Hey Forum Users:

If I plan on moving my entire architecture from VMware/Win2008-2003 to Hyper-V/2012-2008 is it possible to have different chipsets?    In the best of all worlds the answer is yes but think it's no.

One datacenter has AMD servers while the other has Intel.   The goal is to flatten out the network, move to a single domain and use each datacenter as a DR site for the other (and vice versa).

Whether I spend another $20K hinges on the efficacy of cross-processor use.

Thanks, IMSIT

Our Remote Desktop Servers crash on a regular basis. It happens when they are overloaded - high CPU use. There are three of them; the latter two were cloned from the first one. The same behavoir happens in all three.

These machines are Server 2008 64 bit Virtual Machines hosted on a Windows Server 2008 64 bit machine using Hyper-V. They all have 4 processors assigned and more RAM than they need.

The host server has a total of 8 cores available through two quad core 2.8Ghz processors. There are three host servers, all with the same configuration. The crashing happens on the VMs only, never the hosts. The most a host has on it are three VMs, all similarly configured. The only VMs that reboot are the ones with FMP/SQL on them.

There is a blue screen error, which you will see below. The analyzed contents of the crash dump file are below that. I ran it through WinDbg, and posted that info.

The key thing here to understand is this:

RDS1 was stable until FileMaker Pro 11 and the SQL client were installed as part of a maximEyes 11 installation (optimetry software by First Insight).

This crashing has been occurring since the upgrade.

The crashing is obviously related to the FMP and SQL software. Whether it is the software itself, or simply the high CPU use caused by the software I have not been able to figure out. The server can have anywhere from 15 to 20 users logged on, all using the FMP software, some using the SQL.

The Blue Screen message is from the details on the box that appears in Windows after a system reboots.

The Crash Dump is the analysis done in WinDbg.


-----------BLUE SCREEN------------
Problem signature:
  Problem Event Name:      BlueScreen
  OS Version:      6.1.7601.2.1.0.18.10
  Locale ID:      1033

Additional information about the problem:
  BCCode:      ab
  BCP1:      0000000000000020
  BCP2:      0000000000000350
  BCP3:      0000000000000000
  BCP4:      0000000000000001
  OS Version:      6_1_7601
  Service Pack:      1_0
  Product:      18_3

----------CRASH DUMP FILE------------

     Bugcheck Analysis                                    *
*                                                                            *
*******************************************************************************

Use !analyze -v to get detailed debugging information.

BugCheck AB, {1d, 3a0, 0, 1}

Probably pool corruption caused by Tag:  Pool

Followup: MachineOwner
---------

3: kd> !analyze -v
*******************************************************************************
*                                                                            *
*                        Bugcheck Analysis                                    *
*                                                                            *
*******************************************************************************

SESSION_HAS_VALID_POOL_ON_EXIT (ab)
Caused by a session driver not freeing its pool allocations prior to a
session unload.  This indicates a bug in win32k.sys, atmfd.dll,
rdpdd.dll or a video driver.
Arguments:
Arg1: 000000000000001d, session ID
Arg2: 00000000000003a0, number of paged pool bytes that are leaking
Arg3: 0000000000000000, number of nonpaged pool bytes that are leaking
Arg4: 0000000000000001, total number of paged and nonpaged allocations that are leaking.
      nonpaged allocations are in the upper half of this word,
      paged allocations are in the lower half of this word.

Debugging Details:
------------------


CORRUPTING_POOL_TAG:  Pool

DEFAULT_BUCKET_ID:  VISTA_DRIVER_FAULT

BUGCHECK_STR:  0xAB

PROCESS_NAME:  csrss.exe

CURRENT_IRQL:  0

LAST_CONTROL_TRANSFER:  from fffff80001a6e61f to fffff800016cf1c0

STACK_TEXT:  
fffff880`0ddc0ac8 fffff800`01a6e61f : 00000000`000000ab 00000000`0000001d 00000000`000003a0 00000000`00000000 : nt!KeBugCheckEx
fffff880`0ddc0ad0 fffff800`0190b917 : fffff880`04f0bb40 fffff880`04f0b000 fffff880`04f0b000 fffffa80`27f8cb00 : nt!MiCheckSessionPoolAllocations+0x13f
fffff880`0ddc0b10 fffff800`01a08fa5 : fffff880`0ddc0ba8 fffffa80`27f8cb00 ffffffff`ffffffd5 fffff880`04f0b000 : nt!MiDereferenceSessionFinal+0x137
fffff880`0ddc0bb0 fffff800`016a039c : fffff800`01854940 00000000`00000001 00000000`00000000 fffffa80`28066060 : nt! ?? ::NNGAKEGL::`string'+0x25025
fffff880`0ddc0be0 fffff800`019a06da : fffff8a0`1f6da060 00000000`00000000 00000000`00000000 fffffa80`27f8cb00 : nt!MmCleanProcessAddressSpace+0x610
fffff880`0ddc0c30 fffff800`019a0abd : 00000000`00000000 fffff800`01965e01 00000000`00000000 fffffa80`28021b00 : nt!PspExitThread+0x56a
fffff880`0ddc0d30 fffff800`016bfec6 : fffff880`020e1180 00000000`00000080 fffffa80`27f8cb00 00000000`00000200 : nt!PspTerminateThreadByPointer+0x4d
fffff880`0ddc0d80 00000000`00000000 : fffff880`0ddc1000 fffff880`0ddbb000 fffff880`0ddc0710 00000000`00000000 : nt!KxStartSystemThread+0x16


STACK_COMMAND:  kb

FOLLOWUP_IP:
nt!MiCheckSessionPoolAllocations+13f
fffff800`01a6e61f cc              int     3

SYMBOL_STACK_INDEX:  1

SYMBOL_NAME:  nt!MiCheckSessionPoolAllocations+13f

FOLLOWUP_NAME:  MachineOwner

MODULE_NAME: nt

DEBUG_FLR_IMAGE_TIMESTAMP:  4fa390f3

IMAGE_NAME:  memory_corruption

FAILURE_BUCKET_ID:  X64_CORRUPTING_POOLTAG_Pool

BUCKET_ID:  X64_CORRUPTING_POOLTAG_Pool

Followup: MachineOwner

Cathy Burnham, MCSA

Hi,

We've just hit an issue with Dynamic Memory where Hyper-V allocates memory for the guest until the RAM is exhausted on the host. When we look inside the VM (2008 R2 SP1)  there's no process that requires that memory. If we look  There's basically SQL Server which is capped at 10GB and the fulltext process which takes around 6GB but Hyper-V sees the VM as requesting 204GB of RAM and has assigned 238GB to it. Here's a screenshot of the memory allocation in the VM:

Here's what I see on the host:

It looks like a bug in the Integration Services or the dynamic memory allocator.

Has anyone seen this?

Thanks,

Mathieu

Hello,

We run Hyper-V on a couple Server 2008 R2 hosts, and the primary one hosts 11 guests, all Server 2008 R2.  Several of the guests are absolutely critical (namely the one running the Microsoft TMG, which is our firewall / router / VPN server, and the domain controller).  I have only been here since the end of July and have been trying to get things straightened out but am limited in time and resources to set things up the way they need to be.  This VM server is being strangled at the moment in both RAM and disk space, and it has ended up pausing most of the VMs with "paused critical" at least 6 or 7 times, usually late at night or early in the morning, knocking everything offline.  Yes I know we need more disk space, well aware of this and trying to get the new SAN up and running.  Most of the guest VHDs are fixed, some are dynamic (and unfortunately the Exchange server has a terabyte worth of snapshots that I need to merge but cannot afford the massive downtime required to do so at the moment).  I cannot compact any of the VHDs anymore than I already have, cannot be more efficient than I already with conserving disk space, but the data keeps growing (primarily from the Exchange and the MS SQL servers) and growing at a very fast clip.

I'm aware that I'm going to have to deal with Windows pausing the VMs occasionally in the near-term until I get the SAN going for the non database intensive VMs, can't do anything about that right now.  What I want to do is find out if the gateway and domain controller VMs can have *absolute* priority above all the others - interestingly, the gateway usually never pauses, but the domain controller always does so I cannot log into the VPN from home.  I'm working on setting up the DRACs in our Dell servers so I can log in from an entirely independent network but that isn't up yet and I have extremely limited time to work on each of these tasks at hand.  All I want is to ensure the gw and domain controller remain up - the rest can be paused - until I can get the ball rolling elsewhere.  Can I set priority somehow?  Really frustrating that Hyper-V doesn't at least have an option to fire off an alarm email or something, but I have to deal with what I have.

Thanks in advance for any help.

Rhiannon

I was just curious if the configuration I'm attempting is supported or not. I know that with the release of Server 2012, NIC teaming is now supported, but I’m unsure to what level once you add Hyper-V to the mix.

Basically this is what I'm attempting, i have a 1GB NIC, and 10GB NIC in my server. I then created two teams, based on the following:

1GB NIC -> Team1

     Virtual Adapter 1 = VLAN 2

     Virtual Adapter 2 = VLAN 3

     Virtual Adapter 3 = VLAN 4

10GB NIC -> Team 2

     Virtual Adapter 1 = VLAN 5 (dedicated VLAN for Live Migration traffic)

     Virtual Adapter 2 = Trunk (converted to VSwitch)

After setting up this two node cluster, it appears that node 1 works fine, but node 2 does not. Any VM residing on node has no network connectivity.

I've also run into another issue, in that Hyper-V and the VMM agents fail on both servers regularly, after receiving a bunch of the following warnings in the system log:

Log Name:      System
Source:        Microsoft-Windows-MsLbfoSysEvtProvider
Date:          11/6/2012 12:55:13 AM
Event ID:      16946
Task Category: None
Level:         Warning
Keywords:      Classic
User:          N/A
Computer:      Node2.mydomain.com
Description:
Use of Vlan detected on non-default Team Nic Microsoft Network Adapter Multiplexor Driver #7
Event Xml:
<Event xmlns="http://schemas.microsoft.com/win/2004/08/events/event">
  <System>
    <Provider Name="Microsoft-Windows-MsLbfoSysEvtProvider" Guid="{387ed463-8b1b-42c9-9ef0-803fdfd5d94e}" EventSourceName="MsLbfoProvider" />
    <EventID Qualifiers="32768">16946</EventID>
    <Version>0</Version>
    <Level>3</Level>
    <Task>0</Task>
    <Opcode>0</Opcode>
    <Keywords>0x80000000000000</Keywords>
    <TimeCreated SystemTime="2012-11-06T06:55:13.363649100Z" />
    <EventRecordID>20403</EventRecordID>
    <Correlation />
    <Execution ProcessID="4" ThreadID="5152" />
    <Channel>System</Channel>
    <Computer>Node2.mydomain.com</Computer>
    <Security />
  </System>
  <EventData>
    <Data Name="DriverObject">
    </Data>
    <Data Name="Member">Microsoft Network Adapter Multiplexor Driver #7</Data>
  </EventData>
</Event>

Any thoughts or suggestions would be greatly appreciated.

Hi Guys

I have a 3 node Hyper-V CSV cluster using Windows Server 2008 R2. Each node are the same type of servers (HP BL460c G7). I need to add another node to the cluster, a more high spec node (G8). This node will be used to provide RemoteFX capabilities to the virtual desktops that will be hosted on it.

I need to verify what issues may come up by having a CSV cluster with different types of server nodes????

jk

Hi,

I'm trying to upgrade a Hyper-V 2008 R2 installation to Hyper-V 2012:

I've removed all of the snapshots, set the mac addresses to static and exported the existing machines to an external drive.

I've performed a clean install of Hyper-V 2012 and now I'm trying to import the machines back.

I've recreated the virtual switches as they were in the old configuration.

The issue is with importing the VMs back:

If I'm using the exports folder (via Resotre the virtual machine option), the import process hangs at 'Loading configuration'.

If I'm using the existing folder structure for the VM (via Register ...), the import process hangs at 'Performing import operation'.

Any ideas ?

I have successfully gotten three VMs to enable replication; but on my next VM I keep getting the same error message when I try to enable replication.

"Hyper-V failed to enable replication for virtual machine 'IPAM': The operation timed out (0x00002EE2)."

Nothing has changed since I enabled replication for the other VMs, and I am at a loss as to why this VM is having issues.

The only odd thing I have noticed is on the DR node, Hyper-V creates the VM and creates the cluster resource role for the VM but never adds the VM to the resource role.

I have gone through the "Understanding and Troubleshooting guide for Hyper-V Replica" but that has not helped identify the problem.

Any help would be appreciated.

Hyper-V 2012: Perform Live Migration in a Failover Cluster without CSV

In Windows server 2008 R2, when performing Live migration on a VM in a Failover Cluster without CSV, all other VMs in same LUN will also be migrated.

Is it same behavior in Hyper-V 2012? Do I need a dedicated LUN for each VMs in Hyper-V 2012 without CSV?

I am migrating a bunch of physical machines to be hosted on Hyper-V Server 2008 R2. The guest operating systems are a variety of Windows versions. For reasons I won't detail, I need to know which versions of Windows will automatically install a virtual NIC when they are first started as a virtual machine, and which versions of Windows will not have a virtual NIC until I log in to the VM and install Integration Services.

For example, I have already determined that Windows Server 2008 R2 and Windows SBS 2011 automatically install a virtual NIC without requiring an initial login, whereas Windows Server 2003 SP2 does not. What is the exact dividing line or criterion for predicting this behaviour on other versions of Windows? (I need to predict for both desktop and server versions of Windows).

Thanks,
Richard Koett.
 

Hi,

I am very new to Hyper-V so please forgive me if I am asking this in the wrong place.

I installed the Windows 2012 Data Center Edition 6-month trial version, set it up with the Hyper-V role and configured three virtual machines. I decided I wanted to test the Free version of Hyper-V instead so I exported each of those machines to an extra hard drive I installed in my test machine. It only took a few minutes to export each machine.

I then proceeded to install the Hyper-V 2012 Free version in place of Windows 2012 Data Center Edition and have the Hyper-V manager installed and running on a Windows 8 machine I have in my lab. I am able to run the Hyper-V manager and connect to my new server just fine. However, when I attempt to import one of the three machines I had exported previously it starts the process fine, eventually a small window pops up that says "Preparing virtual hard disk destination..." and it seems to be hanging there, even though it only took a few minutes to export this machine it has been stuck on this window for over an hour now.

Does anyone have any idea how to troubleshoot this issue? Or is it normal for this process to take far longer than the initial export?

I appreciate any insight anyone can provide.

Thanks in advance.

We purchased a new Powervault PV3260 with 30 2TB Drives.

I have a Windows 2008 server with two NIC's (Intel PRO/1000 PL and Intel PRO/1000 PM), and one virtual external switch connected to each NIC.

With the 01/14/08 network driver, I get very low throughput on both NIC's (~10KB/sec when downloading from my server when it should be around 11MB/sec).

I've found that disabling "TCP Large Send Offload (IPv4)" on the virtual switch and "Large Send Offload v2 (IPv4)" on the physical adapters solves the problem.

Is this a bug in the driver?

Thank you,

Ricardo Costa

How do I calculate the hardware requirement for a new windows 2008 AD setup for 500 users ?

I know there is not specific hardware requirement stated in MS website.

how do i setup my AD so that it is not overkill and over budget ?

I need to setup a Hyper-V host for 3 VM machine.

2 domain controller and 1 TMG VM.

I'm currently looking at VM resource pools in Hyper-V and I'm wondering how they work. I created a pool using:

New-VMResourcePool "Database Engine Services" Processor,Memory,Ethernet

Now, how can I associate a VM to the pool for metering purposes?

When I use Enable-VMResourceMetering, there's no way to associate a VM to a pool. I checked Set-VM as well but I couldn't find a way using that either.

What I would like to have ultimately is a hierarchy of pools for reporting purposes. i.e.

Servers
Servers\Database Engine Services
Servers\Web Application Services
Servers\Reporting Services
Servers\Reporting Services\SSRS

So I could wrap the stats at different levels, depending on what I want to look at. I see that a hierarchy of pool is possible judging by Set-VMResourcePool.

Any help/guidance on this would be appreciated.

Thanks!

Mathieu

Okay, I'm stumped.  I have tried every combination of formats I can think of, but I cannot seem to find the proper string to remove a network from use by live migration.  Maybe I'm trying to use the wrong cmdlet, but the description of Remove-VmMigrationNetwork seems to be what I am looking for -"The Remove-VMMigrationNetwork cmdlet removes a network from use with migration."  I have tried from a host within a cluster.  I have tried it from a remote host that I used to build the cluster with remote commands.  I have tried wild cards and /24 notation (which results in a syntax error).  I tied specifying a cluster name, a single host name, and all the host names in the cluster.  And on and on.

The primary error I end up with is this (executed from one of the cluster nodes):

Remove-VMMigrationNetwork : Failed to modify service settings.
Cluster migration network tags cannot be modified. Please ensure you are not trying to add or remove a cluster
migration network for '192.168.16.22/32' : One or more arguments are invalid (0x80070057).
A parameter that is not valid was passed to the operation.
At line:1 char:1
+ Remove-VMMigrationNetwork 192.168.16.*
+ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
    + CategoryInfo          : InvalidArgument: (Microsoft.HyperV.PowerShell.VMTask:VMTask) [Remove-VMMigrationNetwork]
   , VirtualizationOperationFailedException
    + FullyQualifiedErrorId : InvalidParameter,Microsoft.HyperV.PowerShell.Commands.RemoveVMMigrationNetworkCommand

What is strange is that I am using a Remove command and the error message asks me to ensure I am not trying to remove.

tim

When we attempt to add English (United States) language support, as per the article (KB2607607), it says the language is not available.

In other words, unless this is a bug with the connection with Microsoft's servers, do we have to get the US English versions of Windows 8 Pro to be able to use these Tools, or have a TechNet Subscription to get the language files which apparently are available, but not for download? But why would we have to get the US version if the OS has language extensions? Licensing issues? Bugs?

Clean install of Windows 8 Pro x64.

<Frustration>Is software design becoming dumber? Developers can't even package the default language with the executable now? Wake-up call: this is dumb, or just another bug. Hopefully a bug.</Frustration>

I have Hyper V 2008 7.6.10, and I have just created 5 virtual machines which I would like to provide 3 different networks to them. WAN, LAN and DMZ. 

The Hyper V original was installed with one NIC "LAN" but now I would like it to show the 3 of these NICs which are physically connected to the HYPER-V already. 

I have tried to browse to the Hyper V Virtual Network management console and click "Add External" then choose the Physical LAN but that would give me error setup installation all the time. 

I haven't done this before and I would appreciate how to do this! 

Thanks 

Mohammed JH

I'm finally stuck.

I've tried every blog post out there. I've tried HVRemote. I've formatted and reinstalled several times.

The only one consistent message is:

RPC server unavailable. Unable to establish communication between 'ServerName' and 'ClientName'

I've allowed Anonymous Remote access in the DCOMCNFG.

Disabled firewalls on both sides

User is administrator on both sides

I've added the server name and ip to the Host file on the client.

I can connect to most services via MMC snap ins.

I can't access the virtual disk manager (Yes it's set to auto and is started)

I've added the user to the Hyper V Administrators (Even tho it's an Administrator)

I've also tried the technet installation guide.

Please help.