Virtualization: Revolution of the evolving datacenter

Simplicity is Key to Private Cloud from Microsoft (cross post)

2012-04-30T19:48:28Z

Source:http://community.brocade.com/community/brocadeblogs/data_center/blog/2012/04/25/simplicity-is-key-to-private-cloud-from-microsoft

By Sue Hartford, Microsoft Global Alliance Manager at Brocade

Simplicity was a key message in Brad Anderson’s keynote at the Microsoft Management Summit in Las Vegas this month. And he’s right on target. Because let’s face it … if the cost and complexity of deploying private cloud are too high, they erase many of the cost benefits of cloud computing.

System Center 2012 delivers on that goal with many improvements to simplify IT’s efforts in building a private cloud, from its simplified licensing (now only two SKU’s which both feature all System Center modules) to simplified deployment of all the technology required to build a private cloud. One example of this is the new System Center 2012 Virtual Machine Manager (VMM) load balancer provider that allows IT admins to more easily deploy and provision network and load balancing services from within VMM.

Brocade presents the new ADX load balancer provider to provide integration between the ServerIron ADX series of application delivery switches and VMM. With this integrated solution, IT admins can use VMM to provision both application and network resources from within the VMM console. For example, IT can efficiently manage application resources, automatically provision service requests, and offer the highest application performance and service levels for users in a cost-efficient manner. The result is an agile, elastic application delivery solution for cloud-optimized environments.

Read more about it in the article by Jason Bovberg in Windows IT Pro: http://www.windowsitpro.com/article/system-center/mms-2012-brocades-adx-load-balancer-provider-offers-toptier-performance-142877 or visit www.brocade.com/systemcenter.

FI-B322 - Scripts used during MMS 2012 PowerShell session

2012-04-16T08:52:00Z

Hello everyone,

For all of those attending MMS 2012, hope you are having a great week!

As promised here is the PowerShell session (FI-B322), script used during the session.

#set script variables
$vmNamePrefix = "COMBINEDDEMO"
#set random number
$global:Rand = (new-object random)
$random = $rand.Next(100,999)
#Get Template by name
($t = Get-SCVMtemplate | where {$_.Name -eq "W2K8R2SP1_Template"}) | fl Name, OperatingSystem, VirtualizationPlatform
#create a copy of the template that you will modify
($template = New-SCVMtemplate -VMTemplate $t -Name ("Temporary" + $vmNamePrefix + "Template")) | fl Name, OperatingSystem, VirtualizationPlatform
##########
#set the OU
(Set-SCVMTemplate -VMTemplate $template -DomainJoinOrganizationalUnit "OU=Test1,DC=contoso,DC=lab") | fl Name, DomainJoinOrganizationalUnit
##########
#get runasaccount which will serve as auto logon credential
($raa = Get-SCRunAsAccount -Name "MMS Build Account") | fl Name, Username, Domain
#set auto logon credential and autologoncount (must be set together)
(Set-SCVMTemplate -VMTemplate $template -AutoLogonCredential $raa -AutoLogonCount 10) | fl Name, AutoLogonCredential, AutoLogonCount
#######
#get unattend settings object
($unattend = $template.UnattendSettings) | Out-Null
#add key/value pairs to unattend settings
#Configuration Pass ID mapping
# windowsPE = 0
# offlineServicing = 1
# generalize = 2
# specialize = 3
# auditsystem =4
# audituser = 5
# oobesystem = 6
#The relevant passes for VM deployment are oobesystem (6) and specialize (3)
$unattend.Add(“3/Microsoft-Windows-International-Core/InputLocale”, “0411:e0010411”)
$unattend.Add(“3/Microsoft-Windows-International-Core/SystemLocale”, “ja-jp”)
$unattend.Add(“3/Microsoft-Windows-International-Core/UILanguage”, “ja-jp”)
$unattend.Add(“3/Microsoft-Windows-International-Core/UILanguageFallback”, “ja-jp”)
$unattend.Add(“3/Microsoft-Windows-International-Core/UserLocale”, “ja-jp”)
$unattend.Add(“6/Microsoft-Windows-International-Core/InputLocale”, “0411:e0010411”)
$unattend.Add(“6/Microsoft-Windows-International-Core/SystemLocale”, “ja-jp”)
$unattend.Add(“6/Microsoft-Windows-International-Core/UILanguage”, “ja-jp”)
$unattend.Add(“6/Microsoft-Windows-International-Core/UILanguageFallback”, “ja-jp”)
$unattend.Add(“6/Microsoft-Windows-International-Core/UserLocale”, “ja-jp”)
$unattend.Add(“6/Microsoft-Windows-Shell-Setup/AutoLogon/Domain”, “contoso”)
$unattend.Add(“3/Microsoft-Windows-OutOfBoxExperience/DoNotOpenInitialConfigurationTasksAtLogon”, “true”)
$unattend.Add(“3/Microsoft-Windows-ServerManager-SvrMgrNc/DoNotOpenServerManagerAtLogon”, “true”)

$unattend | fl Key, Value
#commit unattend to template
(Set-SCVMtemplate -VMtemplate $template -UnattendSettings $unattend) | Out-Null
#######
#Create new VM config from template
($vmconfig = New-SCVMConfiguration -VMTemplate $template -Name ($vmNamePrefix + "_Config")) | fl Name
#Get VMhost object
($vmhost = Get-SCVMHost -ComputerName "VMMLab1823N2.contoso.lab") | fl Name, VirtualizationPlatform
#Set destination to specific host
(Set-SCVMConfiguration -VMConfiguration $vmconfig -VMHost $vmhost -VMLocation ("C:\ClusterStorage\Volume1\VMs\" + $vmNamePrefix) -ComputerName ($vmNamePrefix + $random)) | fl Name, ComputerName, VMLocation, VMHost
#update VM config object with placement recommendations
Update-SCVMConfiguration -VMConfiguration $vmconfig | Out-Null
#create the differencing disk
$VHDService = get-wmiobject -class "Msvm_ImageManagementService" -namespace "root\virtualization" -computername $vmhost.Name
($VHDService.CreateDifferencingVirtualHardDisk(("C:\ClusterStorage\Volume1\VMs\" + $vmNamePrefix + "\" + $vmNamePrefix + ".VHD"),"C:\ClusterStorage\Volume1\BASEVHDs\W2K8R2SP1.VHD")) | fl ReturnValue
#Update VHD information
(Set-SCVirtualHardDiskConfiguration -VHDConfiguration $vmconfig.VirtualHardDiskConfigurations[0] -DestinationLocation ("C:\ClusterStorage\Volume1\VMs\" + $vmNamePrefix) -FileName ($vmNamePrefix + ".VHD") -DeploymentOption "UseExistingVirtualDisk") | fl DeploymentOption, FileName, DestinationLocation
#create VM
($vm = New-SCVirtualMachine -VMConfiguration $vmconfig -Name $vmNamePrefix -StartVM) | fl Name, VirtualizationPlatform, HostGroupPath

#####
#send power state info to VM
(Set-SCGuestInfo -VM $vm -Key "PowerState1" -Value "1. ON")
#####
#Suspend VM and write KVPs
Suspend-SCVirtualMachine -VM $vm | Out-Null
#send power state info to VM
(Set-SCGuestInfo -VM $vm -Key "PowerState2" -Value "2. Pause")
#####
#Save state VM
Resume-SCVirtualMachine -VM $vm | Out-Null
Stop-SCVirtualMachine -VM $vm -SaveState | Out-Null
#send power state info to VM
(Set-SCGuestInfo -VM $vm -Key "PowerState3" -Value "3. Save")
#####
#shutdown VM and write KVPs
Start-SCVirtualMachine -VM $vm | Out-Null
Stop-SCVirtualMachine -VM $vm -Shutdown | Out-Null
#send power state info to VM
(Set-SCGuestInfo -VM $vm -Key "PowerState4" -Value "4. OFF")
#####
#start VM and check
Start-SCVirtualMachine -VM $vm | Out-Null
#use KVP to pass in new IP information
#get basic information
(Read-SCGuestInfo -VM $vm -Key FullyQualifiedDomainName | select KVPMAP).KVPMAP
#get multiple values
$info = @{"FullyQualifiedDomainName"="";"NetworkAddressIPv4"="";"NetworkAddressIPv6"=""}
(Read-SCGuestInfo -VM $vm -KVPMap $info | Select KVPMAP).KVPMAP
#Pass new IP settings
#get mac address of NIC
Read-SCVirtualMachine -VM $vm | Out-Null
($nic1mac = $vm.VirtualNetworkAdapters[0].MacAddress)
#build kvp map with new IP information
$NICIP = @{("["+$nic1mac+"]"+"IPAddress")="10.1.1.34";("["+$nic1mac+"]"+"Mask")="255.255.255.0";("["+$nic1mac+"]"+"Gateway")="10.1.1.1"}
#send ip info to VM
(Set-SCGuestInfo -VM $vm -KVPMAP $NICIP | Select KVPMAP).KVPMAP
#cleanup
Stop-SCVirtualMachine -VM $vm -Force | Out-Null
Remove-SCVirtualMachine -VM $vm | Out-Null
Remove-SCVMConfiguration -VMConfiguration $vmconfig | Out-Null
Remove-SCVMTemplate -VMTemplate $template | Out-Null

This is a registry partial I stored in he Library. This gets called by the guest OS as part of first logon (in script abov)

Windows Registry Editor Version 5.00
[HKEY_CURRENT_USER\Software\Microsoft\Windows\CurrentVersion\Applets\Regedit\Favorites]
"OU"="HKEY_LOCAL_MACHINE\\SOFTWARE\\Microsoft\\Windows\\CurrentVersion\\Group Policy\\State\\Machine"
"GUESTINFO"="HKEY_LOCAL_MACHINE\\SOFTWARE\\Microsoft\\Virtual Machine\\External"
"AUTOLOGON"="HKEY_LOCAL_MACHINE\\SOFTWARE\\Microsoft\\Windows NT\\CurrentVersion\\Winlogon"

This is the unattend.xml partial stored in Library and used in the template.

<?xml version="1.0" encoding="utf-8"?>
<unattend xmlns="urn:schemas-microsoft-com:unattend">
    <settings pass="oobeSystem">
        <component name="Microsoft-Windows-Shell-Setup" processorArchitecture="amd64" publicKeyToken="31bf3856ad364e35" language="neutral" versionScope="nonSxS" xmlns:wcm="http://schemas.microsoft.com/WMIConfig/2002/State" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance">
            <UserAccounts>
                <DomainAccounts>
                    <DomainAccountList wcm:action="add">
                      <Domain>dcmanager.lab</Domain>
                        <DomainAccount wcm:action="add">
                          <Group>Administrators</Group>
                          <Name>mmsbuildaccount</Name>
                        </DomainAccount>
                    </DomainAccountList>
                </DomainAccounts>
            </UserAccounts>
          <FirstLogonCommands>
     <SynchronousCommand wcm:action="add">
              <CommandLine>regedit.exe /s \\vmmteched\demo\MMSDEMOREGFAV.reg</CommandLine>
              <Description>Add favorites to registry</Description>
              <Order>1</Order>
            </SynchronousCommand>
          </FirstLogonCommands>
     <SynchronousCommand wcm:action="add">
              <CommandLine>regedit.exe</CommandLine>
              <Description>Launch registry editor</Description>
              <Order>2</Order>
            </SynchronousCommand>
          </FirstLogonCommands>
        </component>
    </settings>
</unattend>

Analyzing Storage Performance using the Windows Performance Analysis ToolKit (WPT) (cross post)

2012-03-10T19:16:06Z

Excellent post by Robert Smith, PFE:

Introduction

Performance problems, especially as they may relate to the storage subsystem, can be quite difficult to troubleshoot. Enterprise storage technology has come a long way since the SCSI controller with an array of disks. Fortunately, there are some great tools available to help narrow down where to look more closely for storage performance problems. This blog post covers the Windows Performance Analysis Toolkit (WPT), as used for analyzing performance in the storage subsystem.

The facility that enables the analysis I am about to cover is called "Event Tracing for Windows" (ETW). The Performance Analyzer is built on top of the ETW infrastructure. ETW enables Windows and applications to efficiently generate events, which can be enabled and disabled at any time without requiring system or process restarts. ETW collects requested kernel events and saves them to one or more files referred to as "trace files" or "traces." These kernel events provide extensive details about the operation of the system. Some of the most important and useful kernel events available for capture and analysis are context switches, interrupts, deferred procedure calls, process and thread creation and destruction, disk I/Os, hard faults, processor P-State transitions, and registry operations, though there are many others.

One of the great features of ETW, supported in WPT, is the support of symbol decoding, sample profiling, and capture of call stacks on kernel events. These features provide very rich and detailed views into the system operation. WPT also supports automated performance analysis. Specifically, xperf is designed for scripting from the command line and can be employed in automated performance gating infrastructures (it is the core of Windows PerfGates). xperf can also dump the trace data to an ANSI text file, which allows you to write your own trace processing tools that can look for performance problems and regressions from previous tests.

The following information will be mostly about the WPT tool called "Xperf.exe". Xperf.exe is the command line tool used to start, stop, and manage traces. The usage of Xperf.exe is documented thoroughly in the help file included with the WPT titled "WindowsPerformanceToolkit.chm".

Read the full post here: http://blogs.technet.com/b/robertsmith/archive/2012/02/07/analyzing-storage-performance-using-the-windows-performance-toolkit.aspx

SC 2012 - Virtual Machine Manager Storage Automation with the IBM XIV Storage System Gen3

2012-02-09T18:53:26Z

IBM recently announced support for SC 2012 - VMM storage automation with the IBM XIV storage system using SMI-S.

Microsoft® System Center Virtual Machine Manager 2012 Storage Automation With the IBM XIV Storage System Gen3

This technical white paper reviews key concepts and provides step-by-step implementation examples for Microsoft System Center Virtual Machine Manager (SCVMM) 2012 new storage automation features. Microsoft SCVMM 2012 now provides all administrators, including virtualization, database, systems, storage, etc., the ability to perform a full spectrum of core storage management tasks within a true central administrative interface. Within SCVMM’s graphical user interface (GUI), administrators can discover, classify, allocate, provision, map, assign and decommission storage associated with clustered and standalone virtualization hosts. A close examination of popular IBM XIV Storage System Gen3 enterprise-class storage solutions reveals extensibility and performance improvements which align with the Microsoft SCVMM 2012 new features.

http://www-03.ibm.com/support/techdocs/atsmastr.nsf/WebIndex/WP102071

SC 2012 VMM: Collecting storage related traces

2011-12-08T21:32:00Z

SC 2012 VMM ships with a new service called the Storage Management Service which communicates with SMI-S based providers from vendors like NetApp, EMC, HP, and IBM. To help you troubleshoot problems, SCVMM include lots of storage related tracing information in its own logs and whenever possible, the CIMXML output from the providers. In some cases however, you will need to get CIMXML output and trace output from the Storage Management service directly to help you troubleshoot further.

So there are three levels of tracing that you need to keep in mind:

Enable VMM tracing

SCVMM traces will give you the error/exception information. Make sure to collect traces on the VMM server. Hyper-V host traces are only necessary if the failure us on the Hyper-V side (e.g. volume mount issues)

Refer to this KB article for instructions on how to setup SCVMM tracing: http://support.microsoft.com/kb/970066

Enable SCX CIM-XML command tracing

Storage Management Service communicates with SMI providers using CIMXML. This output is the raw call and response interaction between the service and the provider. This information is very verbose so collecting this information only when you reproduce the issue will help minimize the noise.

Add registry subkey

W2K8R2/SP1: HKLM\Software\Microsoft\Storage Management\CIMXMLLog
W2K12: HKLM\SOFTWARE\Microsoft\Windows\CurrentVersion\Storage Management\CIMXMLLog

Add registry DWORD value LogLevel, value 4
Add registry String value LogFileName and assign it to the full path and file name to use for logging. Make sure the directory exists and that the Network Service account has read/write access
Stop and start “Microsoft Storage Management Service”

Here is a sample of the information contained in the trace:

Traceview ETL

Finally, the Storage Management Service has its own trace output which you collect using Traceview.

Download WDK 7.1 http://www.microsoft.com/download/en/details.aspx?displaylang=en&id=11800
and install on VMM server
Unzip traceview.zip (attached to this blog post) to a local folder where VMM is installed
Copy traceview.exe from WDK folder to this folder
Run traceview.exe as administrator
File->Open Workspace, select scx and click ok

With this tracing will start but you might not see anything in the UI until SCVMM communicates with the Storage Management Service. Refresh a provider to ma

The information in the UI also gets logged to storageservice.etl file in same folder as TraceView.

Windows Server 8: Standards-Based Storage Management

2011-11-14T16:51:14Z

Great post from Jeffrey Snover

Windows Server 8 is a cloud optimized OS. You’ve probably heard that phrase or similar comments a lot recently since we introduced Windows Server 8 last month (see Bill Laing’s blog post.) Let’s drill in a bit to explain what that really means and why it matters to you. In the past, Windows Server was a great OS for a server and its devices. Windows Server 8 is a great OS for lots of servers and all the devices connecting them whether they are physical or virtual, on-premise or off-premise.

At the BUILD conference, we showcased for the first time all the new capabilities that deliver on this vision including scalability, availability, Hyper-V, networking, manageability and storage.

Whether we are talking classic architectures or private/public/hybrid cloud architectures, one thing is absolutely clear – there is a large and growing appetite for data. Customers’success is predicated on the efficient and effective management of storage. Windows Server 8 is there to help meet that challenge whether the storage is directly attached to a server or is an external storage array. Working with our storage partners, we are delivering a new set of capabilities, APIs and PowerShell Cmdlets.

Windows Server 8 introduces a new WMI-based API called the Storage Management API (SMAPI) and corresponding set of PowerShell Cmdlets. These provide storage management primitives to manage direct attach storage on the OS as well as external storage arrays. The PowerShell Cmdlets replace tools like diskpart and diskraid. The API is comprised of a WMI object model along with the corresponding set of methods and properties. Storage partners plug into the new API either by:

Implementing the SNIA (Storage Networking Industry Association) industry standard for storage management called SMI-S (Storage Management Initiative –Specification)
Implementing a new provider model called the Storage Management Provider (SMP)

Read the rest here: http://blogs.technet.com/b/server-cloud/archive/2011/10/14/windows-server-8-standards-based-storage-management.aspx

SCVMM 2012: Live from SNIA SMILab Plugfest – Creating 50 VMs using SAN based Rapid Provisioning

2011-10-20T18:54:00Z

Today is the 4th and final day of the SNIA SMILab plugfest. One of the tests we use to validate SMI-S providers with SCVMM is to create 50 VMs in parallel using SAN based Rapid Provisioning. We do this in two modes: 50 all at the same time or 50 in batches of 10 (so 5 batches). I kicked off the former test and was able to create 50 VMs on Dell Compellent array. I will not cover details on time elapsed or maximum scale. I will leave that for a bigger more detailed blog post.

Here are a few screen shots:

50 VMs on 3 hosts

List VMs on one of the hosts

Logical units (writeable snapshots)

50 create VM jobs

This test was run using VMM 2012 RC

SCVMM 2012 Quick Tip: Finding which storage arrays a Hyper-V host can see

2011-10-18T17:19:00Z

SNIA SMI Lab kicked off the 4^th and final plugfest for 2011 this week. Microsoft is here along with several storage partners. Earlier today, we started testing VMM 2012 with a SMI-S based storage provider and we tried to allocate a LUN to a host group that contained a Hyper-V host. This operation succeeded with the following warning 26130:

The recommended action tells me that zoning might not be correct. So I check host properties and notice that the host cannot see any arrays. This means either the iSCSI initiator is not logged into any targets and there is no zone created on the FC switch that includes the host and array WWNs.

I can also access this information using VMM PowerShell. In the SNIA SMI lab, the Hyper-V host is connected to multiple arrays using both FC and iSCSI. Using VMM cmdlets, I can figure out if the host is properly zoned to array and/or logged into target of array and get detailed information about the target names and port WWNs using PowerShell:

#establish connection to VMM server

Get-SCVMMServer localhost | Out-Null

$arrays = Get-SCStorageArray

#collect FC endpoints from array object

$FCarrayendpointaddress = @()

foreach($array in $arrays){

foreach($FCstorageendpoint in $array.StorageEndpoints){

if($FCarrayendpointaddress -notcontains $FCstorageendpoint.Address){

$FCarrayendpointaddress += $FCstorageendpoint.Address

}

#collect target endpoint address from array object

$iSCSIarrayendpointaddress = @()

foreach($array in $arrays){

foreach($portal in $array.StorageiSCSIPortals){

foreach($iSCSIendpoint in $portal.StorageEndpoints){

if($iSCSIarrayendpointaddress -notcontains $iSCSIendpoint.Address){

$iSCSIarrayendpointaddress += $iSCSIendpoint.Address

}

$vmhost = Get-SCVMHost | where {$_.Name -eq "MSSCVMMHost1.msscvmm.com"}

#collect discovered FC portwwns from vmhost object

$vmhostportwwn = @()

foreach($hba in $vmhost.FibreChannelHbas) {

foreach($fibreport in $hba.DiscoveredFibrePorts)
{

$vmhostportwwn += $fibreport.PortWWN

}

#collect target names vmhost iSCSI initiator is logged into

$vmhosttargetnames = @()

foreach($hba in $vmhost.InternetSCSIHbas){

foreach($target in $hba.Targets){

$vmhosttargetnames += $target.targetname

}

#compare endpoints exposed to vmhost using portwwn information

$FCexposedendpoints = @()

foreach($portwwn in $vmhostportwwn){

if($FCarrayendpointaddress -contains $portwwn){

if($FCexposedendpoints -notcontains $portwwn){

$FCexposedendpoints += $portwwn

}

#compare targets exposed to vmhost using hba target name information

$iSCSIexposedendpoints = @()

foreach($targetname in $vmhosttargetnames){

if($iSCSIarrayendpointaddress -contains $targetname){

if($iSCSIexposedendpoints -notcontains $targetname){

$iSCSIexposedendpoints += $targetname

}

#output results

"FC arrays zoned to: " + $vmhost.Name

if($FCexposedendpoints.Count -eq 0){

"`tHost cannot see any FC arrays"

} else{

"`tExposed endpoints:"

foreach($array in $arrays){

foreach($FCstorageendpoint in $array.StorageEndpoints){

if($FCexposedendpoints -contains $FCstorageendpoint.Address){

"`t`t" + $array.Name + ": " + $FCstorageendpoint.Address

}

"iSCSI arrays logged into from: " + $vmhost.Name

if($iSCSIexposedendpoints.Count -eq 0){

"`tHost cannot see any iSCSI arrays"

} else{

"`tExposed endpoints:"

foreach($array in $arrays){

foreach($portal in $array.StorageiSCSIPortals){

foreach($iSCSIendpoint in $portal.StorageEndpoints){

if($iSCSIexposedendpoints -contains $iSCSIendpoint.Address){

"`t`t" + $array.Name + ": " + $iSCSIendpoint.Address

}

Sample output will look like this

FC arrays zoned to: MSSCVMMHost1.msscvmm.com

Exposed endpoints:

EMC VMAX-1SE: 50000972C006251C

EMC VMAX-1SE: 50000972C0062518

iSCSI arrays logged into from: MSSCVMMHost1.msscvmm.com

Host cannot see any iSCSI arrays

If you only need to know which array (by name) the Hyper-V host is exposed to (iSCSI and/or FC), you can also do this:

#establish connection to VMM server

Get-SCVMMServer localhost | Out-Null

$vmhost = Get-SCVMHost | where {$_.Name -eq "MSSCVMMHost1.msscvmm.com"}

$arrays = Get-SCStorageArray -VMHost $vmhost

foreach($array in $arrays){

$vmhost.Name + " is can
see array: " + $array.Name

}

Sample output will look like this

MSSCVMMHost1.msscvmm.com is can see array: EMC VMAX-1SE

In the VMM console, you will see this:

Windows 8 #bldwin - Haters are gonna hate (bye bye iDevice)

2011-09-14T05:02:36Z

Some of these comments I have seen on FB and Twitter are just silly. For those that think "this is too much change"... did you even watch the key note?

1) The new Metro UI is an optimized experience for touch. You still have access to the traditional desktop experience. The user is in control of their experience. The developer is in control of optimizing the experience they want their customers to use

2) The new WindowsRT runtime live SxS with the traditional API stack. You don’t have to compromise either way. You can optimize for both.

3) You still have full keyboard and mouse access to Windows

4) A core fundamental is backwards compatibility. So whatever worked in Win7 will continue to work Win8

5) The experience changes in Windows do not boil down to a simple UI skin. There are fundamental changes made to the entire OS (Kernel up) to optimize for different form factors, hardware devices, battery life, peripherals, etc

6) did you see the HW lineup? Low-end to super high-end... same OS, same experience, optimized for the hardware available.

So either a) You did not watch the keynote (if that is the case then you should do so ASAP http://www.buildwindows.com/) or b) you simply hate Windows and don’t care for any of this (then just say it and stop pontificating how "This is too much of a change" and "it will be a long time for people to adopt") or c) you are paralyzed by change

For those in the b) bucket... did you have the same negative feedback for Apple when they introduced all their new constructs, dropped flash support, locked out all hardware vendors? If yes, then great... at least you are fair. If not, then you will soon be in the minority. Keep enjoying shelling out $500 - $1000 dollars every 9 months to get your iDevice fix.

System Center Virtual Machine Manager (VMM) 2012 RC - now available for download

2011-09-09T05:35:00Z

http://blogs.technet.com/b/keithcombs/archive/2011/09/09/system-center-virtual-machine-manager-vmm-2012-rc-now-available-for-download.aspx?utm_source=dlvr.it&utm_medium=twitter&utm_campaign=hypervking

Keith Combs - System Center Virtual Machine Manager (VMM) 2012 delivers industry leading fabric management, virtual machine management and services deployment in private cloud environments. VMM offers key new features that include hypervisor creation and management, network management, storage management, private cloud creation, self-service usage and service creation. It features deep investments in server application virtualization, service design and service modeling all of which can be used to efficiently offer an on-premises private cloud.

Feature Summary:

Fabric Management
- Setup Upgrade
  - *New in RC: The following upgrade paths are enabled (supported only for TAP customers in production):
    - VMM 2008 R2 SP1 --SC VMM 2012 RC -- SC VMM 2012 RTM
    - SC VMM 2012 RC -- SC VMM 2012 RTM
- Hyper-V and Cluster Lifecycle Management – Deploy Hyper-V to bare metal server, create Hyper-V clusters, orchestrate patching of a Hyper-V Cluster
  - *New in RC:
    - ISO or CD-based OSD for environments with DHCP without WDS
    - OSD will now convert dynamic to fixed type of VHD destination
    - All network adapters on host can be configured during provisioning
  - *New in RC:
    - Ability to bypass cluster validation during cluster creation
    - Run cluster validation reports on-demand
    - New Cluster status tab to view an aggregated status and a cluster validation report
    - Ability to see current CSV owner in the properties of the cluster
- Third Party Virtualization Platforms - Add and Manage Citrix XenServer and VMware ESX Hosts and Clusters
- Network Management – Manage IP Address Pools, MAC Address Pools and Load Balancers
  - *New in RC:
    - Simplification of the logical networks in the Fabric workspace
    - Ability to see IP addresses that are in use from a IP pool
    - Added support for Microsoft Network Load Balancer
    - Gateway and DNS are no longer mandatory fields for logical networks
    - Load balancer can now support affinity to logical networks
- Storage Management – Classify Storage, Manage Storage Pools and LUNs
  - *New in RC
    - Create persistent sessions to iSCSI array and logon initiator to array
    - Better scalability of storage operations - LUN create, snapshot, clone, masking, and unmasking
    - Option to create storage groups per cluster (BETA only supported creation of storage group per node in a cluster)
    - Enablement of MPIO feature when provisioning a new Hyper-V server
    - Automatic MPIO device claim
    - Support for arrays that implement OnePortPerView
- Update Management- Keep your VMM Fabric Servers (VMM roles, hosts, and clusters) up-to-date with patches.
  - *New in RC:
    - Share a WSUS root server between System Center Configuration Manager 2007 R2/ System Center Configuration Manager 2012 Beta
    - Hyper-V Cluster Orchestration- Nodes put into VMM Maintenance Mode can be set to trigger Maintenance Mode in Operations Manager.
- Resource Optimization
  - Dynamic Optimization – proactively balance the load of VMs across a cluster
  - Power Optimization – schedule power savings to use the right number of hosts to run your workloads – power the rest off until they are needed.
    - *New in RC:
      - Set Operations Manager Mode for powered down hosts
  - PRO – integrate with System Center Operations Manager to respond to application-level performance monitors.
    - *New in RC:
      - Support added for System Center Operations Manager 2012 Beta
      - VMM will ship two sample PRO Packs: Cluster scale out and Service scale out MPs
Cloud Management
- Abstract server, network and storage resources into private clouds
- Delegate access to private clouds with control of capacity, capabilities and user quotas
- Enable self-service usage for application administrator to author, deploy, manage and decommission applications in the private cloud
Service Lifecycle Management
- Define service templates to create sets of connected virtual machines, OS images and application packages
  - *New in RC:
    - Service Designer and Specialization UI enhancements
    - Added ability to use Service Template Patterns
- Compose operating system images and applications during service deployment
  - *New in RC:
    - IP-based provisioning
    - New application instance view
- Scale out the number of virtual machines in a service
- Service performance and health monitoring integrated with System Center Operations Manager
- Decouple OS image and application updates through image-based servicing
  - *New in RC:
    - Streamlined ability to enable OS VHD updates to a Service Template
    - Publish updated Service Templates in order to update Service Instances
- Leverage powerful application virtualization technologies such as Server App-V

Whew! Go get it @ http://www.microsoft.com/download/en/details.aspx?displaylang=en&id=27252.

Loading VMM and vSphere cmdlets into the same PowerShell Console

2011-08-24T01:40:00Z

Hello everyone,

This is Hector Linares, Program Manager for storage automation and with help from Erwien Saputra, UI developer on the Virtual Machine Manager team. We want to show you how to load VMM and vSphere cmdlets in the same PowerShell console.

Administrators that need to manage both Hyper-V and ESX in their environment, can use VMM to manage both. However, sometimes it is necessary to configure specific setting only available in vSphere. By loading VMM and vSphere cmdlets into the same console gives you access to both VMM and vSphere from a single console.

To set this up:

1. Download and install VMware vSphere PowerCLI on the same machine that has VMM Administration Console: http://communities.vmware.com/community/vmtn/server/vsphere/automationtools/powercli

2. Create a folder ToolBox in %SYSTEMDRIVE% (this post assumes %SYSTEMDRIVE% but this can be in any location)

3. Create sub-folder VMMCLI in ToolBox folder

4. Create PowerShell script file named LoadScript.ps1 and save it to %SYSTEMDRIVE%\ToolBox\VMMCLI. The file should contain the following script:

if ((Get-ExecutionPolicy) -eq 'Restricted') {

Set-ExecutionPolicy RemoteSigned -Scope Process -Force}

ipmo 'C:\Program Files\Microsoft System Center 2012\Virtual Machine Manager\bin\psModules\virtualmachinemanager\virtualmachinemanager.psd1'

Add-PSSnapIn -NameVMWare.VimAutomation.Core

& "C:\Program Files (x86)\VMware\Infrastructure\vSphere PowerCLI\Scripts\Initialize-PowerCLIEnvironment.ps1"

CLS

$productName = "VMM and vSphere PowerShell Console"

$host.ui.RawUI.WindowTitle = "[$productName]"

5. Create a shortcut to PowerShell console and save it to %SYSTEMDRIVE%\ToolBox\VMMCLI. The shortcut should have the following setting:

Target: C:\Windows\System32\WindowsPowerShell\v1.0\powershell.exe -noe -c "Set-Location -Path %SYSTEMDRIVE%\toolbox\VMMCLI; .\loadscript.ps1"

6. You can now double click on the shortcut and get a PowerShell console with
VMM and vSphere cmdlets

Keep in mind:

1. This script assumes you have the latest VMM build installed. If you are using BETA, you will need to change VMM path:

- VMM 2012 BETA: “C:\Program Files\Microsoft System Center Virtual Machine Manager 2012”

- VMM 2012 RC/RTM: “C:\Program Files\Microsoft System Center 2012\Virtual Machine Manager”

2. Even though all VMM cmdlets have been renamed, some vSphere cmdlets overlap with VMM alias (e.g. Get-VM). The cmdlets loaded last take precedence so if you execute Get-VM for example, the vSphere implementation of Get-VM will be invoked. VMM’s new name for get-VM is Get-SCVirtualMachine.

Happy Scripting!

List all the VMs hosted on a specific SAN array

2011-07-27T05:17:29Z

Hey everyone,

While preparing for my TechReady session, I reached out to a fellow dev on the team to help me craft a script to list all the VMs hosts on a specific SAN array. This is what he came back with. Thanks Erwien!

Enjoy

($array = Get-SCStorageArray | where {$_.Name -eq "FAS2040-SCVMM"}) | fl Name, TotalCapacity, IsCloneCapable, IsSnapshotCapable, IsSanTransferCapable, StoragePools
($pools = $array.StoragePools | where {$_.IsManaged -eq $True}) | fl Name, TotalManagedSpace, Classification, StorageLogicalUnits

$luns = @()
foreach ($pool in $pools) {
$assignedLuns = $pool.StorageLogicalUnits | where {$_.IsAssigned -eq $True -and $_.HostDisks.Count -ne 0}
$luns += $assignedLuns
}

$vmHosts = @()
$disks = @()
foreach($lun in $luns) {
    $vmHost = $lun.HostDisks[0].VMHost
    if ($vmHosts -notcontains $vmHost) {
        $vmHosts += $vmHost
    }

    foreach($disk in $lun.HostDisks) {
      if ($disks -notcontains $disk) {
          $disks += $disk
      }
    }
}

$vms = @()

foreach($vmHost in $vmHosts) {
if($vmHost -ne $null) {
    $vms += Get-VM -VMHost $vmHost
}
}
$vmWithArray = @()
foreach($vm in $vms) {
    foreach($passthroughDisk in $vm.PassThroughDisks) {
        if ($disks -contains $passthroughDisk) {
            $vmWithArray += $vm
        }
    }

    foreach($vdd in $vm.VirtualDiskDrives | where {$_.IsVHD}) {
        $vhd = $vdd.VirtualHardDisk

        if ($vhd.HostVolume -ne $null -and $vhd.HostVolume.StorageDisk -ne $null) {
            if (($disks -contains $vhd.HostVolume.StorageDisk) -and
                ($vmWithArray -notcontains $vm)) {
                $vmWithArray += $vm
            }
        }
    }
}
$vmWithArray | fl Name, VMHost

How much faster is IE10 compared to IE9, Chrome 11, and FireFox 4

2011-04-13T08:09:00Z

So I downloaded and installed the IE10 Platform Preview today and started to check out what's new. One of the tests Microsoft provides is called PaintBall. I ran this test using IE10 and noticed that the test displays "seconds" and "paitballs per minute" after the test. Since I had nothing to compare the results to, I decided to run a few tests myself.

I first installed the following browsers in addition to having IE9 (9.0.8112.16421) RTM and IE10 (10.0.1000.16394) Platform Preview:

- Chrome (11.0.696.43 beta)

- FireFox (4.0)

I ran the PaintBall test 3 times in each browser:

IE9	- Average seconds to complete test: 19.60 - Average paintballs per minute: 266.38
IE10	- Average seconds to complete test: 18.11 - Average paintballs per minute: 288.20
Chrome 11	- Average seconds to complete test: 120.90 - Average paintballs per minute: 43.17
FireFox 4	- Average seconds to complete test: 122.00 - Average paintballs per minute: 42.79

So the obvious winners are Chrome and FireFox... NOT. IE9 and IE10 performed well in this test. What is more telling however, is the visual end result of this test. Both IE9 and IE10 render the final visual very well. Chrome and FireFox fail.

IE9
IE10
Chrome 11
FireFox 4	Interesting about FireFox is that the visual did not render correctly with the black background. When I clicked on the link to restart the test, the blackground swept in finally revealing the seconds and paintballs per minute results

I noticed that after completing the test in FireFox, my system was reduced to a crawl. Processor Explorer revealed the culprit - DirectX:

XPERF provided more information about the behavior:

Congrats to the IE team for an excellent set of IE releases these past few months!!!

Storage Automation in VMM 2012 - Overview

2011-04-07T22:13:57Z

Hi, I am Hector Linares, Senior Program Manager on the SCVMM team. This blog is an overview of the storage automation feature set in System Center Virtual Machine Manager (VMM) 2012. The goal of this feature is to simplify and automate the discovery, allocation, and assignment of storage assets in a virtualized environment.

Summary

VMM 2012 introduces enhanced networking, storage, and physical server management that enables enterprises to create on premise private clouds. The key investments in storage include modeling and automation. Modeling helps administrators understand what storage assets they have and how the underlying SAN infrastructure relates to their Hyper-V environment. Automation helps streamline common storage workflows – provisioning new storage, adding capacity to a Hyper-V or cluster, and rapid provisioning of new VMs using the SAN.

The automation of storage in the context of a Hyper-V environment helps reduce the time to market of new virtual servers, reduce the complexity of using the SAN in a virtualization environment, and take advantage of all the capabilities of the underlying SAN while delivering excellent performance. This level of integration will help customers fully leverage the SAN for their Hyper-V environment.

VMM 2012 uses a new storage service that communicates to SMI-S based storage providers for active management of storage arrays. Leveraging an industry standard like SMI-S helps VMM deliver value-add functionality to customers across a wide range of storage devices while minimizing the engineering effort and time to market

Funny stuff on the internet (positive demos of Hyper-V Dynamic Memory by the competition)

2011-04-02T00:28:57Z

Hey everyone,

There is so much random stuff on the web these days. Some content is good for a laugh but then there is that content that leaves you scratching your head. Over at v_Critical, looks like their investigatory prowess has uncovered that something's afoot with Hyper-V.

According to the article, the Hyper-V dynamic memory demo at TechEd this year presented by Bryon Surace did a very BAD thing. It showed 20 VMs created on a single host with 14 powered on, but only one actually showing dynamic memory demand. Naturally it was time to conjur up conspiracy theories about why this was the case. Ultimately the writer landed on the conclusion that this was a "fake" demo.

The writer of the article goes off and installs Hyper-V in his lab, creates 12 VMs with Win7 installed, and dynamic memory enabled to prove only three things:

Dynamic memory works as expected ("Dynamic Memory is actually enabled and working here, as the assigned memory for each VM has increased")
Hyper-V can run many VMs with dynamic memory enabled ("In fact, a dozen Windows 7 SP1 VMs– running in my own lab")
Win7 uses memory ("SP1 VM is started up on Hyper-V it almost immediately demands more than the meager 384MB")

So then (and here is the funny part), the writer finishes the article with:

"Clearly the Dynamic Memory demo at MMS 2011 was misleading, but why? Could Hyper-V Dynamic Memory have accommodated that same VM density if operating systems were actually running?"

Based on the slides presented at TechEd, the content was about explaining dynamic memory in depth. The demo focused on showcasing dynamic memory "in action" in line with the content presented.

However, maybe the writer of the article can present Hyper-V dynamic memory at the next Microsoft event since he has so much experience showing how well it works and how well it scales with many VMs running.

So I guess thanks for demoing Hyper-V :-)

Using SCVMM 2012, NetApp SMI-S provider, and Visio to visualize storage

2011-03-30T10:14:00Z

Hello everyone,

So the VMM team finally announced BETA at MMS in March. One request we heard from you during the event is the need to visualize storage. VMM 2012 goes a long way to integrate storage automation into VMM using SMI-S based providers. Through these providers, VMM gets a lot of great data. You can use this data to visualize your storage environment. Below is one example of how I modified an existing NetApp PowerShell script that generates a Visio diagram with aggregate, volume, and LUN information.

Check out NetApp for Microsoft Environments blog for the latest and greatest news from NetApp, including scripts, tools, and best practices - http://blogs.netapp.com/msenviro/

For this example, I downloaded the following:

Download the NetApp Visio stencils and extract them all into your "My Documents\My Shapes" directory. if the link is broken then go to http://www.visiocafe.com/downloads/netapp/ too get the latest.

Make sure to launch this script from a client that has the VMM administrator console installed on it. This script assumes LOCALHOST.

Here is a sample of a small environment:

And a more complicated environment:

##script##

$shpFile1 = "\NetApp-Logic-Icons.vss"
$shpFile2 = "\NetApp-Equipment-Icons.vss"

function connect-visioobject ($firstObj, $secondObj) {
$shpConn = $pagObj.Drop($pagObj.Application.ConnectorToolDataObject, 0, 0)

#// Connect its Begin to the 'From' shape:
$connectBegin = $shpConn.CellsU("BeginX").GlueTo($firstObj.CellsU("PinX"))
#// Connect its End to the 'To' shape:
$connectEnd = $shpConn.CellsU("EndX").GlueTo($secondObj.CellsU("PinX"))
}

function add-visioobject ($mastObj, $item) {
Write-Host "Adding $item"
# Drop the selected stencil on the active page, with the coordinates x, y
$shpObj = $pagObj.Drop($mastObj, $x, $y)
# Enter text for the object
$shpObj.Text = $item
#Return the visioobject to be used
return $shpObj
}

#Establish connection to VMM Server (assumes the VMM server is running on the same machine executing the script)
$vmm = Get-VMMServer localhost | Out-Null

#list the arrays under VMM's management
Get-SCStorageArray | ft Name, Model, FirmwareVersion

$FASName = read-host "Enter the Name of your NetApp array"
If ($FASName -eq "") { Write-Host "No selection made, script now exiting." ; exit }

#Get array, pools, LUNs
$array = Get-SCStorageArray -Name $FASName.Trim()
$allPools = @()
$allPools = Get-SCStoragePool | where {$_.StorageArray -eq $array}
$allLUNs = @()
$allLUNs = Get-SCStorageLogicalUnit | where {$allPools -contains $_.StoragePool}

#construct the volumes manually from the list of LUNs and their containing pool
$allVols = @()
#get all volumes
Foreach ($lun in $allLUNs)
{
    $volName = $lun.Name.split("/")
    $volume = New-Object PSObject
    Add-Member -MemberType NoteProperty -Name Name -Value ($volName[2] + "/") -InputObject $volume
    Add-Member -MemberType NoteProperty -Name StoragePool -Value $lun.StoragePool -InputObject $Volume
    $hashmatch = $false
    $allVols += $volume
}
$allVols = $allVols | Select-Object -Unique Name, StoragePool

# Create an instance of Visio and create a document based on the Basic Diagram template.
$AppVisio = New-Object -ComObject Visio.Application
$AppVisio.AutoLayout = $true
$docsObj = $AppVisio.Documents
$DocObj = $docsObj.Add("Basic Network Diagram.vst")
# Set the active page of the document to page 1
$pagsObj = $AppVisio.ActiveDocument.Pages
$pagObj = $pagsObj.Item(1)

# Load a set of stencils and select one to drop
$stnPath = [system.Environment]::GetFolderPath('MyDocuments') + "\My Shapes"
$stnObj1 = $AppVisio.Documents.Add($stnPath + $shpFile1)
$FlexVOLObj = $stnObj1.Masters.Item("FlexVol")
$poolObj = $stnObj1.Masters.Item("Raid Grp Aggregate Storage")
$LUNObj = $stnObj1.Masters.Item("Cylinder")
$stnObj2 = $AppVisio.Documents.Add($stnPath + $shpFile2)
$FASObj = $stnObj2.Masters.Item("FAS3000 Double controllers")

#Draw objects
$y = $allPools.Count * 1.50 / 2
$x = 1.50
$FASObj = add-visioobject $FASObj $FASName

$x = 3.50
$y += 2

Foreach ($pool in $allPools) {
$poolObj = add-visioobject $poolObj $pool.Name
connect-visioobject $FASObj $poolObj
$y += 1.5

Foreach ($volume in $allVols) {
If ($volume.StoragePool -eq $pool) {
$x += 2.50
$volInfo = "Volume Name: " + $volume.Name # + "`r`n" + "Total Size (GB): " + "{0:n2}" -f + ($volume.SizeTotal / 1gb) + "`r`n" + "Size Used: " + "{0:n2}" -f ($volume.SizeUsed / 1gb)
$FlexVOLObj = add-visioobject $FlexVOLObj $volInfo
connect-visioobject $poolObj $FlexVOLObj

$t_x = $x
$t_y = $y

Foreach ($lun in $allLUNs) {
If (($lun.StoragePool -eq $pool) -and ($lun.Name -match $volume.Name)) {

$x += .5
$y += .25
$lunInfo = "LUN Name: " + $lun.Name + "`r`n" + "Total Size (GB): " + "{0:n2}" -f ($lun.TotalCapacity / 1024 / 1024)
$LUNObj = add-visioobject $LUNObj $lunInfo
connect-visioobject $FlexVOLObj $LUNObj
}
}
$x = $t_x
$y = $t_y
}
}
$x = 3.50
$y += 2.50
}

# Resize to fit page
$pagObj.ResizeToFitContents()

You can find the original script here: http://communities.netapp.com/docs/DOC-6411

--Hector

Case of the Unexplained: Lost Client Connections

2010-12-07T02:30:00Z

At TechEd Berlin I had the opportunity to present a VMM performance troubleshooting case study in Iftekhar Hussain's VIR311 session. You can hear the presentation http://bit.ly/aQ11bP starting at 50:10.

Symptoms Reported

On our production server, we observed that VMM server became very slow. On one of the clients the connection to the server was lost, and the attempt to create a new connection was also not successful. On inspecting the VMM server, the VMM service was taking around 2 GB of memory and was consuming ~25% CPU - which is much higher the usual range of ~500 MB. I collected the memory dump of the VMM service at that time and the logs of the last few minutes at the server. I restarted the service to make it usable. After the service restart, the memory foot-print was around 250 MB and the server was behaving normally again.

A few days back also, I had observed that a few calls like Get-VM (-all or –VMHost), Get-VMHostGroup (-all) were taking times in excess of 1 minute. After some time, the client lost the connection and after initiating the new connection – the things were back to normal. I could not collect logs or any other data at that point.

Troubleshooting Steps (raw log)

I RDP’d into the machine about 2 weeks ago and left the session open with Process Explorer and Process Monitor ready to run once the issue repro’d. Two days ago (7/13/2010), I finally noticed the high CPU utilization and sluggish performance. I first looked at TaskMgr

The system is still usable… but sluggish. This is what I see in Task Manager:

VMMService process is not very interesting… lots of this threads consuming 1-3% CPU:

However, in the System process, the top CPU hog is VMBus:

I tried to get as much info as possible using xperf and TTTracer. I ended up killing the VMMService process accidently after detaching the tracer. This did result in a drop in CPU

CPU utilization during repro:

CPU Utilization right after repro (VMM service restarting):

CPU Utilization after repro (system at “steady state”):

At this point I thought I had lost the repro… until yesterday (7/15/2010) when the symptoms returned.

Typically this issue takes 5-10 days to reproduce. Since this is a production system, the Visual Studio team who owns this machine simply reboot the guest OS. This time however, I was the one that simply restarted the VMMService. Now I got the repro in less than 24 hours.

So with the repro in hand I went off an collected Performance Monitor information. After playing around with loading and unloading counters for disk, network, processes… I narrowed down the list to a few heavy hitters.

- I/O Other Operations/Sec (Process: vmmservice.exe)

- I/O Other Bytes/sec (Process: vmmservice.exe)

“I/O Other” guided me down the path of looking at the network load on the box:

I kicked off Process Monitor and captured a few seconds of activity on the box. After reviewing the information I extracted the following summary for 60 seconds of activity:

All operations:

- Time range: 3:35 – 3:35 (1 min)

- All events: 26687

All operations (EXCLUDING vmmservice.exe):

- Time range: 3:35 – 3:35 (1 min)

- All events: 13179

- Network (TCP/UDP): 1778

ONLY operations for VMMservice.exe:

- Time range: 3:35 – 3:35 (1 min)

- All events: 13508

- Network (TCP/UDP): 10278 (171 TCP Send/Receive)

Looking at Resource Monitor, VMMService.exe has about ~160 TCP connections:

At this point I started searching online for similar symptoms and landed on http://support.microsoft.com/kb/981836. Network connectivity for a Windows Server 2003-based Hyper-Vvirtual machine is lost temporarily in Windows Server 2008 R2

However, this article specifically points to W2K3 guests. So I looked up the WINSE bug and it specifically states that this issue should not occur with W2K8 because the issue is specific to NDIS5.

Information reported to Hyper-V team on 07/22/2010:

· The following changes were made to the environment before 7/22/2010

o KB9881836 has been applied on the parent partition

o The updated integration services have been applied to the SCVMM VM. They are now: 6.1.7600.20683

o The reg key has been set on the parent partition & VM

§ “HKLM\System\CurrentControlSet\Control\Session Manager\Memory Management”, create a DWORD “DisablePagingExecutive” with value 1

§ The kernel debug switch has been set on the parent partition & VM. (bcdedit /debug on)

Logged into machine and opened TaskMgr.

Notice how “CPU usage” graph on the left is not properly reporting the % used.

So I jump to Processes tab and notice that VMMSERVICE is consuming 30-50% of the CPU:

So I fire up Process Explorer to confirm this and notice that VMMSERVICE is one of the CPU hogs along with DPCs (device drivers use DPCs to complete I/O requests):

Considering that last time I observed high CPU was on VMBus.sys, I decided to collect xperf info

C:\>XPERF.EXE –on DIAGEASY

Focusing on DPC info only… I loaded CPU usage summary and noticed that NETVSC60.sys is consuming the most CPU.

Quick, search online tells me that each DPC should not be taking more than 100ms. I am not sure if this is a fact but if you look at the top entries with >100 ms… they are all network related

Also collected the following trace

C:\> xperf -on base -stackwalk profile

After walking the stack trace for 30 minutes

netio is consuming spinning on some spinlocks

Turns out that this is due to WFP filters used by ForeFront

· WFP filters causing cache invalidation
· Forefront uses WFP filters for its network monitoring driver
· Forefront team provided notified and they provided workaround
· ForeFront is still Investigation this issue for a proper root cause and fix

Lesson Learned

· Issue required ~4 minutes of traces to help isolate the culprit
· However, the issue went unsolved for over 6 months
· Supplement your system knowledge with Windows tools
· Capture as much information in case you need to call PSS/CSS
· Waiting for a performance issue to repro is risky
· Minimize exposure by adding performance tools to your toolbox
· Understand how to read a stack trace

How to list VMs by the virtual network they are connected to

2010-06-26T06:28:05Z

This is from an internal question I replied to. This is a simply script given as a starting point that can be extended.

----------------------------------------------------------------------------------------------------------------------------------

I have a host that has 4 physical network cards, 3 of which are mapped to virtual switches that are External.

I would like to, programmatically,

(1) get these 3 virtual network adapters for the given host

PS C:\ > $virtnets = Get-VirtualNetwork -VMHost myhost

PS C:\ > $virtnets.Count

PS C:\ > $virtnets | ft name

Name

----

BLAH

3BLAH3

PRODUCTION

Broadcom - Virtual Network

Test1

(2) figure out how many VMs are connected to each of them

$vmhost = Get-VMHost | where {$_.Name -match "myhost"}

$vmhost.VMs.Count

PS C:\> $count = 1

PS C:\> foreach($vm in $vmhost.VMs){foreach($vnic in $vm.VirtualNetworkAdapters) {if($virtnets -eq $vnic.virtualnetwork){Write-Host ("VM: " + $vm.Name + ", NIC: " + $vnic.VirtualNetwork + ", Count: " + $Count++)}}}

VM: 32w2k8r2, NIC: Test1, Count: 1

VM: hectorl3, NIC: Test1, Count: 2

VM: VI3Servers, NIC: Test1, Count: 3

VM: vi3servers01, NIC: Test1, Count: 4

VM: vi3vc2501, NIC: Broadcom - Virtual Network, Count: 5

VM: VirtuozzoVM01, NIC: Test1, Count: 6

VM: VirtuozzoVM02, NIC: Test1, Count: 7

VM: VM_fromblank, NIC: Test1, Count: 8

VM: VM1_DUPMAC, NIC: Broadcom - Virtual Network, Count: 9

VM: VM1_DUPMAC, NIC: Broadcom - Virtual Network, Count: 10

VM: VMM2008R2RTM, NIC: Test1, Count: 11

VM: VMM2008R2TEST, NIC: Test1, Count: 12

VM: vmm2008test, NIC: Test1, Count: 13

VM: VMNIC, NIC: 3BLAH3, Count: 14

VM: VMNIC, NIC: BLAH, Count: 15

VM: VMNIC, NIC: PRODUCTION, Count: 16

PS C:\>

Domain Join on a Win2k3 VMWare VM from template does not work

2010-06-26T06:05:00Z

The product team is still investigating this issue to come up with a viable solution. Its worth noting that this is a change in VMware's API behavior and not a directly VMM bug.

While helping CSS root cause this issue, I found this forum post:

http://communities.vmware.com/message/1505758 :

“Our own VMware Support call
We used customized sysprep files in Lab Manager v3 to do the domain join and it worked great. A VMware Tech Support Analyst confirmed that this option is no longer supported in Lab Manager v4 and that the customization script should be used. He said that the sysprep files should remain clean”

So I decided to look into VMM code to understand the interfaces VMM uses to generate the VMware answer file :

In the customization subtask use deploying a VM, the there is a method that calls calls VMware API “CustomizationSysprepText” to create the answer. This all seems ok, so I looked at VMware’s VISDK 2.5 API reference doc:

Data object – CustomizationSpec
1. Has a property CustomizationIdentitySettings
Data Object - CustomizationIdentitySettings
1. Is extended by CustomizationLinuxPrep, CustomizationSysprep, CustomizationSysprepText

Looking back at our code

First we create CustomizationSysprepText sysprepText
Then CustomizationGlobalIPSettings globalIpsettings, and specify CustomizationAdapterMapping
Finally call CustomizationSpec, with the following input

1. CustomizationSpec.identity = sysprepText;

2. CustomizationSpec.globalIPSettings = globalIpsettings;

3. CustomizationSpec.nicSettingMap = nicSettings;

Since identity can take CustomizationSysprep and CustomizationSysprepText, I figured this something the vCenter UI mimics.

In vCenter, the Windows Guest Customization manager allow you to create a New Customization Spec in two modes:

1. vCenter managed properties in UI (no sysprep file)

2. Custom sysprep answer file

Even the left nav menu is different for each option.

Looking deeper into CustomizationSysprep, it allows you to explicitly specify CustomizationIdentification, which contains:

1. domainAdmin

2. domainAdminPassword

3. joinDomain

4. joinWorkgroup

P2V fails, if App-V is installed on source machine

2010-06-26T05:59:20Z

VMM P2V agent does not like the App-V virtual drive. I was able to repro the following issue using App-V 4.6 package.

Analysis

P2V agent installs correctly on source machine… but when it enumerates local volumes, an exception gets raised in the trace logs:

00000323 77.26648712 [11200] 2BC0.2B90::06/19-18:30:24.971#17:hwscout.cpp(130): Disk Properties

……

00000343 77.50833893 [11200] 2BC0.2B90::06/19-18:30:25.213#17:systeminfo.cpp(2813): 0: disk 0, [0xc35100000 0x2540a00000) bytes

00000344 77.50843048 [11200] 2BC0.2B90::06/19-18:30:25.213#17:systeminfo.cpp(2851): partition 0, [0x100000 0x6500000) bytes

00000345 77.50848389 [11200] 2BC0.2B90::06/19-18:30:25.214#17:systeminfo.cpp(2851): partition 1, [0x6500000 0xc35100000) bytes

00000346 77.50856018 [11200] 2BC0.2B90::06/19-18:30:25.214#17:systeminfo.cpp(2851): partition 2, [0xc35100000 0x2540a00000) bytes

00000347 77.50861359 [11200] 2BC0.2B90::06/19-18:30:25.214#17:systeminfo.cpp(2504): ==>GetVolumeLetterByVolumeGuid

00000348 77.50897217 [11200] 2BC0.2B90::06/19-18:30:25.214#17:systeminfo.cpp(2504): <--GetVolumeLetterByVolumeGuid

00000349 77.50902557 [11200] 2BC0.2B90::06/19-18:30:25.214#17:systeminfo.cpp(2563): QueryDosDevice 'D:' returned '\Device\HarddiskVolume3'

00000350 77.50916290 [11200] 2BC0.2B90::06/19-18:30:25.214#17:systeminfo.cpp(2748): GetLDrivePartition: Volume \\?\Volume{1a0ecb70-74cd-11df-8ccd-00123f4cfc49}\

00000351 77.50929260 [11200] 2BC0.2B90::06/19-18:30:25.214#00:exceptions.cpp(77)[011BE388]: CarmineException::CarmineException: CarmineError: 1051488, hr: 0x80070005

00000352 77.50942993 [11200] 2BC0.2B90::06/19-18:30:25.214#17:systeminfo.cpp(2882): <--GetLDrivePartitions …

This leads to the 2910 error.

The agent install will try to clean up the installation… but the code will for some reason enumerate the volumes again and fail again…

00000456 130.53585815 [1932] 078C.1F8C::06/19-18:31:18.239#08:agentca.cpp(338): ==>OfflineP2VCleanup

00000457 130.54528809 [1932] 078C.1F8C::06/19-18:31:18.248#17:systeminfo.cpp(3728): ==>CSystemInfoOnline::FindBootVolume

00000458 130.56547546 [1932] 078C.1F8C::06/19-18:31:18.268#17:systeminfo.cpp(2962): GetBootDevice : volume [\Device\HarddiskVolume1]

00000459 130.56553650 [1932] 078C.1F8C::06/19-18:31:18.268#17:systeminfo.cpp(3736): Get letter and guid for boot device [\Device\HarddiskVolume1]

00000460 130.56561279 [1932] 078C.1F8C::06/19-18:31:18.268#17:systeminfo.cpp(2563): QueryDosDevice 'A:' returned '\Device\Floppy0'

00000461 130.56568909 [1932] 078C.1F8C::06/19-18:31:18.269#17:systeminfo.cpp(2563): QueryDosDevice 'C:' returned '\Device\HarddiskVolume2'

00000462 130.56575012 [1932] 078C.1F8C::06/19-18:31:18.269#17:systeminfo.cpp(2563): QueryDosDevice 'D:' returned '\Device\HarddiskVolume3'

00000463 130.56581116 [1932] 078C.1F8C::06/19-18:31:18.269#17:systeminfo.cpp(2563): QueryDosDevice 'E:' returned '\Device\CdRom0'

00000464 130.56588745 [1932] 078C.1F8C::06/19-18:31:18.269#17:systeminfo.cpp(2563): QueryDosDevice 'Q:' returned '\Device\SftVol'

00000465 130.56594849 [1932] 078C.1F8C::06/19-18:31:18.269#00:exceptions.cpp(77)[00CEEED4]: CarmineException::CarmineException: CarmineError: 1051488, hr: 0x80070490

…

00000486 130.77815247 [1932] 078C.1F8C::06/19-18:31:18.481#17:systeminfo.cpp(2813): 0: disk 0, [0xc35100000 0x2540a00000) bytes

00000487 130.77828979 [1932] 078C.1F8C::06/19-18:31:18.481#17:systeminfo.cpp(2851): partition 0, [0x100000 0x6500000) bytes

00000488 130.77835083 [1932] 078C.1F8C::06/19-18:31:18.481#17:systeminfo.cpp(2851): partition 1, [0x6500000 0xc35100000) bytes

00000489 130.77841187 [1932] 078C.1F8C::06/19-18:31:18.481#17:systeminfo.cpp(2851): partition 2, [0xc35100000 0x2540a00000) bytes

00000490 130.77847290 [1932] 078C.1F8C::06/19-18:31:18.481#17:systeminfo.cpp(2504): ==>GetVolumeLetterByVolumeGuid

00000491 130.77894592 [1932] 078C.1F8C::06/19-18:31:18.482#17:systeminfo.cpp(2504): <--GetVolumeLetterByVolumeGuid

00000492 130.77900696 [1932] 078C.1F8C::06/19-18:31:18.482#17:systeminfo.cpp(2563): QueryDosDevice 'D:' returned '\Device\HarddiskVolume3'

00000493 130.77917480 [1932] 078C.1F8C::06/19-18:31:18.482#17:systeminfo.cpp(2748): GetLDrivePartition: Volume \\?\Volume{1a0ecb70-74cd-11df-8ccd-00123f4cfc49}\

00000494 130.77934265 [1932] 078C.1F8C::06/19-18:31:18.482#00:exceptions.cpp(77)[00CEEB44]: CarmineException::CarmineException: CarmineError: 1051488, hr: 0x80070005

00000495 130.77944946 [1932] 078C.1F8C::06/19-18:31:18.482#17:systeminfo.cpp(2882): <--GetLDrivePartitions

00000496 130.77969360 [1932] 078C.1F8C::06/19-18:31:18.482#17:systeminfo.cpp(2995)[00CEF660]: <--CSystemInfoOnline::GetFilteredDiskProperties

00000497 130.77976990 [1932] 078C.1F8C::06/19-18:31:18.483#17:systeminfo.cpp(3728): <--CSystemInfoOnline::FindBootVolume

00000498 130.78016663 [1932] 078C.1F8C::06/19-18:31:18.483#08:agentca.cpp(357): Restore error, fail uninstall

If you try to uninstall the agent manually… you get the 408 error.

Workaround

Since 1) the issue is related to enumerating volumes, 2) App-V has the virtual volume (typically Q: ) 3) and uninstalling App-V fixed the issue…. there is a good chance this is the culprit. On my machine I installed App-V 4.6 to get the repro and test the following workaround.

I went into the Services list and DISABLED

1. Application Virtualization Client

2. Application Virtualization Service Agent

I then REBOOTED the machine. After reboot the Q: drive is no longer visible. Now I can successfully uninstall the agent and kick off P2V successfully.

Mastering Virtual Machine Manager 2008 R2

2009-12-11T05:10:00Z

Mastering Virtual Machine Manager 2008 R2 by Michael Michael and Hector Linares is now available on Amazon.com. Order your copy today.

http://www.amazon.com/Mastering-Virtual-Machine-Manager-2008/dp/0470463325/ref=sr_1_1?ie=UTF8&s=books&qid=1260507554&sr=8-1-catcorr#noop

One-of-a-kind guide from Microsoft insiders on Virtual Machine Manager 2008 R2!

What better way to learn VMM 2008 R2 than from the high-powered Microsoft program managers themselves? This stellar author team takes you under the hood of VMM 2008 R2, providing intermediate and advanced coverage of all features.

Walks you through Microsoft's new System Center Virtual Machine Manager 2008, a unified system for managing all virtual and physical assets; VMM 2008 not only supports Windows Server 2008 Hyper-V, but also VMware ESXas well!
Features a winning author team behind the new VMM
Describes all the new and enhanced features of VMM 2008 R2 and devotes ample time to how it also supports top competitors VMware ES
Uses a hands-on approach, giving you plenty of practical examples to clarify concepts

Open this in-depth guide and discover techniques and processes you can put to immediate use.

From the Back Cover

The Technical Foundation You Need—from Two Microsoft Insiders!

This book shows you how to use Microsoft's System Center Virtual Machine Manager to effectively manage both the virtual and physical assets of your enterprise. Learn from two virtualization insiders at Microsoft how to install, configure, deploy, monitor, and troubleshoot VMM; handle backup and recovery; and much more. Going well beyond the basics, this book provides the technical know-how, best practices, and actual code you'll need to tailor a VMM deployment and keep things running smoothly.

Coverage includes:

Understanding the big picture, including Hyper-VTM 2.0 and Operations Manager 2007
Thoroughly examining all features, capabilities, and architecture you'll need for a successful deployment
Interacting with VMware's vCenterTM for management of ESXTM hosts
Deploying virtual machines to Hyper-V hosts
Building rich automation through Windows PowerShellTM and extending the capability of your VMM
Using the Performance and Resource Optimization (PRO) infrastructure to create new PRO packs and dynamic IT
Backing up and planning for recovery of the VMM server; also backup and recovery of Hyper-V hosts using Hyper-V VSS Writer

Master Microsoft's System Center Virtual Machine Manager 2008 R2

Learn Essential Techniques and Best Practices from Microsoft Experts

Implement Advanced, Effective Virtualization Solutions

Manage Hyper-V and Virtual Server as well as VMware's ESX Infrastructure

Reinforce Your Skills with Real-World Examples

Comparison: SCVMM P2V and Sysinternals Disk2VHD

2009-10-27T00:40:00Z

Sysinterals recently released a free P2V. The comparison table below is for your reference to help highlight the difference in features and functionality between the two. The table is not meant to pin one tool against the other (i.e. who is better). The free tool is not meant to compete against VMM P2V. The free tool does however address the following gaps:

1. Support technician that needs to capture a VM from a physical machine for troubleshooting purposes (VMM is not deployed)

2. Customers that do not want deploy VMM just for P2V or pay for another product

3. End-user converting their own desktop to VMs without requiring a server admin

*Table updated to reflect features in v1.62 of Disk2VHD

Features	VMM P2V	Sysinternals Disk2vhd
Disable services offline after conversion	Yes	No
Disable drivers offline after conversion	Yes	No
Offline P2V support (uses WINPE)	Yes	No
Operating system re-configuration (install IDE driver, fix up HAL, fixup boot)	Yes	Yes
Automation through scripting	Yes (POSH)	Yes(Command shell)
Job progress and job audit	Yes	No (can code something in POSH for job progress)
Supports remote execution	Yes	No
Supports no user interaction on source	Yes	No
Enhanced placement technology to know which host to place the VM workload on	Yes	No
Carry over network configuration from source NICs (static IP)	Yes	No
Skip copy of empty space on source to improve transfer time	Yes	Yes
Uses VSS for online P2V	Yes	Yes
Supports BITS for transferring files (both online and offline) - secure, re-start	Yes	No
Support streaming of data between the source and destination	Yes	No
Support disk size expansion	Yes	No
Automatically install Integration Services	Yes	No
Support W2KSP4	Yes	No
Support client and server OS - XP, W2K3, Vista, W2K8, Win7, W2K8 R2	Yes	Yes
Automatic creation of VM as part of conversion	Yes	No
Free	No	Yes
Requires Management Tool (additional infrastructure)	No	Yes
End-user standalone interface	No	Yes

Recovering VM that has snapshots but no Hyper-V Export file (last resort)

2009-08-24T21:36:00Z

If you need to quickly recover a VM that has snapshots but no export file, you can use the commands in SCVMM R2 that enable rapid provisioning. In this case you are creating a new VM from the latest AVHD.

NOTE: This approach does not recreate the snapshot tree. This is simply a way of recovering your VM if you need immediate access to it.

To get the latest AVHD:

PS C:\>$avhds = get-childitem \\vmhost01\e$\SnapshotRecover01 -Filter "*.avhd" -Recurse | sort-object @{Expression={$_.LastWriteTime-$_.CreationTime}; Ascending=$true}

The first AVHD in this list should be the latest one. From the object you want to tease apart the location of the AVHD and the file name

PS C:\> $avhds[0].directoryname, $avhds[0].name

\\ vmhost01\e$\SnapshotRecover01

SnapshotRecover01_disk_1_37BD309A-C4B3-4196-A698-485B7BADF8AE.avhd

PS C:\>

Now that you have the Path to the AVHD and the AVHD file name, use SCVMM to create a new VM from the existing disk

(1) PS C:\> $guid = [guid]::NewGuid()

(2) PS C:\> $vmhost = Get-VMHost | where {$_.Name -match "vmhost01"}

(3) PS C:\> $vhd = @(Get-VirtualHardDisk | where {$_.Name -Match "large"})

(4) PS C:\> New-VirtualDiskDrive -VirtualHardDisk $vhd[0] -IDE -BUS 0 –LUN 0 -Path "E:\SnapshotRecover01" -Filename "SnapshotRecover01_disk_1_37BD309A-C4B3-4196-A698-485B7BADF8AE.avhd" -JobGroup $guid

(5) PS C:\> New-vm -VMHost $vmhost -Name "SnapshotRecover01_01" -Path "E:\snapshotrecover01" -JobGroup $guid -UseLocalVirtualHardDisks -Description "" -Owner "contoso\user01" -CPUCount 1 -MemoryMB 512 -VMMServer localhost –SkipInstallVirtualizationGuestServices

Step by step

1. Create a guid that will be used for the jobgroup

2. Get the vm host object where you want to create the vm

3. Get ANY vhd from the library. Doesn’t matter which one. The next cmdlet simply needs a placeholder. Use the blank VHDs that ship with VMM for example. This VHD does NOT need to be the snapshot aVHD!

4. Pass the path to the AVHD and the AVHD file name (retrieved from the host earlier) to New-VirtualDiskDrive. Use the VHD placeholder here. The absolute paths you specify here are from the VM host that contains the AVHD (nothing to do with Library).

5. Call the New-VM cmdlet with –UseLocalVirtualHardDisks parameter. This instructs New-VM to skip the BITS deployment and look for the disks you specified in New-VirtualDiskDrive. Use –SkipInstallVirtualizationGuestServices to avoid installing IC.

Disclaimer: This method is simply a quick way of getting the VM up and running. This method does not allow you to restore snapshot information or save state. There is always risk of using the wrong AVHD, so keep that in mind when you do this.

SCVMM 2008 R2 RTM is now available

2009-08-24T21:33:00Z

Download the eval of SCVMM 2008 R2 RTM - http://blogs.technet.com/systemcenter/archive/2009/08/24/system-center-virtual-machine-manager-vmm-2008-r2-rtms.aspx

Windows 7 and Windows Server 2008 R2 Volume Activation Deployment Guide Published

2009-07-21T01:08:00Z

http://technet.microsoft.com/en-us/library/dd772269.aspx

Includes KMS client setup keys. Useful when creating SCVMM Guest OS profiles.

--snippet--

Manually Activate a KMS Client

By default, KMS clients automatically attempt to activate themselves at preset intervals. To manually activate KMS clients (for example, disconnected clients) before distributing them to users, use the Control Panel System item, or run slmgr.vbs /ato at an elevated command prompt. The Slmgr.vbs script reports activation success or failure and provides a result code. To perform activation, the KMS client must have access to a KMS host on the organization’s network.

Converting MAK Clients to KMS and KMS Clients to MAK

By default, Windows 7 and Windows Server 2008 R2 operating systems use KMS for activation. To change existing KMS clients to MAK clients, simply install a MAK key. Similarly, to change MAK clients to KMS clients, run:

Copy Code

slmgr.vbs /ipk <KmsSetupKey>

where KmsSetupKey is one of the setup keys shown in Table 4. After installing the KMS setup key, activate the KMS client by running cscript slmgr.vbs /ato.

Table 4 KMS Client Setup Keys

Operating System Edition	Product Key
Windows 7
Windows 7 Professional	FJ82H-XT6CR-J8D7P-XQJJ2-GPDD4
Windows 7 Professional N	MRPKT-YTG23-K7D7T-X2JMM-QY7MG
Windows 7 Enterprise	33PXH-7Y6KF-2VJC9-XBBR8-HVTHH
Windows 7 Enterprise N	YDRBP-3D83W-TY26F-D46B2-XCKRJ
Windows 7 Enterprise E	C29WB-22CC8-VJ326-GHFJW-H9DH4
Windows Server 2008 R2
Windows Server 2008 R2 HPC Edition	FKJQ8-TMCVP-FRMR7-4WR42-3JCD7
Windows Server 2008 R2 Datacenter	74YFP-3QFB3-KQT8W-PMXWJ-7M648
Windows Server 2008 R2 Enterprise	489J6-VHDMP-X63PK-3K798-CPX3Y
Windows Server 2008 R2 for Itanium-Based Systems	GT63C-RJFQ3-4GMB6-BRFB9-CB83V
Windows Server 2008 R2 Standard	YC6KT-GKW9T-YTKYR-T4X34-R7VHC
Windows Web Server 2008 R2	6TPJF-RBVHG-WBW2R-86QPH-6RTM4