John Howard - Senior Program Manager in the Hyper-V team at Microsoft

Senior Program Manager, Hyper-V team, Windows Core Operating System Division.

Blogcast on VSMT (Virtual Server Migration Toolkit) - Part 5

Blogcast on VSMT (Virtual Server Migration Toolkit) - Part 5

  • Comments 14
  • Likes

As promised, here's the last blogcast installment walking through a VSMT migration of an NT4 server into a virtual machine running under Microsoft Virtual Server 2005. This blogcast walks through deploying the disk image into the blank VM and seeing the final result. It can be viewed by clicking here (10 minutes 6 seconds).

If you missed any of the previous parts:

 Part one can be viewed here
 Part two can be viewed here
 Part three can be viewed here
 Part four can be viewed here
 

Edited by John: 3rd Nov 2005 - Rehosted WMV File

Comments
  • Hello,

    We are trying to migrate a physical NT4.0 server to a Virtual server exactly like is being done in this article. Everything is working up until we try to create the Virtual server with createVM.cmd. At the end of this process we are getting the following error.

    <===Remote job output
    Add device in ADS:(XOCOFASTEST, 00010245927E)
    Take control of device: XOCOFASTEST
    set patch path variable to: "D:\VSMT\XOCOFASTEST\Patches"
    set script path variable to: "D:\VSMT\XOCOFASTEST"
    set windows partition variable to: \device\harddisk0\partition1
    set windows directory variable to: WINNT
    set windows drive variable to: C:
    set boot partition variable to: \device\harddisk0\partition1
    set the name of the source device used to generate scripts: XOCOFASTEST
    Use discovery to populate created VM info into ADS
    File or assembly name Microsoft.VirtualServer.Interop, or one of its dependencie
    s, was not found.
    Error: Result Code: 2
    Error occured, Cleaning up.:(2)
    release control of device XOCOFASTEST
    remove device XOCOFASTEST
    Delete Virtual Machine:XOCOFASTEST on vsmt01.vsmt.local

    Does anyone know why this is accuring?

    --
    Darrell

  • Darrell

    I don't know the answer, not a common problem. However, from the description above, it looks like there is a problem with an element of VS or VSMT missing (hence the Interop or dependency not found error). I would recommend that you reinstall VS and VSMT to make sure that everything is in place, or alternatively, maybe post this up on the microsoft.public.virtualserver newsgroup.

    Regards,
    John.

  • For those souls who made it here (like me) that had problems with the DeployVM script with "network circuit disconnected" errors in their job history details or other errors.. Try this hotfix that John mentioned (although I was unable to find in the White Paper):

    http://support.microsoft.com/?kbid=875533

    It appears to have fixed my issues! Kudos to John for an excellent blogcast!

    |\\arco..
    marconielsen@msn.com

  • Thanks! Glad you've got everything going. I should have mentioned exactly where the troubleshooting information can be found. I'll have to listen again, if I did say a white paper, I apologise - it's actually in the VSMT help file. Unfortunately the VSMT.chm help file isn't directly linked from the start menu, so most people don't know it's there.

    If you navigate to %Program Files%\Microsoft VSMT\Help and open the file VSMT.chm, there's a troubleshooting section under "Virtual Server 2005 Migration Toolkit Deployment Guide".

    There's a section titled "The deployment step failed with a 'Timeout while waiting for the device to boot to DA. Exiting' Message" section which mentions this KB article.

  • Hello John,
    When running the DeployVM.cmd, I get as far as the <server_hostname>_Deployvm.xml command and then it hangs and then I receive an error code 8 and fails. I have turned on -trace and -verbose in the .cmd file, but don't know why it is failing. Has anyone else come across this problem?

    Thanks,
    Karl

  • Karl - can you email me or post up the contents of the deploy step output. Are there any other clues in there which might point to something. Have you tried turning off network unloading as documented in the troubleshooting section (see my previous feedback comment), and installed the hotfix for DHCP? I can't specifically recall what error number I was getting before I installed the hotfix, and can't see it mentioned in the guide. Hence any further info would make it much easier to track down.

    Thanks,
    John.

  • John,
    I did apply the hotfix this morning, but it did not seem to help. I don't receive the network error that the hotfix references, so it did not seem to help me with this issue. The <virtual_machine>_deployvm.cmd command is the last script as you well know, the "Deploying:<virtual_machine>_DeployVM.xml" message is the last that is displayed after the DA connection is made when running the .cmd script above, then the ERROR code 8 appears.

    As I write this I did make a change...I enabled a second NIC on the host and allowed all out-of-band mgmt occur through this NIC (like RDP). The virtual guest OS (from the physical server) is still using the VM0 network card that points to a different internal NIC on the host server (make sense?). With this configuration I am able to at least get a restore progress counter when viewing the virtual OS and it seems that the Deploy.xml piece of the code is working....but it is painfully SLOWWWW!!! My .vhd is 35GB, the two captured partitions are 6GB total. It has been running for 3 hours and I am just at 13%. Does this seem normal? The host is a 4 CPU Dell 6650 with 8GB of RAM and there are no other guest OS'es running.

    Thanks,
    Karl

  • I wouldn't expect it to be quite that slow - are the two boxes on the same network segment or are you being impacted by other traffic? I'd leave it to run for the moment and see how far it gets by the morning. FWIW, my VSMT trial with 1/2GB of data took about 10 minutes to deploy on a very slow laptop drive, but this was local network traffic. Extrapolating, 6GB should be a lot less than 3 hours! Offloading network functionality can make a difference as I understand it.

  • John,
    I let it run all night...going on 16 hours. It is at 90% and still on the Deployvm.xml line stating that it is executing. To answer your latest question, yes the server is on its own switch (in our lab) and both NIC's go into the switch and the NIC's are on the same private subnet. I verified the duplex and negotiation is at 100/Full from the NIC's to the switch ports. Through task manager on the host, there is no NIC saturation and about 1% utilization.

    The source images (2 partition) are on this server locally and the .vhd is on a fiber attached EMC CX300 SAN. I don't see disk I/O as an issue here.

    Do I have to run 2 NIC's for this process to work, or can this all occur within the same NIC on the machine? It just so happens that the only way I could get this to work is with 2 NIC's.

    Thanks,
    Karl