Déjà vu Bug: The network connection of a running Hyper-V virtual machine may be lost under heavy outgoing network traffic on a computer that is running Windows Server 2008 R2 SP1

Anyone who’s been doing virtualization with Hyper-V on Windows 2008 R2 has a good change of having seen the issue described in http://support.microsoft.com/kb/974909/en-us

You install the Hyper-V role on a computer that is running Windows Server 2008 R2.

  • You run a virtual machine on the computer.
  • You use a network adapter on the virtual machine to access a network.
  • You establish many concurrent network connections, or there is heavy outgoing network traffic.

In this scenario, the network connection on the virtual machine may be lost. Additionally, the network adapter is disabled.
Note You have to restart the virtual machine to recover from this issue.

We’ve seen this one on VM’s that have indeed a lot of outgoing traffic.  In our environment the situation looks like this:

  • You can access the VM with Hyper-V Manager or SCVMM but not via RDP as all Network connectivity is lost.  The status the  guest NIS is always “Enabled” but there is no traffic/connectivity
  • You can try to disable the NIC but this tales a  very long time and when you try to enable it again this never succeeds. Disconnecting the NIC form the virtual network and connecting it again doesn’t help either.
  • You need to shut down the host but this takes an extremely long time, so long you really can’t afford to wait if it ever succeeds. It seems to hang at shutting down with a “non whirling whirly”.  So finally you’ll power off the VM and start it up again. Apart from entries related to having not connectivity the event logs are “clean” and there is no indication as to what happened.

Well this exact same issue is back with Windows 2008 R2 SP1. That’s the bad news. The good news is there is a hotfix for it already so you can fix it. You can read up on this issue in Knowledge Base article 2263829  and request the hotfix here. Instructions to get the hotfix are in there as well as a reference to the previous fixes for Windows 2008 R2 RTM.

Consider the following scenario:

  • You install the Hyper-V role on a computer that is running Windows Server 2008 R2 Service Pack 1 (SP1).
  • You run a virtual machine on the computer.
  • You use a network adapter on the virtual machine to access a network.
  • You establish many concurrent network connections. Or, there is heavy outgoing network traffic.

In this scenario, the network connection on the virtual machine may be lost. Additionally, the network adapter may be disabled.
Notes

  • You must restart the virtual machine to recover from this issue.
  • This issue can also occur on versions of Windows Server 2008 R2 that do not have SP1 installed. To resolve the issue, apply the hotfix that is described in one of the following Microsoft Knowledge Base articles:

    974909 (http://support.microsoft.com/kb/974909/ ) The network connection of a running Hyper-V virtual machine is lost under heavy outgoing network traffic on a Windows Server 2008 R2-based computer
    2264080 (http://support.microsoft.com/kb/2264080/ ) An update rollup package for the Hyper-V role in Windows Server 2008 R2: August 24, 2010

Oh yeah, people often seem confused  as to where to install the hotfix. Does it go on the Hyper-V hosts or and/or on the guest?  It’s a hyper visor bug in Hyper-V so it goes on the hosts. Have a nice weekend.

Upgrading a Hyper-V R2 Cluster to Windows 2008 R2 SP1

For all you people waiting to roll out Windows 2008 R2 SP1 to your Hyper-V cluster here’s a quick screenshot rich run through of the process. Some people claim you need to shut down the cluster services and shut down the guests but this is not the case.  You can do a rolling upgrade and your guests can stay on line on the other nodes, just use live migration to get them there. Now I do recommend to upgrade all the nodes tot SP1 as soon as possible and not staying a mixed Windows 2008 R2 / Windows 2008 R2 SP1 situation in your cluster. But this mixed situation makes upgrades of the nodes in the cluster possible without any down time for the guests (if you have live migration), which is the aim of having a high availability cluster.

Walk Through

Live migrate all the guests from the node you wish to upgrade to SP1. Make sure the host is fully patched and disable any antivirus services if you are running any. I always reboot the node before a major upgrade to make sure we have the server in a clean state with  no lingering reboots waiting  or processes can cause issues.

Navigate to the service pack 1 file for Windows 2008 R2, it’s called windows6.1-KB976932-X64.exe and start it up:

You’ll have to accept the license terms:

And then the preparation process starts:

It is now ready to start the upgrade and yes we want it to reboot automatically when needed:

The upgrade process takes a while (about 17 minutes on my lab servers):

 

When it’s done it will reboot and bring you back to the logon screen. Multiple reboots might be needed to complete the upgrade process depending on what’s running on your server. In this case, we are dealing with dedicated Hyper-V nodes.

View when connected to the console

View when connected via RDP

After logging on you are greeted with this window:

And yes this is indeed the case

Reboot included the entire process took about 22 to 23 minutes. In the setup event log you’ll find these messages:

  • Initiating changes for package KB976932. Current state is Absent. Target state is Installed. Client id: SP Coordinater Engine.
  • Package KB976932 was successfully changed to the Installed state.

Note: if an extra reboot is required you’ll see an extra entry in between these stating: A reboot is necessary before package KB976932 can be changed to the Installed state.

When you have a cluster with nodes running both W2K8R2 TM and W2K8R2 SP1, mixed situation so to speak,  you’ll see the following notification in the cluster events:

You can live migrate the guest from the next node to the node already upgraded to SP1 and than repeat the process. You keep doing this until all your nodes are upgraded.

As a final recommendation, I would suggest waiting until you get the SCVMM2008R2 SP1 bits is you use this product before you upgrade your clusters especially when using this with SCOM2007R2 PRO Tips. Otherwise, you don’t need to wait just realize that until you have SP1 for SCVMM2008 R2 you won’t be able to use the new functionality for Hyper-V. In production, I would not recommend using the RC1 for this.

Please do not forget to update your guests with the new SP1 version of the Hyper-V Integration Components. This is needed to be able to use the new features like Dynamic Memory & Remote FX. The Windows 2008 R2 RTM version of the Integration Components is  6.1.7600.16385:

image

You can do this using Hyper-V Manager through selecting “Insert Integration Services Setup Disk”  and running the setup, this will require a reboot.

Click to start the upgrade process.

It will ask to upgrade or repair the previous version:

Work in progress:

Done and asking for a reboot:

SCVMM2008R2 can also be used, here you shut down the guest before updating the virtual guest services as it’s called in SCVMM2008R2. It can be annoying that the nomenclature differs. The good thing here is that you can upgrade multiple guest using VMM2008R2. Hans Vredevoort did a blog post on this here: http://www.hyper-v.nu/blogs/hans/?tag=windows-server-2008-r2-sp1.  After the upgrade you can see that the version of the Integration Components  for Windows 2088 R2 SP1 is  6.1.7601.17514:

Windows 2008 R2 / Windows 7 is RTM today (for real this time)

After an initial false start (https://blog.workinghardinit.work/2011/01/14/windows-2008-r2-windows-7-sp1-rtm-today/) Window 2008 R2/Windows 7 Service Pack 1 has been officially RTMed today . SP1 brings Dynamic Memory& RemoteFX to hyper-V virtualization. I’m probably doing a SQL Server virtualization project and thus I’m very interested in the ability to disable NUMA spanning (watch Ben Armstrong’s Tech Ed 2010 presentation here) http://blogs.msdn.com/b/virtual_pc_guy/archive/2010/06/10/talking-about-dynamic-memory-the-movie.aspx) when beneficial to do so . Which is good news. Until now SQL Server Hyper-V host where better of using machine with lesser CPU sockets & and SSL server VM’s that don’t consume more RAM than that CPU socket can address directly to avoid this.  Until now, for the  environment at hand, I’m leaning to virtualizing SQL Server on it’s own Hyper-V cluster for that reason. It will have to be confirmed in a test environment to see how big the impact is. Systems do differ and get better every year. Perhaps I’ll get back to that subject later. Anyway The bits should be on TechNet/MSDN on February 16th 2001 and available to the general public on February 22 2011. Read the announcement by Microsoft here Windows Server 2008 R2 and Windows 7 SP1 Releases to Manufacturing Today

Shameless Plug For Mastering Hyper-V Deployment By Aidan Finn

In October 2010 Aidan Finn (MVP) his book “Mastering Hyper-V Deployment” was released and in November three copies of this book landed on my desk. I bought them (pre order) via Amazon. Nope I did not get them as a gift or anything. Why Three? Well that’s the number of people I wanted to get up to speed about Hyper-V and virtualization management and operations in a Microsoft environment.

His book takes you along a journey through a Hyper-V project that will teach you about virtualization in all it’s aspects. It also touches on many supporting technologies and products such as System Center Virtual Machine Manager 2008 R2, System Center Essentials 2010, Data Protection manager 2010 and System Center Operations Manager 2007 R2. No one book can be the only source of knowledge and understanding, but using this book as a start for both new and experienced IT Pros to learn about virtualization with Hyper-V will give you the best possible start. Consider it going to an Ivy league college on a scholarship paid for by Aidan’s experience and hard work. The subsidized tuition fee is the price of the book.

We feel a bit sorry that Aidan only got one copy so we made a group picture of the gang of three on the desk of our newest team member. He got a copy of the book together with 4 recycled PC’s and a TechNet subscription to build a lab.

If you know people who want or need to learn about Hyper-V, you’d do well to make sure they get this book and have them set up a lab to play with the technologies. Those efforts will pay off big time when they implement their solutions in the wild. If Ireland is doomed it won’t be because of smart & hardworking Irish IT professionals like Aidan. You see when you design, build and support IT solutions that your customers depend on 24/7 you can not hide behind false promises, you can’t fake away from the fact when “stuff” doesn’t work or hide behind vast amounts of papers & documents void of any substance. Nope, you are responsible for everything and anything you build. Aidan backed and supported by some very knowledgeable colleagues has made that burden a bit lighter for you to bear with this book. Aidan’s blog lives here: http://www.aidanfinn.com/