive/2013/10/21/storage-spaces-how-to-configure-storage-tiers-with-windows-server-2012-r2.aspx
“Storage Spaces Frequently Asked Questions (FAQ)” http://social.technet.microsoft.com/wiki/contents/articles/11382.storage-spaces-frequently-asked-questions-faq.aspx