Hyper-V 2016 OS redundancy


  • Hello All!

    I'm testing out Storage spaces direct on a 2 server cluster using Hyper-V 2016. The servers are set to use HBA mode(bypass) and I've installed the OS on a dedicated SSD on each host. The remaining drives will be used for the shared cluster storage. My question comes on the OS redundancy side. While I could lose a few drives utilized by the shared storage, If the SSD containing my OS fails, I would not only lose the host but the storage it provides to the cluster. Perhaps I've missed something or just don't understand how this is supposed to be setup but my understanding is that SDS requires HBA mode, thus disabling my ability to RAID the OS drive. Is this correct and am I just left to reimage the host if my drive fails? How does everyone else deal with these situations? Thanks for any input!

    Tuesday, May 2, 2017 3:16 PM

All replies

  • Hi Sir,

    >>The servers are set to use HBA mode(bypass) 

    If you mean "pass-though" mode , it should cover the requirement of S2D .


    >>If the SSD containing my OS fails, I would not only lose the host but the storage it provides to the cluster.

    Please connect these disk directly to each node as local attached disk ,not shared disks .

    It is a cluster , it allows proper nodes failure .

    Please check the following article regarding cluster quorum:

    Hope it is helpful to you .

    Best Regards,


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

    Wednesday, May 3, 2017 10:07 AM
  • As Elton notes, S2D requires an HBA, not a RAID controller, for the data disks.  You can use a RAID controller for the OS disks, but you need a separate, supported HBA for the data disks.  Talk with your system vendor about which HBA they have tested for use in S2D clusters.  (S2D is a new technology and puts a significant amount of 'stress' on the systems that was not there previously.  Microsoft even has a separate testing procedure for vendors to 'prove' their solutions.  Though it is possible to roll your own solution, it makes a lot of sense to use components the vendors have already tested.  The vendors spent months testing their components and solutions.  If you use untested/unproven components, you will, in essence, be performing the months of testing on your own.)


    Wednesday, May 3, 2017 12:32 PM