Tags:
Hyper-V
When you deploy multiple virtual RDSH Servers on a Hyper-C cluster you want in most cases some control in the placement of these servers. If you have a two-server RDS Farm you don't want those two nodes to be running on the same Hyper-V Node.
Now it is possible to define anti-affinity for clustergroups. You can create an anti-affinity object called NLBCLuster1 and try to prevent running both RDSH server on the same Hyper-V node by running the following commands:
Cluster.exe group "RDSH01" /prop AntiAffinityClassNames="NLBCluster1"
Cluster.exe group "RDSH02" /prop AntiAffinityClassNames="NLBCluster1"
The anti-affinity can be overridden in some cases where a cluster with reduced capacity needs to keep the VM's running.
September 10, 2010
·
Posted by Erik den Burger
·
0 Comments
Tags:
Hyper-V
After doing some tests with dynamic on my Home lab I ran into a problem that suddenly the performance of the parent partition was well below normal. So after some troubleshooting I figured out that my VM's were using too much memory. Normally I would just stop one of the VM's and reduce the amount of memory to that VM. But after installing SP1 this fix didn't give me the desired results. I soon found out that the dynamic memory feature was the cause. There just wasn't enough memory available for all VM's I had running.
Normally Hyper-V will calculate a parent partition reserve but this calculation is done assuming you are not running anything else on your parent partition. Off course this is best practice in any production system but on my home lab I was running other software on my parent partition.
Fortunately Microsoft has a registry key (HKLM\software\microsoft\windows nt\currentversion\virtualization\memoryreserve, a DWORD) that will let you tweak the parent partition memory reserve. Now I can reserve more memory for my parent partition and keep the software running. Downside is that I have less memory available for my VM's. But off course dynamic memory isn't a magic tool that will give you more memory, it just will use the memory more efficiently.
September 9, 2010
·
Posted by Erik den Burger
·
0 Comments
Tags:
Performance · Hyper-V
One of the big improvements in windows 2008R2 regarding the performance of networking in a Hyper-V environment is the implementation of Virtual Machines Queues (VMQ). VMQ uses hardware packet filtering to deliver packet data from the external network to your virtual servers directly and doing so, reduces the overhead of routing packets between the management OS and virtual servers.
To make us of VMQ you should be running Windows 2008R2, the physical NIC should support VMQ (or VMDq as Intel calls it) and your virtual machines should be running Windows 7, Windows Server 2008R2 or Windows 2008/Vista with the windows 2008R2 Integration Services installed. Older versions cannot make us of VMQ.
Most Intel 10Gig NIC do support VMDq. In the gigabit range you should go for an Intel ET, ET2, EF or VT Based card. Other vendors have announced they will support VMQ in the near future, at this moment I'm not aware of any vendor, other than Intel that has support for VMQ.
When VMQ is enabled a dedicated queue is established on the physical NIC for each virtual NIC that has requested a queue. Because queues are allocated on a first-come-first-served basis it would be a bad idea to enable VMQ for all your VM's. So identify your heavy traffic VM's and enable VMQ for those VM's only. Because VMQ works primary to improve receive side performance, providing the VM's that receive the most packets will benefit most from enabling VMQ.
To enable VMQ for a Virtual Machine, you first need to enable VMQ for a physical NIC. After you have enabled it you can enable VMQ on the Hyper-V Server. Finally you need to configure the virtual NIC of your virtual machine to use VMQ.
These steps will be covered in Part 2. In part 3 we will fine tune VMQ a bit more and see some results of VMQ in action. All I can say at this moment that VMQ is well worth having on your Hyper-V implementations.
August 25, 2010
·
Posted by Erik den Burger
·
0 Comments
Tags:
Hyper-V · Windows Infrastructure
This weekend I got this message from Dell:
Not the first time Equallogic has troubles with their firmware....
Dear Customer,
Do Not Install Dell EqualLogic PS Series Firmware Versions 5.0.0 or 5.0.1
According to Dell’s records, you have recently downloaded v5.0.0 or v5.0.1 of the Dell EqualLogicTM firmware. Dell has become aware of potential issues with the installation and operation of these versions of the firmware. Under certain conditions, any or all of the following might occur:
• Volumes might not come online properly immediately after the install.
• Replication might not occur properly.
• VMware V4.1 Zero offload performance might be affected.
If you have not yet installed the V5.0.0 or V5.0.1 firmware updates, Dell recommends that you do not install at this time. Dell is working to resolve these issues in future releases.
If you have either of these firmware updates installed and are experiencing any issues with your
array(s), please contact your local Dell EqualLogicTM technical support team for assistance.
Sincerely,
Dell EqualLogic
July, 2010
August 2, 2010
·
Posted by Erik den Burger
·
0 Comments
Tags:
Hyper-V · Windows Infrastructure
During the World Partner Conference Microsoft announced the public beta of Service Pack 1 for Windows 2008R2 and Windows 7. Actually this release was supposed to be available in a couple of weeks so the announcement surprised me. The final version is supposed to be out somewhere next year but now the beta is out earlier than expected, who knows?
Besides fixing bugs SP1 introduces two new exciting technologies.
Dynamic memory gives Hyper-V admins a pool available memory on a physical host and dynamically distribute the memory to any virtual guest running on that host. An interesting concept that will make Hyper-V even more flexible to use.
RemoteFX will enhance the MS Desktop virtualization. It will deliver a better user experience for users of desktop virtualization solutions (RDS, VDI). It will allow users to use full motion video, 3D-Applications and rich Silverlight content. And because the power of the server is being utilized these experiences will be available to a wide array of devices. Another feature that will be welcomed by a lot is the ability to forward USB ports of the local client to the virtual machine just like you are using client based printers now.
As you see, a lot of new technologies that will provide new opportunities. So go out and download SP1 here.
July 12, 2010
·
Posted by Erik den Burger
·
0 Comments
Tags:
Windows Infrastructure · Hyper-V
Last week Microsoft issued a press release that had a lot of new statements on the MS front. In the press release a lot of changes were announced, especially for the VDI licensing. But there were also some technical statements.
One of these was not a surprise as this feature was supposed to have made it to the RTM of Windows 2008R2 but just didn't do so. The upcoming SP1 for Windows 2008R2 will bring memory over commitment to Hyper-V. Off course Microsoft has given this feature another name but Dynamic Memory is basically the same as Memory over commitment. So it is possible to reserve more memory for your guest than you physically have in your hosts.
Besides the announcement of Dynamic memory another new technology was announced. The RemoteFX also was supposed to be in the RTM but also was delayed until SP1. This will enable users of virtual desktops (either VDI or RDS) to use the graphic power of the clients to create a rich 3D and multimedia experience.
One other announcement was a little surprise for me. XP mode will no longer require CPU Assisted virtualization. This will make it more accessible to many more older PC's and therefore this feature will be more used in small and medium businesses.
The last technical announcement was HDX technology that will enhance and extend the RemoteFX platform in Citrix Xendesktop. This will give citrix again an edge (at a price) over standard 2008R2 RDS.
March 25, 2010
·
Posted by Erik den Burger
·
11 Comments
Tags:
Hyper-V
During his keynote at the recent VMWare Partner Exchange, Paul Jackson made some interesting comments.
The chief marketing officer of VMWare told the attendees that: "Microsoft is painting a beautiful picture about cloud computing, but according to its own internal documents is not using its own Hyper-V virtualization platform because it cannot easily pool CPU, memory, and networking resources."
Now he has a point, the current version of Hyper-V doesn't shine in the possibilities to pool CPU, memory and network resources and yes, Microsoft doesn't use Hyper-V for its azure platform. But what supervisor does it use in its Azure platform and why does Azure use it?
The foundation of the Azure platform is a newly developed hypervisor that was built with three basic principles in mind. The hypervisor should have a small footprint, be efficient and provide tight integration.
- The small footprint means that any features that are not needed in the cloud scenario are removed. That way you remove the need for fixing unnecessary code. Having a small footprint reduces the amount of patching and rebooting of your servers and making your hypervisor less vulnerable for potential attacks.
- The efficiency is obvious. Any percentage gain will be significant once multiplied by tens of thousands machines. The result is that more CPU cycles are available for the hosted applications.
- To achieve the required level of performance and scalability it is necessary to have a tight integration between the Azure Hypervisor and the Azure Kernel.
As you see, a lot of the requirements for the Azure hypervisor are specific for the needs of azure only, and therefore there is a need for a separate supervisor. However a lot of the innovations in the Azure hypervisor will be incorporated in the next releases of Hyper-V
February 13, 2010
·
Posted by Erik den Burger
·
0 Comments
Tags:
Hyper-V · Security
For the first time since its official launch 18 months ago, Microsoft has issued a critical security patch for Hyper-V. The fix is available for all windows 2008 and 2008R2 installations.
The update resolves a vulnerability in Hyper-V that could allow a denial of services attack if a malformed sequence of instructions is run on a virtual guest running on the Hyper-V Server. However, the attacker must have valid logon credentials and be able to logon locally into the virtual machine.
The impact of the vulnerability is limited, in most scenarios, administrator access to the virtual hosts is limited to trusted administrators. In proper designed SBC cases the users have limited rights and cannot launch the malformed sequence of instructions. In some VDI scenarios it is possible that users have local administrator rights within their virtual machine and thus are capable to attack the underlying Hyper-V Server.
For more information about the vulnerability you can visit KB977894: VERY Important Hyper-V Security Update
February 11, 2010
·
Posted by Erik den Burger
·
0 Comments
Tags:
Hyper-V · Windows Infrastructure
Ever since the release of Windows 2008R2 we have been waiting for this new release of DPM. After testing severals Beta's I think it is great news that DPM 2010 now has the release candidate status.
now you get better performance, possibilities to backup to the cloud and expecially CSV support! So we can backup out Hyper-V Clusters in an efficient manner.
February 9, 2010
·
Posted by Erik den Burger
·
11 Comments
Tags:
Active Directory · Hyper-V · Windows Infrastructure
With Hyper-V becoming widely used the question in a lot of designs is weather to virtualize your AD infrastructure as well. The answer isn't always as easy and clear as it looks.
First let's see what Microsoft has to say about virtualizing your ADDS Servers. In article KB888794 you can read that it is indeed possible to run your ADDS server in a virtual environment when you follow the guidelines about pausing, snapshotting, backupping and performance, you should have no problems at all.
Now let's see what would happen if you're using a Hyper-V cluster and because you want to take advantage of your High Availabily features of your Hyper-V Cluster, all of your ADDS Servers are running virtualized. As long as your cluster is up and running everything is looking good. But what is for whatever reason your cluster needs to be restarted after a shutdown? You have a problem. The cluster service isn't starting, telling you it cannot contact a domain controller. So you have a cluster that isn't starting because there is no ADDS server available and you have a ADDS Server that isn't starting because the cluster it is running on cannot be started. This is a major problem. To fix this problem take a while, you have to build a standalone Hyper-V machine and manually copy the VHD of the Cluster-LUN. Now you can start the ADDS server and start your cluster.
So when using a Hyper-V cluster it is really important to keep at least one of your ADDS server running on a physical machine or if you want on a standalone Hyper-V server.
February 6, 2010
·
Posted by Erik den Burger
·
9 Comments
Next >>