How-To: Rapid Deploy VM Using Powershell

I have created a Powershell script to rapidly deploy VMs on Microsoft Hyper-V host and this was demo during the Hyper-V Workshop.


The demo Clustered Hyper-V infrastructure are as below diagram:

Clustered Hyper-V infrastructure with MS Hyper-V 2008 R2 SP1

On of my interesting demo is to rapid deploy VM on Microsoft Hyper-V host without using SCVMM. In achieving this I have created a Powershell script by referring to existing script found on http://social.technet.microsoft.com/wiki/contents/articles/176.aspx.

The script which I created is not perfect to suit everyone yet, but its enough for most of us out there who in need to create n numbers of VM given a short period. During this event, I have conducted a demo in creating 6 VMs with just a hit on 'enter' and all 6 VMs booted. And right now, I will share with you in details how to achieve this.

What the Powershell script does (In general speaking)

1. Read VMs configuration from CSV line-by-line

2. Create VM into 'C:\clusterstorage\volume1\vname'

3. Set the processor count of each VMs

4. Set the memory for each VMs

3. Create and attach a 'Differencing Disk' to each VMs

Future development of this script

1. To be able to set 'Dynamic Memory' for each VMs

2. To be able to set 'Fixed/Dynamic/Differencing Disk' for each VMs

3. To be able to set desire sysprep/preconfigured or mix OS platform base VMs

4. To be able to add each VMs to CSV cluster if the Hyper-V host is clustered
5. To add and improve comments into the script

CREATEVM.ps1

Import-Module "C:\Program Files\modules\HyperV"
$vmdefaultpath = "C:\ClusterStorage\volume1"
$ParentVHD = "win2k8r2sp1"
$path = "c:\createvm\VM.csv"
 import-csv -path $path|ForEach-Object {
$erroractionpreference = 0    
$vmName = $_.Name
$vmmemory = $_.Memory
$vmcpucount = $_.Cpucount
$vmSwitch = $_.Network
$vmpath = $vmdefaultpath
New-VM -Name $vmname
Set-VMMemory -VM $vmName -Memory $vmmemory
Set-VMCPUCount -VM $vmname -CPUCount $vmcpucount
Add-VMNIC -VM $vmName -VirtualSwitch $vmSwitch
New-VHD -VHDPaths $vmpath\$vmname\$vmname.vhd -ParentVHDPath $vmpath\vmbase\$ParentVHD.vhd
Add-VMDisk -VM $vmname -ControllerID 0 -Path $vmpath\$vmname\$vmname.vhd
Write-Host -BackgroundColor Green -ForegroundColor Black "Virtual Machine $vmname has been successfully created"     
}

VM.csv

CSV file which predefined VMs settings

How to use this script.

1. Create a CSV and list all necessary configuration of each VM as shown sample above.

2. Save the CSV as 'vm.csv'.

3. Copy 'vm.csv' to the Hyper-V host in 'c:\createvm\'

4. Copy the script above and save it as 'createvm.ps1' in 'c:\createvm\' (you may have to edit the variable to suit your requirement and configuration of your Hyper-V host.

5. Execute the PowerShell management Library for Hyper-V by codeplex.

6. Type 'powershell -file createvm.ps1'.

Requirement and caution
1. A syspreped OS image is required
2. Windows Server 2008 R2 and Windows 2008 R2 SP1 with Hyper-V role installed
3. Basic knowledge of Powershell script
4. Changing of variables in the script to suit your Hyper-V environment
5. Use at you own risk as you have been warned this may affects you Hyper-V server. KNOWING WHAT YOU ARE ABOUT TO EXECUTE IS IMPORTANT. 


A video worth thousand words

How-To: Compact Hyper-V Dynamic Expanding VHD

When you do housekeeping on your target VM which using 'Dynamic Disk', the free space that you free up in your VM will not added back to your Hyper-V host. If space is the concern for you Hyper-V host, then you may have to manually compact the VM's VHD.

To compact a Dynamic Expanding VHD:

Step 1: Right-click the target VM. Select 'Settings...'. Highlight the target Hadr Drive. Click 'Edit'

 Step 2: Select 'Compact'.

Compact - Applies to dynamically expanding virtual hard disks and differencing virtual hard disks. Reduces the size of the .vhd file by removing blank space that is left behind when data is deleted from the virtual hard disk. If the virtual hard disk is not NTFS formatted, the blank space must be overwritten with zeroes so that the compact action can reduce the file size by removing sectors that contain only zeroes.

If the virtual hard disk is not NTFS formatted, you must prepare the virtual hard disk for compacting by using a non-Microsoft disk utility program to replace the blank space with zeroes.

Step 3: Click 'Finish'

Wait for the compacting process to finish.

Below show the size of before and after VHD compacting process.

How-To: Enabling Data Co-Location on DPM

Microsoft's Data Protection Manager (DPM) 2010 allows backup administrator to co-locate protection groups on tape. In previous version of DPM, tape co-location is not available. This means to say that if you have a protection group with backup size maybe 2Gb a day and 5 LTO4 tapes and the backup strategy is to backup 5 days a week. You will end up with DPM 2010 uses each LTO4 1.6GB (compress) to backup your 2GB everyday. This only utilize 0.1% of your tape capacity, and the worst is that other protection group will not be able to write on the same tape.

With tape co-location available now you can optimizes the tape usage in case you have many small protection groups. The backup to tape will backup multiple protection group on the same tape depending on the backup strategy you configured and consideration on "TapeWritePeriodRatio" and "ExpiryToleranceRange".

Enabling Data Co-Location on DPM

1. Open DPM Management Shell.

Set-OptimizeTapeUsage to True using the Set-DPMGlobalProperty cmdlet.

Set-DPMGlobalProperty -DPMServerName <name of DPM server> -OptimizeTapeUsage $True

Take note that a dataset will be collocated only if both the below conditions are true.

The expiry date of the current dataset should fall in between the following dates: 

Upper bound: furthest expiry date among all the datasets on the tape - (furthest expiry date among all the datasets on the tape - current date) * ExpiryToleranceRange

Lower Bound: furthest expiry date among all the datasets on the tape + (furthest expiry date among all the datasets on the tape - current date) * ExpiryToleranceRange.

Current time should be less than first backup time of the dataset on the media + TapeWritePeriodRatio * RetentionRangeOfFirstDataset.

TapeWritePeriodRatio

It is a global property for the DPM which needs to be set using DPM CLI command. Here is the command to set it

Set-DPMGlobalProperty –DPMServerName <dpm server name> -TapeWritePeriodRatio <fraction>

WritePeriodRatio indicates the number of days for which data can be written on to a tape as a ratio of the retention period of the first data set written to the tape.

WritePeriodRatio value can be between 0.0 to 1.0

Default value is 0.15 (i.e. 15%)