Error: Unable to enumerate all disks for virtual machines with RDM disks
search cancel

Error: Unable to enumerate all disks for virtual machines with RDM disks

book

Article ID: 424749

calendar_today

Updated On:

Products

VMware vSphere ESXi

Issue/Introduction

Virtual machines configured with Raw Device Mapping (RDM) disks may experience sudden power-on failures or become unresponsive during critical operations. This behavior typically follows an ESXi host reboot, kernel patching, or a vMotion migration where the destination host lacks proper storage presentation. When the underlying storage LUN is not visible to the host storage stack, the VM cannot resolve its disk descriptor files, resulting in a boot failure. This guide addresses how to identify storage path failures, verify NAA ID visibility via the CLI, and restore access to affected virtual disks in a vSphere environment.

Symptoms

  • A virtual machine (VM) fails to power on after a host reboot, maintenance, or being migrated (vMotion) to a new ESXi host.
  • The power-on task fails with the error message: Unable to enumerate all disks.
  • In the vSphere Client:
    1. Select the affected VM.
    2. Click Edit Settings.
    3. Multiple VMDKs are listed, but one or more Raw Device Mapping (RDM) disks show a disk size of 0 MB.
  • Host logs (vmkernel.log) show All Paths Down (APD) or path selection failures: 
    WARNING: vmw_psp_rr: psp_rrSelectPath:2295: Could not select path for device "naa.####".
  • vmkernel.log contains Permanent Device Loss (PDL) entries: [vob.scsi.device.state.permanentloss] Device :naa.#### has been removed or is permanently inaccessible.
  • SCSI sense data reports 4h/3Eh/01h.
  • Connectivity issues on TCP port 5044 (if using IO splitters).

Environment

  • VMware vSphere ESXi 7.x / 8.x / 9.x.
  • Virtual Machines using RDM disks (Physical or Virtual  compatibility mode).

Cause

This issue occurs when the backend storage device for one or more Raw Device Mapping (RDM) disks is not visible or accessible to the ESXi host.

Common triggers include:

  • Storage Zoning/Masking: The LUN is not presented to the host on the storage array (e.g., PowerMax, Unity).
  • Connectivity: Physical or fabric issues leading to an APD (All Paths Down) state.
  • vMotion: The target host in a migration does not have access to the same shared RDM LUNs as the source host.

As a result, the ESXi storage stack cannot resolve the RDM descriptor file, preventing the VM from enumerating the disks during power-on.

Resolution

Ensure all backend storage devices for the RDM disks are properly attached and visible to the ESXi host on which the VM resides.

  1. Identify the affected RDM disk in the VM Edit Settings and record the NAA ID.
  2. Coordinate with the storage team to verify LUN masking and zoning for the affected NAA ID.
  3. Log in to the vSphere Client, navigate to the ESXi host > Configure > Storage Devices.
  4. Locate the device and click Attach if the status is "Detached".
  5. Perform a storage rescan: esxcli storage core adapter rescan --all
  6. If the VM still fails to power on, remove the RDM from the VM configuration (Remove from virtual machine only) and re-add it as an Existing Hard Disk to refresh the pointer files.
  7. Verify port 5044 is open if using an IO splitter.

Additional Information