locked
Does data corruption also occur on WHS running in VM? RRS feed

  • Question

  • Was the data corruption ever observed in WHS running in a Virtual Machine? If not does WHS team consider it likely / possible to happen in VM?

    Friday, May 23, 2008 10:53 PM
    Moderator

Answers

  •  brubber wrote:

    Was the data corruption ever observed in WHS running in a Virtual Machine? If not does WHS team consider it likely / possible to happen in VM?

     

    I don't think you'll get an official response from MS, brubber (since WHS on any VM machine is completely unsupported).  Which is a shame, since I think that's the only way they're going to get a real good eval of PP1 (I don't think most people are going to set up another real machine).

    Friday, May 23, 2008 11:42 PM
    Moderator
  • Brubber, to the best of my knowledge, it's possible for the same conditions that cause corruption on real hardware to cause it in a virtual machine. As a practical thing, I wouldn't expect to see it; it's hard enough to recreate on real hardware.

    As for deploying on hardware, kariya, I expect most people who test will do so on their production servers, hopefully after following Microsoft's guidance regarding backing up their server. Some of us have additional hardware on which we can test, without disrupting our production servers, but I think you're right to believe that we're a small fraction of the total user base.
    Saturday, May 24, 2008 1:47 AM
    Moderator

All replies

  •  brubber wrote:

    Was the data corruption ever observed in WHS running in a Virtual Machine? If not does WHS team consider it likely / possible to happen in VM?

     

    I don't think you'll get an official response from MS, brubber (since WHS on any VM machine is completely unsupported).  Which is a shame, since I think that's the only way they're going to get a real good eval of PP1 (I don't think most people are going to set up another real machine).

    Friday, May 23, 2008 11:42 PM
    Moderator
  • Brubber, to the best of my knowledge, it's possible for the same conditions that cause corruption on real hardware to cause it in a virtual machine. As a practical thing, I wouldn't expect to see it; it's hard enough to recreate on real hardware.

    As for deploying on hardware, kariya, I expect most people who test will do so on their production servers, hopefully after following Microsoft's guidance regarding backing up their server. Some of us have additional hardware on which we can test, without disrupting our production servers, but I think you're right to believe that we're a small fraction of the total user base.
    Saturday, May 24, 2008 1:47 AM
    Moderator
  •  Ken Warren wrote:

    >snip<

    As for deploying on hardware, kariya, I expect most people who test will do so on their production servers, hopefully after following Microsoft's guidance regarding backing up their server.

     

    I searched Connect & these forums & didn't see any mention by the WHS Team or you experts on exactly how one Backs Up their server.

     

    Is there a link someone can post for the instructions to do this?

     

    TIA

    Saturday, May 24, 2008 11:32 PM
  • For your shares:
    • Plug in an external drive.
    • If you connected it to the server, log in to the server using Remote Desktop.
    • Copy files from your shares to the external drive. The easiest way to access the server shares is via the shorcut on every client machine's, and the server's, desktop.
    • disconnect the external drive.
    • Repeat as needed until you have all your files.
    My recommendation is to treat your backup database as replaceable and ignore it. If you want to back it up, you'll find directions in the Home Computer Backup technical brief.
    Sunday, May 25, 2008 4:59 AM
    Moderator
  •  Ken Warren wrote:
    Brubber, to the best of my knowledge, it's possible for the same conditions that cause corruption on real hardware to cause it in a virtual machine. As a practical thing, I wouldn't expect to see it; it's hard enough to recreate on real hardware.

     

    Thanks Ken, and also Kariya21. This is more or less the answer I was expecting and afraid of. If you're not able to reproduce the bug using well defined conditions, there's NO way you can ever be sure you fixed it.

    Monday, May 26, 2008 1:10 AM
    Moderator
  • Oh, no, I think you misunderstand. I'm not saying that it can't be recreated. I'm saying that it's extremely hard to recreate, and as far as know nobody outside Microsoft can give chapter and verse on how to do so. I certainly can't. What I know of the causes from conversations with Microsoft, though, tells me that Microsoft believes they understand the cause of KB946676 pretty well. The scenario they have described (in general terms) is plausible, and includes enough information for me to be pretty certain that it takes just the right circumstances. I don't have enough information to identify those circumstances, however.

    I don't think there needs to be any worry that Microsoft has failed to fix this issue. And the upcoming beta isn't, I think, about whether the issue is fixed, it's about whether new issues have been introduced. I haven't seen any, but widening the pool of potential testers by a couple of orders of magnitude will do wonders for everyone's confidence. Smile

    Monday, May 26, 2008 1:45 AM
    Moderator
  • A condition to reproduce the bug is to have more than one disk in the server. With VMWare, I don't see how you will be able to have 2 disks to reproduce the conditions (Unless I'm not looking hard enough).

     

    Monday, May 26, 2008 2:25 AM
  • True Image by Acronis works for me.

    Monday, May 26, 2008 8:33 AM
  • Xilikon, every virtualization tool I've ever used allows you to define multiple virtual disk drives for your VM.
    Monday, May 26, 2008 12:54 PM
    Moderator