none
"Missing" VMs in SCVMM 2007? RRS feed

  • Question

  • I can't seem to find anything on this.

    I have several virtual machines with "missing" status. Repair options are all grayed out. This is in VM2005 and SCVMM 2007 is the front end. This happened after a coworker rebooted the host without shutting the virts down (grr).

    The VM folder is still there where it's always been. The files all appear normal.

    Where do I go to bring these back?
    Sunday, March 7, 2010 7:50 PM

Answers

  • Typically I see this occur when one of the cluster nodes experiences an unexpected outage.  When that happens, failover cluster performs by design and moves the highly available virtual machines (HAVMs) to other nodes.  When the node that had the outage is brought back online, then VMM "remembers" that it used to own "virtual_machine_name" and since it is no longer found on that host, the status displayed is "missing".

    Now, the question becomes how can I remove that VM from the VMM console if it truly does not exist on that cluster node?

    There are a number of ways that I have had success with in correcting this issue.
    1.    In Failover Cluster Manager (FCmgr), fail the VM back to the original owner.  And then from the Actions pane select "refresh virtual machine configuration".  Then back in the VMM console perform a Refresh of the cluster.  This seems to have more success if attempted very soon after the outage.
    2.    In the VMM console, remove the host cluster and re-add it back.  This should refresh all of the VM information for each host.  Kind of like a hard reset.
    3.   Perhaps the easiest way to resolve this is to simply delete the missing VM from within the VMM console.  However, before attempting this method you must be absolutely positive that the VM is indeed not on this same Hyper-v host and is running on another cluster node.

    As I said, I've used each of these methods to recover from this scenario.  Please reply back and let us know which one you used and the results.  Also mark this post as "Answer" if it resolves your issue.



    Thanks,

    Mike Briggs [MSFT]
    http://blogs.technet.com/mbriggs


    ** This information provided "AS IS" with no warranties, and confers no rights **

     

    Sunday, March 7, 2010 11:06 PM
    Moderator

All replies

  • Typically I see this occur when one of the cluster nodes experiences an unexpected outage.  When that happens, failover cluster performs by design and moves the highly available virtual machines (HAVMs) to other nodes.  When the node that had the outage is brought back online, then VMM "remembers" that it used to own "virtual_machine_name" and since it is no longer found on that host, the status displayed is "missing".

    Now, the question becomes how can I remove that VM from the VMM console if it truly does not exist on that cluster node?

    There are a number of ways that I have had success with in correcting this issue.
    1.    In Failover Cluster Manager (FCmgr), fail the VM back to the original owner.  And then from the Actions pane select "refresh virtual machine configuration".  Then back in the VMM console perform a Refresh of the cluster.  This seems to have more success if attempted very soon after the outage.
    2.    In the VMM console, remove the host cluster and re-add it back.  This should refresh all of the VM information for each host.  Kind of like a hard reset.
    3.   Perhaps the easiest way to resolve this is to simply delete the missing VM from within the VMM console.  However, before attempting this method you must be absolutely positive that the VM is indeed not on this same Hyper-v host and is running on another cluster node.

    As I said, I've used each of these methods to recover from this scenario.  Please reply back and let us know which one you used and the results.  Also mark this post as "Answer" if it resolves your issue.



    Thanks,

    Mike Briggs [MSFT]
    http://blogs.technet.com/mbriggs


    ** This information provided "AS IS" with no warranties, and confers no rights **

     

    Sunday, March 7, 2010 11:06 PM
    Moderator
  • Hi Mike, thanks for your reply. I have seen that post elsewhere, but wonder about how it applies to me. The host is up. It's just a few specific VMs on the host, others are fine. And we don't use Clusters for our VMs (yet?) I just have 3 individual hosts.

    Thanks!
    Tuesday, March 9, 2010 1:43 AM
  • So in that case #1 definitely does not apply in your scenario since you're host is not a member of a failover cluster.  But the known work-arounds or resolutions apply to anytime the "missing" status is displayed for a VM.  So options #2 & #3 do apply and either would resolve this issue.

    If you opened a case with CSS for this issue, and we got on the phone or remote session, one of the first things that I would look at is how Hyper-v manager displayed the VMs on this particular host.  If the VMs that have the "missing" status in VMM are not listed in Hyper-v manager, then we would either delete them from the VMM console or remove / re-add the host in VMM.  Options #2 & 3 above :-)

    Either one of these work-arounds should fix the issue.  Unless somehow I've completely misunderstood!  Check in Hyper-v manager and reply back and let me know if the "missing" VMs are really there.  And if not, follow either suggestion to remove them from VMM and let me know the outcome.


    Thanks,

    Mike [MSFT]
    Wednesday, March 10, 2010 1:44 AM
    Moderator
  • Hi Duroc,

    If I understand your issue correctly you are running Virtual Server 2005 managed by SCVMM 2007.  I no longer use this in my environment but I seem to remember having a similar problem.  The issue should be able to be resolved by one of the two methods below...

    1)  Use Virtual Server Administration page to update (I seem to remember it telling you when there was something wrong) and correct the component (remove connections to missing files etc).

    2)  Create another Virtual Machine, then using VS Admin Page and Windows Explorer copy the vhd files into place, and use the VS admin page to connect to those vhds.

    After both you will want to refresh the VM using SCVMM it will then show the correct information.

    Also I would encourage you to move forward towards SCVMM 2008 R2 + Hyper-V 2008 R2.

    -matt

    If this post is helpful, please mark it as such.
    Thursday, March 11, 2010 10:26 PM
  • Thanks Matt.

     

    Sorry it's been so long since I replied. I can't find where to set email notifications for Technet and didn't save the link. Today is the first chance I had to log in again and look my post up.

    This environment was 'inherited'. It seems the previous admin disabled much that I would need to follow your instructions. When I attempt to launch the admin page, I get "...cannot display the webpage". IIS is started and running, but I can't access the page.

    There are other bits I am running into (HOST NOT RESPONDING happens frequently when managed by the SCVMM server, libraries don't work right, etc). I tried installing SCVMM on another 2k3 server, but whenever I try to add the host

    I think I'd like to rebuild this from scratch. There's no budget for a SCVMM 2008 purchase (we're a church and k-8 school), so I have to work with what I've got. I can't see how to import an existing virtual machine if I back them up then rebuild. Can't find a thing in the help files or online. Can I get a link or something to get me able to do that? This host has 30 WinXP virtuals to provide IE and Windows services to the Mac using school admin staff here.

    Thanks!

    Don

     

    EDIT: I think (hope) I found my notifications buried under my Live settings. Sigh...

    Sunday, March 28, 2010 8:46 PM