The customer encounters an 'insufficient disk space' error when they attempt to clone a massive virtual machine

Products

• VMware vCenter Server

Issue/Introduction

The customer attempts to clone a 50 TB virtual machine using the standard vSphere Client Clone wizard. The operation fails during the initial pre-flight checks, returning an 'insufficient disk space' error before the data transfer begins.

Environment

vCenter 7.0

Cause

The target datastore does not have enough contiguous free space to support a 50 TB VM copy. Furthermore, attempting to clone a 50 TB VM using the standard vCenter UI wizard is highly inefficient and not recommended. Even if the space requirement is met, pushing 50 TB of data through the ESXi host data mover places a massive, sustained I/O penalty on the storage fabric, potentially takes 3 to 7 days to complete, and severely degrades production storage latency.The 'insufficient disk space' error occurs because the vCenter cloning wizard attempts to reserve the full storage footprint (up to 50 TB, depending on thin/thick provisioning and swap files) on the destination datastore prior to initiating the transfer. If the target datastore lacks this immense amount of raw free capacity, the hypervisor's pre-flight check immediately fails the task to prevent filling the datastore and crashing running workloads

Resolution

Because the goal is to spin up an identical copy of the VM within the same environment, we highly recommend abandoning the vCenter UI Clone wizard entirely. Instead, use one of the following industry-standard methods to duplicate massive workloads:

Method 1: Storage Array Cloning (The Gold Standard)

If the underlying storage is an enterprise SAN or NAS (e.g., Pure Storage, Nimble, NetApp, Dell PowerStore), leverage the storage hardware to do the work.

1. Log into the storage array management interface.

2. Locate the datastore/LUN that holds the 50 TB VM.

3. Trigger a Clone or Snapshot-to-Volume operation on that LUN.

4. Present the newly cloned LUN to the ESXi hosts.

5. In vCenter, rescan storage, browse the new datastore, right-click the .vmx file, and select Register VM.

6. When you power it on, vCenter prompts if the VM is "Moved" or "Copied." Select I Copied It so vCenter generates new MAC addresses to avoid network conflicts.

Method 2: Instant VM Recovery via Backup Software

Modern enterprise backup software (e.g., Veeam, Rubrik, Cohesity) is specifically designed for this scenario (e.g., Veeam's "Instant VM Recovery").

1. Access the backup software and select the most recent backup of the 50 TB VM.

2. Choose the option to restore/publish it to vSphere as a new VM.

3. The backup software temporarily mounts the backup repository directly to the ESXi host via an NFS datastore and powers the VM on.

Note: This boots the VM in minutes without copying 50 TB of data back into production datastores, making it ideal for patch testing or verification.

Method 3: The vCenter Clone Wizard (Last Resort)

If advanced backup software or array access is unavailable, you must ensure the environment supports VAAI (vStorage APIs for Array Integration) before attempting the clone.

• Specifically, the XCOPY primitive must be enabled. XCOPY offloads the block transfer to the storage array internally, saving ESXi CPU and network resources.

• To verify VAAI: In vCenter, select the host > Configure > Storage Devices. Check the "Hardware Acceleration" column; it must say Supported.

Critical Requirement: For XCOPY to work, the source and destination datastores must reside on the exact same physical storage array and share the same VMFS block size. If you attempt to split the 50 TB across completely different arrays, XCOPY fails, ESXi falls back to the slow software data mover, and the operation likely times out.