You installed CentOS 6.2 on Hyper-V and you want to install Linux Integration Services v3.2.
Let’s read documentation and see if we can.

So let’s do it on CentOS server.

You may need to uninstall Hyper-V Linux Integration Services v3.2 due to some reasons.
Let’s read documentation and see if we can.

After my first look, i don’t see any difference between x86 and x64.
Lets check it on VM console and see if they are correct.

If you have a public Hyper-V Cloud and use SCVMM, your customers can install Linux VM and install Hyper-V LIS v3.1 on it, there is nothing you can do to stop them. If they install it, you will not be able to use SCVMM due to LIS v3.1 bug.

There are 2 options to prevent this.
1) You can provision all new virtual machines without Data Exchange offer.
2) You can disable Data Exchange offer only from Linux VMs to prevent that kind of problem.

If you have 0 – 50 VMs, it’s easy to deal with. But if you are a large Cloud provider, you can’t do it manually, you need scripting. So if you execute this command on SCVMM Powershell Interface, it’ll disable Data Exchange offer on all Linux VMs and save your SCVMM.

Get-VM * | where {$_.OperatingSystem -notlike "*Windows*"} | Set-VM -EnableDataExchange $false

As you see, it just looks for non-Windows machines and disable their Data Exchange offer. Good luck!

I want to show you how to install Citrix Netscaler VPX on Hyper-V step by step.

Step 1: Create a VM on Hyper-V with synthetic network card and 2 vCPU.
It’s very important to use 2 vCPU and synthetic network card, or you can get kernel panic.

Step 2: Download VHD type of Netscaler image from Citrix, mount it to VM and start.

If you have a Hyper-V cluster in a HP Server farm then you may face with low live migration performance. We realized that HP 300 series G6 and G7 servers shipping with energy saving mode. Let’s check that from HP website.

In the ProLiant 300-series G6 and G7 RBSU, the HP Power Profile defines three possible configurations of some of the power features identified earlier in this section. The HP Power Profile provides a simple mechanism for users to configure the power management options of their system based on their tolerance to power versus performance without having to individually configure each option.

There are three possible settings for the HP Power Profile: Maximum Performance; Balanced Power and Performance; and Minimum Power Usage. An additional Custom setting is simply any combination of user settings that do not match the pre-sets for the three categories listed below.

If you want to increase performance of live migration process, you need to disable c-states from BIOS. C-States has a huge impact on live migration. Since HP comes with Balanced Power and Performance mode, it also activates C6 state on BIOS. That’s reason of the issue. So changing HP Power Profile to “Maximum Performance“, you disable C-States from BIOS and get maximum performance from Live migration. As you see our servers can utilize 99% of bandwidth.

So it’s not a good thing to use “Balanced Power and Performance” on Virtualization. Get away from it :)

You may get this error when you try to start maintenance mode on a Hyper-V host.

SCVMM can not live migrate virtual machines if a shared ISO file is attached. So that can pause maintenance mode if you choose live migration mode.
You can execute following Powershell codes to find VMs which have shared ISO files.

Get-VM | Get-VirtualDVDDrive | where {$_.ISO -ne $null} | ft Name

If you’re looking for a specific Hyper-V host called Host01:

Get-VMHost Host01 | Get-VM | Get-VirtualDVDDrive | where {$_.ISO -ne $null} | ft Name

Deattach shared ISO files from VMs and start maintenance mode again.

You may get this error when you try to refresh virtual machine.

You should refresh cluster configuration to fix this issue. Go to one of the Cluster nodes.
Execute the commands on Powershell:

1
2

Import-Module FailoverClusters
Get-ClusterResource -c CLUSTERNAME | where {$_.resourcetype.name -eq 'virtual machine configuration'} | Update-ClusterVirtualMachineConfiguration

After that, you can refresh Cluster and VM from SCVMM console. That should fix the issue.