Are your Windows 2003 VMs reporting just 1 processor?

Yesterday I’ve been asked to get some information on a Windows 2003 VM, since this is a remote VM managed with OpsMgr the natural choice has been to ask OpsMgr for such data. I must admit this is not the easiest task of it all, given the fact that OpsMgr misses an holistic view of discovered data for a specific managed device (this is an addon I must find the time to develop). The way to go is to figure out in which class the information is hidden and then use the discovered inventory view properly targeted. Starting points are typically the Windows Computer and Windows Operating Systems classes:


When I checked for the number of Logical Processors, with my surprise, just 1 was reported, while I was pertty sure the VM had 2 cores assigned.

It turned out WMI in Windows 2003 has an issue with multicore or hyperthreading processors and Hyper-v allocates the guest VMs one virtual multicore processor. The VM is effectively using the allocated cores, but WMI reports just one processor. Luckily there’s a fix for this behavior: KB 932370 The number of physical hyperthreading-enabled processors or the number of physical multicore processors is incorrectly reported in Windows Server 2003.

– Daniele

This posting is provided "AS IS" with no warranties, and confers no rights.

  1. #1 by Brian Pavnick on September 17, 2011 - 6:37 pm

    Hello Daniele,

    Nice post! Assuming this VM is running on VMware, you can leverage Veeam’s nworks MP for VMware to gather information about that VM such as vCPUs, vCPU Reserves and Limits, vRAM, vRAM Reserves and Limits, vDisk, vNIC. Mind you, this method of data collection is not dependant on the SCOM Agent or OS running within the VM. The VM could be a non-managed Linux server, and we will stil get this level of detailed information for it.

    Additionally, you will have the ability to collect configuration, performance, and availability infromation about the VMware Clusters, Hosts, Resource Pools, vApps, and vCenters.

    If interested, feel free to drop me an e-mail and I can explain in more detail on how Veeam’s nworks MP makes this happen.

    Keep up the good posts!

    Brian Pavnick
    Sr. Solutions Architect w/ Veeam Software

    • #2 by Daniele Grandini on September 19, 2011 - 8:55 am

      Hi Brian, thanks for reading my blog.
      You’re right, we use Veeam nworks for ESX monitoring and it is comprehensive indeed, much better than what we have for Hyper-v. Nevertheless being able to treat all servers (physical or virtual) the same way it is an important value of System Center in general and OpsMgr in particular. Regarding nworks, immo, to make it even better you should think to start with fewer enabled rules and monitors: some kind of basic monitoring with only KHIs and then an ovveride MP to enable the full power when needed. In loosly manged environment the risk is to have too much noise. Another annotation on discovery, you should take some step to make it more reliable, we have discovery spikes when the xml document generation fails (for any reason, but it happens) and the MP undiscover and rediscover everything. Currently nworks is tha major discovery generator in our environment. Scaled this issue to support, alas we hadn’t being able to ge through the first level.

      – Daniele

  1. OU, Logical Processor, Physical Processor, and other class properties are not discovered with OpsMgr 2012 on Windows Server 2003 - Kevin Holman's System Center Blog - Site Home - TechNet Blogs

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s

%d bloggers like this: