none
hyper-v windows 2012 R2 - Move VM from one drive to another drive on the same Hyper-v server questions?

    Question

  • Goal - move VM's from one drive to another drive on the same hyper-v host 2012 R2.


    If we use the "move" command move a VM to another drive on the same server will this process make another copy while this move occurs to another drive.  The reason I am asking this is that the drive we are moving from does not have enough space for a second copy to be made during the move.  There is less than 20 GB free space on this drive with multiple VM's.   I do not want to break any VM on this drive while trying to move to another drive on the same server.***Does it make a difference whether we shut down the VM during the move or if the VM is still powered on.    Will this move occur quicker if the VM is shut down versus if the VM stays turned on?

    So we likely will be moving more than one VM to this new drive.

    VM1 - two IDE controller 0 attached vhdx's and one SCSI attached vhdx.

    VM2 - one IDE controller 0 attached vhdx(OS) and two SCSI attached vhdx's.

    VM3 - one IDE controller 0 attached vhdx(OS) and five SCSI attached vhdx's.

    Note - VM1 has two IDE controller attached vhdx's of which one is for the OS and the other is another data vdhx.  Will both of these vhdx move together or is there an additional step?

    How does the move command work with mutiple vdhx?  Are there any setting changes that might optimize this move?  Does the move option require me to move the SCSI attached vhdx in a separate move?




    dsk



    • Edited by kimdav111 Thursday, March 9, 2017 4:11 AM addiitonal questions
    Thursday, March 9, 2017 3:40 AM

All replies

  • Hi,

    Running VM will allways take longer time to migrate since you additionally need to copy content of actual used RAM. 

    It will not take additional disk space on source location.

    You can optimize move speed by using SMB direct with RDMA on WIndows server 2016.

    No, you don't need to do separete steps when moving disks on different controllers (IDE, SCSI).

    Radek

    Thursday, March 9, 2017 8:09 AM
  • Are you positive that it will not take additional space at the source drive?   From my experience with VMM live migrations that it creates a snap at the source which will revert if there is failure or a network interruption during the migration. If there is no space required at the source lun then what happens if there is an interruption like an accident reboot?   What allows it to revert to the original state at the source location with the hyperv move command.  And we do not have space for backup of this VM.

    dsk


    • Edited by kimdav111 Friday, March 10, 2017 8:59 PM clarificatino
    Friday, March 10, 2017 8:58 PM
  • You can find something yourself: https://youtu.be/IPWIS6zj8jU

    In a short:

    admin. initiates live migration (full vm or storage)

    hypervisor creates a container of VM or storage

    hypervisor fills container with info

    hypervisor replicates delta (info that was changed after pervious step started)

    hypervisor switches resources to new location and deletes old resources.

    you can cancel a job in hyper-v manager or failover cluster manager by right-clicking the res.

    If you think of accident reboot... well, it's not normal operation of server.

    • Proposed as answer by Serge Poyarkov Friday, March 10, 2017 10:37 PM
    Friday, March 10, 2017 10:37 PM
  • It does not have to be a reboot.  A service could get hung and you have to restart the service or reboot the server.  What happens to the MOVE option?,... I am assuming it rolls back automatically.   What space is required at the source LUN for the roll back to occur properly so that there is no data corruption?  My understanding that MOVE creates a snap at the source to roll back if there is a disruption in the MOVE.  How much space do we need for this snap??,... we are moving slightly over 2.3 TB when there is only 67 GB free space at the source LUN.

    dsk


    • Edited by kimdav111 Saturday, March 11, 2017 6:32 PM correction
    Saturday, March 11, 2017 6:23 PM
  • There is nothing to 'roll back' on the source.  The source continues to run as is no migration was taking place.  If anything happens at the destination, such as the VM not able to start, the whole process is aborted and the source just continues to run.  If the source crashes, it is the same as if there was no migration in process.  It simply aborts the migration and the source crashes

    . : | : . : | : . tim

    Sunday, March 12, 2017 7:23 PM
  • Hi,
    Are there any updates?
    You could mark the reply as answer if it is helpful.
    Best Regards,
    Leo

    Please remember to mark the replies as answers if they help.
    If you have feedback for TechNet Subscriber Support, contact tnmff@microsoft.com.

    Monday, March 27, 2017 2:02 AM
    Moderator