Monday, March 17, 2014

Add a virtual switch to Hyper-V with Desired State Configuration

In my last post I was really simple, I added the Hyper-V Role and the Hyper-V PowerShell module to a Server 2012 R2 machine using the Server Manager Desired State Configuration provider (this is a capability that is provided in-box).

Lets step it up a bit.

In the community resources MSFT has released the xHyper-V Module as a component of the Desired State Configuration Resource Kit.

If you follow the instructions you will see that adding this to your Windows Server is as easy as:

  • Download the xHyper-V module to the server that you are testing with
  • unzip the package to C:\Program Files\WindowsPowerShell\Modules
  • It should create a folder ‘xHyper-V’ under which you will find the necessary parts
    • “ C:\Program Files\WindowsPowerShell\Modules\xHyper-V”
    • including a DSCResources folder containing the classes of this module
    • and the xHyper-V.psd1 file.

But that is all you need to do to add this, nothing more.  You are done, simply begin using it.  If you want to see that your server has ‘found’ the module; open your PowerShell command window and type Get-DscResource – you will see a list of all of the providers and modules.

Back to the xHyper-V module..  It includes resources for handling VMSwitches, VHDs, VMs, file directories, and VHD files. 

If you think the module grouping is strange for the features, it just might be; a Resource Provider is designed to act upon a particular object beginning with ‘Present” and “Absent” and then most allow applying other properties to that object to extend it.  And in a configuration multiple modules can be called to mix and match and produce a complete configuration.

In the previous article I only used one module – Server Manager.  In this example I will use two.  And in the next article I will use three and add more dependencies.

So, how do you use the VMSwitch class from the xHyper-V module?  Just like before but with a little addition. Since VMSwitch is not an in-box resource provider, you actually have to tell the configuration to load it.

configuration Sample_cVMHost

    WindowsFeature hypervRole
    {
        Ensure = 'Present'
        Name = 'Hyper-V'
    }
    WindowsFeature hypervManagement
    {
        Ensure = 'Present'
        Name = 'Hyper-V-PowerShell'
        DependsOn = '[WindowsFeature]hypervRole'

    } 

   Import-DscResource -module xHyper-V

        xVMSwitch ExternalSwitch
        {
            Ensure         = 'Present'
            Name           = 'VMs'
            Type           = 'External'
            NetAdapterName = (Get-NetAdapter)[0].Name
            AllowManagementOS = $true 
        }

Sample_cVMHost

Start-DscConfiguration -Wait -Verbose -Path .\Sample_cVMHost

Import-DscResource brings in the custom module named ‘xHyper-V’ to allow it to be used.  Following that is a reference to xVMSwitch which is a member of that module.  And ExternalSwitch is just my arbitrary name for this section.

One tricky thing that I am doing here – that I just decided to try on my own – is declaring that the first Network Adapter would be used for the switch.  That is what (Get-NetAdapter)[0].Name does for me on the fly. 

Since this is a required field for the resource provider and for the creation of an External Virtual Switch, I have to know what it is, and I don’t want to be bothered by testing it – my test machine has one NIC, I just want it to happen. 

You could default this to the second NIC (Get-NetAdapter)[1].Name and remove the AllowManagementOS setting.

Now, you have a Hyper-V Server, with the PowerShell module, and an External Virtual switch using your first NIC that shares the management OS and is called ‘VMs’

Next time, another dependency, and my community provider – You can configure the default paths now…

No comments: