Configuring and Managing RAID 5 on Windows Server 2012

An Overview of RAID 5

RAID 5 consists of three or more volumes each located on a separate physical disk. As with RAID 0, RAID 5 also uses disk striping, whereby blocks of data are divided up into stripes with each stripe written to a different disk. RAID 5, however, differs quite considerably from RAID 0. Under RAID 5, not only are the data stripes written, but also parity information relating to the data. The key to RAID 5 fault tolerance is the fact that the parity information for a particular data stripe is always written to a different drive from the drive containing the corresponding data stripe. This means that if a disk fails, the corresponding parity information stored on another disk can be used for error detection and data correction (also referred to as regeneration).
Whilst RAID 5 has considerable advantages over RAID 0 there are one or two drawbacks that should be taken into consideration when considering this storage option. Firstly, there is the inevitable performance overhead inherent in calculating and storing parity information for each data stripe written to disk. Secondly, the loss of more than one disk in a RAID 5 array will leave insufficient parity data on the remaining disks to regenerate the original data. That said, RAID 5 does provide considerable advantages that generally outweigh the disadvantages.

Configuring RAID 5 Using Windows Server 2008 Disk Management

A Windows Server 2008 RAID 5 configuration may be set up using the Disk Management snap-in. This is accessed either from the Server Manager or Computer Management tools. To launch the Server Manager, open the Start menu and click on the Server Manager option, or click on the Server Manager icon in the task bar. Alternatively launch Computer Management from Start -> All Programs -> Administration Tools -> Computer Management or run compmgmt.csc at the command prompt or in a Run dialog. In all cases the Disk Management tool can be found under the Storage category.
As previously noted, RAID 5 implementation requires a minimum of 3 disk drives. For the purposes of this tutorial a system containing four disk drives is assumed. In this scenario, disk 0 is the system disk and disks 1 through 3 are available for use in the RAID 5 configuration. Before proceeding the disks will need to be initialized using either the MBR or GPT partition style. Assuming these prerequisites are met the first step is to right click on one of the 3 disks in the Disk Management graphical view. In the resulting popup menu select the New RAID-5 Volume... option to invoke the New RAID-5 Volume wizard. On the wizard's welcome page click on the Next button to proceed to the Disk Selection screen. This screen contains a list of disk drives available for inclusion in the disk array together with a list of selected disks. Currently only the current disk is included in the Selected list. Two more disks must be added to the selected disks before the RAID 5 array can be built. Select disks from the Available list and click on the Add> button to add the disk to the selected list. Once sufficient disks (in this case disks 1, 2 and 3) are selected the Next button will activate to allow the remainder of the configuration to be completed. Note that disk 0 (the system disk) is not included in the RAID 5 array:

With the disk selections completed, the Next button proceeds to the drive letter and mount point assignment screen. Once these settings are configured click Next to proceed to the Format Volume screen. Select the appropriate file system and compression options and click on Next to proceed to the Summary screen. Review the information displayed and click on Finish to initiate the RAID 5 creation process. During this process the Disk Management graphical view will list the disks as Formatting and then Resynching. The amount of time these phases will take depends on the size of the volumes in question. Once the process is complete the status will change to Healthy and the RAID 5 volume is ready for use.

Configuring RAID 5 from the Command Prompt using DiskPart

In addition to configuring RAID 5 from within Disk Management, the configuration may also be implemented from the command prompt using DiskPart. DiskPart may be launched either from a command prompt or a Run dialog simply by typing diskpart. Once invoked, DiskPart will display the DiskPart> command prompt and is ready to receive commands.
The first step in the configuration process is to identify the disks attached to the system using the list disk command:

DISKPART> list disk

  Disk ###  Status      Size     Free     Dyn  Gpt
  --------  ----------  -------  -------  ---  ---
  Disk 0    Online        30 GB    15 GB   *
  Disk 1    Online         8 GB  8189 MB
  Disk 2    Online         8 GB  8189 MB
  Disk 3    Online         8 GB  8189 MB

For the purposes of this chapter disks 1, 2 and 3 will be used to create a RAID 5 configuration. Each of these disks needs to be converted to dynamic disks before the configuration can proceed. This is achieved by selecting each disk in turn and executing the convert dynamic command:

DISKPART> select disk 1

Disk 1 is now the selected disk.

DISKPART> convert dynamic

DiskPart successfully converted the selected disk to dynamic format.

DISKPART> select disk 2

Disk 2 is now the selected disk.

DISKPART> convert dynamic

DiskPart successfully converted the selected disk to dynamic format.

DISKPART> select disk 3

Disk 3 is now the selected disk.

DISKPART> convert dynamic

DiskPart successfully converted the selected disk to dynamic format.

Once the disks have been converted to dynamic disks the next step is to create the RAID 5 volume using the create volume raid command. This command also takes as a parameter the disk= directive followed by a list of disks to be used in the array. The size= directive may also be specified to declare the size of the volume. If this value is omitted the volume will be sized to match the smallest contiguous block of unallocated space on the designated disk drives:

DISKPART> create volume raid disk=1,2,3

DiskPart successfully created the volume.

Once the command has completed and displayed the DISKPART> prompt the system will have begun the resynching process. This can take a considerable amount of time depending on the size of the volume. During this process, the status of the volume will be listed as Rebuild when the list volume command is executed:

DISKPART> list volume

  Volume ###  Ltr  Label        Fs     Type        Size     Status     Info
  ----------  ---  -----------  -----  ----------  -------  ---------  --------
* Volume 0                      RAW    RAID-5        16 GB  Rebuild
  Volume 1     C                NTFS   Simple        15 GB  Healthy    System
  Volume 2     D                       DVD-ROM         0 B  No Media

Once the resynch is complete the volume status will change to Healthy at which point the volume may be formatted (the RAW type listed above indicates the volume has yet to be formatted). The volume may be formatted using the format command as follows:

DISKPART> format fs=ntfs label="RAID 5 Vol"

  100 percent completed

DiskPart successfully formatted the volume.

Once the volume has been formatted the RAID 5 configuration is ready for use.

Fixing RAID 5 Problems

In order to function, all the disks in a RAID 5 configuration must be online and healthy. If the RAID 5 set displays Failed Redundancy and the volume is listed as Offline, Missing or Online (Errors) there is a problem which needs to be resolved.
If the status is Offline or Missing, check that the disk is connected and powered up. If the problem is resolved select Rescan Disks from the Actions menu. Once the rescan is complete right click on the problematic drive and select Reactivate. All being well the drive status will change to Regenerating as the data is rebuilt using the parity information on the other drives in the array. If, once regeneration is complete, the status does not return to Healthy, right click once again and select Regenerate Parity.
If a disk in a RAID 5 configuration is listed as Failed or Online (Errors) repeat the above steps. If this fails the disk may be unrecoverable and will need to be replaced. To achieve this, right click on the volume on the failed disk and select Remove Volume from the popup menu. Once the volume has been removed, right click on a suitable block of unallocated space on another dynamic disk which is not already part of the RAID 5 array, has sufficient space and has a matching partition style and select Repair Volume. This will rebuild the RAID 5 configuration using the space on new disk drive together with the remaining healthy disks from the original RAID 5 configuration.

Readmore...

How to Team Your Network Cards in Windows Server 2012

Windows Server 2012

The great thing about NIC teaming is that it is included in the box, free, and will work on any NIC that works with Windows Server 2012. It is remotely configurable and remains the same process no matter what card you use. Obviously you need a server and at least two NICs.

A team in Server 2012 can comprise two or more NICs from one or more vendors. The team can be added to after creation, and this is achievable from a Windows 8 client remotely on a Server 2012 machine - with or without a GUI on the server. The whole process is underpinned by PowerShell 3.0.
For the purpose of this guide, I will have three servers:
1. DC and DHCP server (single NIC)
2. Hydrangea (to be teamed, member of the domain)
3. Dahlia (to be teamed, member of the domain)
Of course all of this can be achieved using Virtual Servers. My setup was carried out on a Dell Laptop using Windows 8 RTM with Hyper-V.

Where to begin

It would be normal to use static IP addresses for all servers (or most of them), and certainly on a team of NICs. It is just as easy to use DHCP too, Server 2012 will allow that and I will use DHCP to ease the configuration.
So make sure you have a functional Server 2012 with a DHCP scope available to it (can be installed as a role on the server, or be available somewhere else on your network). If you have no DHCP, then you will need to manually assign IP addresses to your teams when you create them.
As a quick test of your current network configuration, run PowerShell 3.0 and enter the following commands:
RESOLVE-DNSNAME Hydrangea
RESOLVE-DNSNAME Dahlia
(The results of these Cmdlets show that each server has two NICs and two IP addresses - showing that the NICs are NOT teamed).

Now team the NICs on each server

This can be performed remotely from the DC. Run Server Manager and select All Servers, right-click Hydrangea and select Configure NIC Teaming, the dialog box that is presented allows all teaming activities.
Ensure the server is selected under Servers, and under Teams select Tasks and click on New Team.
Enter the Team Name - I used the creative name of "Team1" - then simply select the NICs you want to include in your team (minimum of two). Then select Additional Properties to choose how you want to set-up your team.
Now here are the really powerful settings. You can choose to be Switch dependent or independent, if redundant switches are important to you. You can choose if you want a standby adaptor or if you want all adaptors included and active. Finally, you can choose the VLAN you want this Team to operate on.
Select your options (Switch independent, None [for standby] and Default VLAN), then click OK. This will take a couple of minutes and will cause a loss of connectivity to the server you are managing (you are changing its IP address). To reconnect, you need to delete your local DNS cache and refresh server manager and all will be well (PowerShell – enter Clear-DNSClientCache and Resolve-DNSName ). This will show that you now have only one IP address and a single interface.
Repeat this set of steps on your second server. You now have a pair of teamed servers.

The way to test the teaming is quite simple, but does involve more PowerShell 3.0

First, on your equivalent of my Hydrangea server, run a PowerShell window and type Ping -t DC (where DC is your DC server name) this will keep pinging the DC until you tell it to stop, leave the window open.
Open another PowerShell window and set it side-by-side with the ping window. My NICs were called LAN and LAN2, you will need to use your correct naming in the below commands:
Get-netadapter
Disable-NetAdapter –Name “LAN” –Confirm:$False
After the first disable command – notice that the ping is still working fine
Disable-NetAdapter –Name “LAN2” –Confirm:$False
and only fails after the second disable command.
Enable-NetAdapter –Name “LAN” –Confirm:$False
Enable-NetAdapter –Name “LAN2” –Confirm:$False
Once NICs are re-enabled, all goes back to normal (in-fact comms are restored when the first NIC is re-enabled).

Summary

Prior to the release of Windows Server 2012, to be able to implement NIC teaming required several prerequisites:
1. NICs of the same make and model, and often of the same firmware revision
2. Third-party drivers built specifically for teaming
3. Skills in configuring the specific NICs and drivers.
With this groundbreaking release, Microsoft has provided all the functionality in the box for free – and removed all of the above.
Your NICs can be any make, model or revision; provided it works with Server 2012, the drivers all have the functionality built in.
The real USP is the fact that this can be done remotely on any number of servers through Server Manager and of course through

Readmore...

NIC Teaming in Windows Server 2012

NIC Teaming in Windows Server 2012

The great thing about NIC teaming is that it is included in the box, free, and will work on any NIC that works with Windows Server 2012. It is remotely configurable and remains the same process no matter what card you use. Obviously you need a server and at least two NICs.

The Logicalis Enterprise Cloud

Download Now

A team in Server 2012 can comprise two or more NICs from one or more vendors. The team can be added to after creation, and this is achievable from a Windows 8 client remotely on a Server 2012 machine - with or without a GUI on the server. The whole process is underpinned by PowerShell 3.0.
For the purpose of this guide, I will have three servers:
1. DC and DHCP server (single NIC)
2. Hydrangea (to be teamed, member of the domain)
3. Dahlia (to be teamed, member of the domain)
Of course all of this can be achieved using Virtual Servers. My setup was carried out on a Dell Laptop using Windows 8 RTM with Hyper-V.

Where to begin

It would be normal to use static IP addresses for all servers (or most of them), and certainly on a team of NICs. It is just as easy to use DHCP too, Server 2012 will allow that and I will use DHCP to ease the configuration.
So make sure you have a functional Server 2012 with a DHCP scope available to it (can be installed as a role on the server, or be available somewhere else on your network). If you have no DHCP, then you will need to manually assign IP addresses to your teams when you create them.
As a quick test of your current network configuration, run PowerShell 3.0 and enter the following commands:
RESOLVE-DNSNAME Hydrangea
RESOLVE-DNSNAME Dahlia
(The results of these Cmdlets show that each server has two NICs and two IP addresses - showing that the NICs are NOT teamed).

Now team the NICs on each server

This can be performed remotely from the DC. Run Server Manager and select All Servers, right-click Hydrangea and select Configure NIC Teaming, the dialog box that is presented allows all teaming activities.
Ensure the server is selected under Servers, and under Teams select Tasks and click on New Team.
Enter the Team Name - I used the creative name of "Team1" - then simply select the NICs you want to include in your team (minimum of two). Then select Additional Properties to choose how you want to set-up your team.
Now here are the really powerful settings. You can choose to be Switch dependent or independent, if redundant switches are important to you. You can choose if you want a standby adaptor or if you want all adaptors included and active. Finally, you can choose the VLAN you want this Team to operate on.
Select your options (Switch independent, None [for standby] and Default VLAN), then click OK. This will take a couple of minutes and will cause a loss of connectivity to the server you are managing (you are changing its IP address). To reconnect, you need to delete your local DNS cache and refresh server manager and all will be well (PowerShell – enter Clear-DNSClientCache and Resolve-DNSName ). This will show that you now have only one IP address and a single interface.
Repeat this set of steps on your second server. You now have a pair of teamed servers.

The way to test the teaming is quite simple, but does involve more PowerShell 3.0

First, on your equivalent of my Hydrangea server, run a PowerShell window and type Ping -t DC (where DC is your DC server name) this will keep pinging the DC until you tell it to stop, leave the window open.
Open another PowerShell window and set it side-by-side with the ping window. My NICs were called LAN and LAN2, you will need to use your correct naming in the below commands:
Get-netadapter
Disable-NetAdapter –Name “LAN” –Confirm:$False
After the first disable command – notice that the ping is still working fine
Disable-NetAdapter –Name “LAN2” –Confirm:$False
and only fails after the second disable command.
Enable-NetAdapter –Name “LAN” –Confirm:$False
Enable-NetAdapter –Name “LAN2” –Confirm:$False
Once NICs are re-enabled, all goes back to normal (in-fact comms are restored when the first NIC is re-enabled).

Summary

Prior to the release of Windows Server 2012, to be able to implement NIC teaming required several prerequisites:
1. NICs of the same make and model, and often of the same firmware revision
2. Third-party drivers built specifically for teaming
3. Skills in configuring the specific NICs and drivers.
With this groundbreaking release, Microsoft has provided all the functionality in the box for free – and removed all of the above.
Your NICs can be any make, model or revision; provided it works with Server 2012, the drivers all have the functionality built in.
The real USP is the fact that this can be done remotely on any number of servers through Server Manager and of course through PowerShell 3.0.

Readmore...