locked
Unexpected shutdown of VMs after Hyper-V host is rebooted RRS feed

  • Question

  • Hi,

    we are running Hyper-V 2012 and 2012R2s host with Windows Server 2012 VMs.

    VMs are configured to go into saved state.

    But after Windows updates (SCCM) and after some scheduled reboots we are experiencing bad issues
    within the OS of the VMs one can see "unexpected shutdown" events which causes trouble  with DFS-R for example

    So my questions:
    1. If we schedule reboots with the  "shutdown" command could it be that some VMs are not completely put into saved state and the hosts reboots anyway?

    2. If there is the "/force" Parameter is this a difference?
    So does the host behave differently then regarding the VMs?

    3. Do DFS-R configurations get into trouble when using "saved state"? So would it be recommended to use "shutdown" instead?

    Thanks in advance!!

    Kind regards

    /bkpfast


    My postings are provided "AS IS" with no warranties and confer no rights

    Wednesday, February 17, 2016 3:57 PM

Answers

  • Some details that might help -

    The VM Saving State is not a guarantee.  It is a best attempt and there are ways that prevent it from happening.

    First of all, Saving the VM running memory state is attempted only when there is a clean shutdown / reboot of the host initiated.  A best attempt is made to save each VM.  This process takes time, especially if the VMs have large memory allocations.

    However, if the saving process is taking too long on a particular VM the Host will reboot anyway in an attempt to reduce the overall downtime for all VMs on that Host.

    Also, if the shutdown is initiated on the command line and forced, then the shutdown / reboot is immediate.  Yes a faster reboot.

    Personally, I have all my VMs set to shutdown And then power on.  I never save the memory state.  This way I always get application consistency for whatever applications are running in the VMs (assuming the applications in the VM can handle a clean shutdown and power on).


    Brian Ehlert
    http://ITProctology.blogspot.com
    Learn. Apply. Repeat.

    • Proposed as answer by Elton_Ji Monday, February 29, 2016 3:45 PM
    • Marked as answer by Elton_Ji Monday, March 7, 2016 1:39 AM
    Wednesday, February 17, 2016 4:34 PM

All replies

  • Some details that might help -

    The VM Saving State is not a guarantee.  It is a best attempt and there are ways that prevent it from happening.

    First of all, Saving the VM running memory state is attempted only when there is a clean shutdown / reboot of the host initiated.  A best attempt is made to save each VM.  This process takes time, especially if the VMs have large memory allocations.

    However, if the saving process is taking too long on a particular VM the Host will reboot anyway in an attempt to reduce the overall downtime for all VMs on that Host.

    Also, if the shutdown is initiated on the command line and forced, then the shutdown / reboot is immediate.  Yes a faster reboot.

    Personally, I have all my VMs set to shutdown And then power on.  I never save the memory state.  This way I always get application consistency for whatever applications are running in the VMs (assuming the applications in the VM can handle a clean shutdown and power on).


    Brian Ehlert
    http://ITProctology.blogspot.com
    Learn. Apply. Repeat.

    • Proposed as answer by Elton_Ji Monday, February 29, 2016 3:45 PM
    • Marked as answer by Elton_Ji Monday, March 7, 2016 1:39 AM
    Wednesday, February 17, 2016 4:34 PM
  • Hi Sir,

    I would like to check the current state of that issue .

    Best Regards,

    Elton


    Please remember to mark the replies as answers if they help and unmark them if they provide no help. If you have feedback for TechNet Subscriber Support, contact tnmff@microsoft.com .

    Friday, February 19, 2016 11:53 AM