none
Various errors - The wrong diskette is in the drive + Failover cluster fails creating new virtual machines. RRS feed

  • Question

  • Hello

    Our rig.

    2 nodes - hp dl360 connected to a HP mSA 2040 SAN directly with FC cards.

    Running Failover cluster roles and senveral virtual hyper-v vm's.

    We have 2 LUNS, clusterstorage 1 and 2, several weeks ago, one of the hosts (node2) lost connection to the SAN, we suspect during Veeam backup, node 2 is offhost-proxy for Veeam also, we restarted and it came online again after apporx 30 minutes, and suddenly it happened a couple of days later. And all of a sudden we couldn't bring it online, and failover didn't work, which means, the vm's didn't failover to node1.

    So all the vm's on that LUN got stuck, closed/seized by node2, but we could manually mount that LUN on node1, and start the vm's, but all vm's are now not in failover mode because node2 cannot see that particular LUN/CLusterStorage\Volume1. VOlume2 works on both nodes.

    After many days of searching/hotfixes and so on, we found some references to this:

    There was an error loading the disk information for disk Cluster Disk 2 - Microsoft.FailoverClusters.Framework.ClusterControlCodeException: Failed to execute control code '16777713'. ---> System.ComponentModel.Win32Exception: The wrong diskette is in the drive.
    Insert %2 (Volume Serial Number: %3) into drive %1
       --- End of inner exception stack trace ---
       at MS.Internal.FailoverClusters.Framework.ClusApiAdapter.ResourceAdapter.ExecuteOnControlCode(SafeClusterResourceHandle resourceHandle, Int32 controlCode, String resourceName, UnmanagedBuffer inBuffer, Action`2 controlCodeCallBack, Action invalidFunctionCallback)
       at MS.Internal.FailoverClusters.Framework.ClusApiAdapter.ResourceAdapter.LoadDiskFromCluster(SafeClusterResourceHandle resourceHandle, PResource resource, Boolean includeMountPoints)
       at MS.Internal.FailoverClusters.Framework.ClusApiAdapter.ResourceAdapter.<>c__DisplayClass2d4.<LoadDisk>b__2d1(SafeClusterResourceHandle resourceHandle)
       at MS.Internal.FailoverClusters.Framework.ClusApiAdapter.ResourceAdapter.ExecuteOnResource(Guid id, String name, Action`1 actionOnResource)
       at MS.Internal.FailoverClusters.Framework.ClusApiAdapter.ResourceAdapter.LoadDisk(PResource resource)


    And many people resolved the issue by deleting the volume, and enable it again, and change from MBR to GPT. We on the other hand, annot delete that volume because it active and in production manually on node1. So we decides to buy a new set of discs to the SAN, create LUN3 and backup/restore the manually mounted vm's on defect lun to the new lun3, we did so, ClusterStorage\Volume3 is visible from both node1 and node2, but..... We see the same errors on the new vdisk/lun/volume3...

    When trying to create a VM in failover cluster:

    There was a failure configuring the virtual machine role for 'test4'.
    An error occurred retrieving the disk information for the resource 'Cluster Disk 2'.

    The wrong diskette is in the drive.
    Insert %2 (Volume Serial Number: %3) into drive %1

    This error happens on a brand new set of discs on tha hp msa 2040 san, and a new volume.

    What the hell happened on node2, that causes failover cluster to not owrk on a brand new set of LUN?

    This is critical... 

    Any help is greatly appreciated!

    Best regards 

    Aksel

    Saturday, October 10, 2015 9:46 AM

Answers

  • Hi Aksel,

    What does LUN/CLusterStorage\Volume1 contain, can't you move the data over to VOl3 that both VMs are able to see.

    Are you directly sharing the CSVs or using a shared VHDx. If not try LUN\CLusterStorage\Volume4\Shared.VHDx\Expose it as a disk to the VMs.

    This should work better than using a pass-thru.

    Let me know about the problematic VMs roles,application and data.


    Regards,

    Satyajit

    Please“Vote As Helpful” if you find my contribution useful or “MarkAs Answer” if it does answer your question. That will encourage me - and others - to take time out to help you.



    Hello Satyajit

    We solved the issue, when we deleted the clustervolume which was causing the errors, and brought it up again.

    Monday, October 12, 2015 8:09 PM

All replies

  • Hi Aksel,

    What does LUN/CLusterStorage\Volume1 contain, can't you move the data over to VOl3 that both VMs are able to see.

    Are you directly sharing the CSVs or using a shared VHDx. If not try LUN\CLusterStorage\Volume4\Shared.VHDx\Expose it as a disk to the VMs.

    This should work better than using a pass-thru.

    Let me know about the problematic VMs roles,application and data.


    Regards,

    Satyajit

    Please“Vote As Helpful” if you find my contribution useful or “MarkAs Answer” if it does answer your question. That will encourage me - and others - to take time out to help you.



    • Edited by Satyajit321 Monday, October 12, 2015 11:57 AM
    Monday, October 12, 2015 11:55 AM
  • Hi Aksel,

    What does LUN/CLusterStorage\Volume1 contain, can't you move the data over to VOl3 that both VMs are able to see.

    Are you directly sharing the CSVs or using a shared VHDx. If not try LUN\CLusterStorage\Volume4\Shared.VHDx\Expose it as a disk to the VMs.

    This should work better than using a pass-thru.

    Let me know about the problematic VMs roles,application and data.


    Regards,

    Satyajit

    Please“Vote As Helpful” if you find my contribution useful or “MarkAs Answer” if it does answer your question. That will encourage me - and others - to take time out to help you.



    Hello Satyajit

    We solved the issue, when we deleted the clustervolume which was causing the errors, and brought it up again.

    Monday, October 12, 2015 8:09 PM
  • Hi Aksel,

    Good to know, but you said "We on the other hand, annot delete that volume because it active and in production". How did you manage to keep the data.


    Regards,

    Satyajit

    Please “Vote As Helpful” if you find my contribution useful or “Mark As Answer” if it does answer your question. That will encourage me - and others - to take time out to help you.

    Tuesday, October 13, 2015 4:08 AM