"Invalid configuration for device '0'" when migrating VM with RDM disks to new hosts.
search cancel

"Invalid configuration for device '0'" when migrating VM with RDM disks to new hosts.

book

Article ID: 440553

calendar_today

Updated On:

Products

VMware vSphere ESXi

Issue/Introduction

  • vMotion or cold migration to a new host of a VM with RDM disks fails with error:

    "Invalid configuration for device '0'"

  • Following removal of the RDM disks from VM, migration completes. However the above error is reported when attempting to re-add the RDM disks to the VM (with the same RDM pointer file)

Environment

VMware vSphere ESXi (any version)

Cause

  • The RDM LUN is presented with different LUN IDs on the source and target hosts.

  • When a new RDM disk is added to a VM, an RDM pointer file is created. This records the RDM LUN VML ID whcih is generated on ESXi level incorporating both the LUN ID and naa/eui ID.

  • If the RDM LUN is presented to the target host from the storage with a different LUN ID, then it will have a different VML ID for the RDM LUN. On migration, of the VM the target host therefore cannot identify the RDM LUN specified in the pointer file, and migration fails.

Resolution

Ensure that the RDM LUN is presented to all ESXi hosts with the same LUN ID.

To do so for an RDM LUN already in use:

  • Shut down the virtual machine in which the RDM disk is incorrectly appearing. 
  • Remove the RDM(s) from any virtual machine accessing it.
    • From VM "Edit settings" remove and RDM disk from the VM, checking the "Delete from disk" check box. (This will delete the pointer file. It will not delete the data on the LUN.)
  • Unpresent the LUN from the ESXi hosts, to which it was presented. For instructions, see How to detach a LUN device from ESXi hosts (323128)
  • Perform a rescan of the storage on the ESXi hosts. For more information, see Performing a rescan of the storage on an ESX/ESXi host (308546).
  • Re-present the LUN to the ESXi hosts with a consistent LUN ID.
  • Perform a rescan of the storage on the ESXi hosts.
  • Re-attach the RDM disk to the virtual machine, as a new RDM disk.

A workaround:

  • Remove the RDM LUN from the VM (as above)
  • Migrate the VM to the new host
  • Add the LUN back to the VM as a new RDM (recreating the pointer file).

    This workaround will allow the VM to power on the target host. However, future migration of the VM to another host may fail. if the there is a mismatch between the LUN ID with which the LUN is mapped to the two hosts.
  •  

Additional Information

See Correlating VML ID with disk ID.