IBM FlashSystem

IBM FlashSystem

Find answers and share expertise on IBM FlashSystem


#Storage
 View Only

  • 1.  IBM PBHA volume Microsoft cluster validation fails

    Posted Thu November 20, 2025 10:57 AM

    Hi,

    Currently we have several IBM flashsystems (FC connected) in a hyperswap setup with Microsoft hyper-v clusters (2022) as hosts. This setup is working great for us, but since we need to move on to pbha we ran into an issue.

    we have configured two new system setups with each two Flashsystems in a PBHA configuration and some Microsoft server 2022 hyper-v hosts.

    We mapped a shared volume to the new hosts from the PBHA storage partition, but the Microsoft cluster validation fails with error "The disks are unsuitable for clustering"
    If we map a disk from outside of the pbha partition as local only on the same flashsystems, the cluster disk validation is successful. (This Microsoft cluster validation is also successful on a hyperwap shared volume)

    The Hyper-v cluster is working with the disks (we tried, but since the validation is failing we cannot run this in production)

    Anyone who has expirienced the same issue? (We already have a support case open with IBM and Microsoft but not making any progress for the moment)

    We tried this with spectrum virtualize code 7.7.3.3 and with another system running the new 9.1.0.1 code level.

    Regards

    Kurt



    ------------------------------
    Kurt De Wit
    ------------------------------


  • 2.  RE: IBM PBHA volume Microsoft cluster validation fails

    Posted Thu November 20, 2025 12:14 PM

    Hello Kurt,

    For what it's worth, we also performed a (FC) Hyper-V migration from Hyperswap to PBHA (9.1.0.1). This without any issues. However, this involved one Flashsystem per site, not, as you indicated, "Two new system setups with two flash systems each." Perhaps something is wrong there (Clustered Flashsystems).



    ------------------------------
    TMasteen
    ------------------------------

    Original Message


  • 3.  RE: IBM PBHA volume Microsoft cluster validation fails

    Posted Fri November 21, 2025 04:32 AM
    Edited by Kurt De Wit Fri November 21, 2025 04:34 AM

    Hi,

    Meanwhile we have received an update from support.
    It seems to be related to the fact that PHBA consists of two separate systems and hyperswap was actually one cluster.
    More information: https://learn.microsoft.com/en-us/troubleshoot/windows-server/high-availability/failover-cluster-storage-validation-tests-dont-discover-all-shared-luns



    ------------------------------
    Kurt De Wit
    ------------------------------

    Original Message


  • 4.  RE: IBM PBHA volume Microsoft cluster validation fails

    Posted Fri November 21, 2025 08:01 AM
    Edited by Thiago Lucas Fri November 21, 2025 09:58 AM

    When you presented the PBHA shared volume to the hosts, storage firmware was already at 8.7.3.3 ?



    ------------------------------
    Thiago Lucas
    ------------------------------

    Original Message


  • 5.  RE: IBM PBHA volume Microsoft cluster validation fails

    Posted Tue November 25, 2025 02:59 AM
    Edited by Kurt De Wit Wed November 26, 2025 12:33 AM

    We tried with two different PBHA setups, one setup with 2 systems on 8.7.3.3 and one setup with 2 systems on 9.1.0.1 before we configured the HA partition.



    ------------------------------
    Kurt De Wit
    ------------------------------

    Original Message


  • 6.  RE: IBM PBHA volume Microsoft cluster validation fails

    Posted Wed November 26, 2025 02:36 PM

    Hi.

    Yes! We've faced the same issue.

    The VPD Serial Number is not the same in both hosts. In HyperSwap it is indeed, because all controllers acts as an array. But this changes in PBHA. 

    If you force the LOCATION parameter on the hosts on the partition and you set the same, the validations pass. 

    It seems it is an issue with MS validation tool, because MPIO works as pretended. 

     
     
    Cause
    The storage validation tests that the wizard runs select only shared LUNs. For a shared LUN, the disk signatures, device identification number (page 0x83), and storage array serial number are the same on all cluster nodes. When you use site-to-site mirroring, a LUN in one site (site A) has a mirrored LUN in another site (site B). These LUNs have the same disk signatures and device identification numbers (page 0x83). However, the storage array serial numbers are different. Because of the difference, the wizard doesn't recognize that the LUNs are shared.
    Regards. 
    =============
    Sergio Rivas
    =============


    ------------------------------
    Sergio Rivas
    ------------------------------

    Original Message


Related Content