VMware ESXi and vSphere Cluster Management

Use vSphere Storage vMotion to Move Virtual Machine Files Between Datastores

Learn how to use vSphere Storage vMotion to move a running VM's disks and configuration files between datastores, choose disk formats, verify results, and troubleshoot failures.

Storage vMotion is vSphere's live storage migration feature. It moves a virtual machine's storage files to another datastore while the guest operating system and applications can remain online. The VM continues to run on its current ESXi host, which is the hypervisor server that executes virtual machines.

A datastore is a vSphere storage container backed by a supported storage type. It holds VM files, while the ESXi host supplies compute resources such as CPU and memory. Storage vMotion changes the storage location; it does not primarily change the host running the VM.

What Storage vMotion Moves

A virtual machine's VM home is the datastore location containing its configuration and associated files. Depending on the selected workflow and placement choices, Storage vMotion can move:

  • Virtual disk files, commonly represented by VMDK files.
  • VM configuration and metadata files.
  • Snapshot files and their dependent delta disks.
  • Virtual machine swap files.
  • Other related VM home files, logs, and auxiliary files where applicable.

A virtual disk is a VM disk device represented by one or more files. The wizard can often move all VM files together or place individual virtual disks on separate compatible datastores.

When to Use Storage vMotion

  • Move workloads away from a datastore scheduled for maintenance, replacement, retirement, or migration.
  • Balance capacity across datastores when one datastore is approaching its usable limit.
  • Move a VM or selected disks to a higher-performance or lower-cost storage tier.
  • Place disks on storage with a different protection policy or storage technology.
  • Change a virtual disk provisioning format during the move when the destination and VM support the conversion.

For example, an application VM can remain powered on while its files move from an aging shared datastore to a newly provisioned datastore. Storage vMotion is not primarily a host migration operation. Use vMotion when the goal is to change the ESXi host while retaining the storage location.

Requirements and Pre-Migration Checks

vCenter Server is the management platform that inventories and orchestrates vSphere resources. The relevant ESXi host and VM should be managed by vCenter Server before starting a Storage vMotion operation.

Storage vMotion Prerequisite Checklist

Check: vCenter management
Why it matters: The migration workflow is orchestrated through vCenter Server.
How to verify: Confirm the VM and its ESXi host appear in the vSphere inventory.

Check: Permissions
Why it matters: The account needs migration privileges and permission to allocate space on the destination datastore.
How to verify: Review the assigned role at the VM, host, cluster, and datastore levels.

Check: Licensing and edition
Why it matters: Storage vMotion availability depends on the deployed vSphere licensing and edition model.
How to verify: Check the license assigned to the relevant vCenter and hosts and confirm the feature is available.

Check: VM power state
Why it matters: A powered-on VM uses live Storage vMotion; a powered-off VM uses a cold migration workflow.
How to verify: Inspect the VM power state before opening the migration wizard.

Check: Destination visibility
Why it matters: The ESXi host running the VM must be able to access the destination datastore.
How to verify: Confirm the datastore is mounted and visible from the current host.

Check: Capacity headroom
Why it matters: The destination needs room for VM files, snapshots, metadata, overhead, and temporary migration activity.
How to verify: Compare free capacity with actual disk consumption and expected growth, not just nominal disk size.

Check: Compatibility and policy
Why it matters: Datastore type, storage policies, encryption, disk format, and VM configuration can restrict placement.
How to verify: Review compatibility checks and storage policy compliance.

Check: VM state and I/O
Why it matters: Snapshots, mounted media, attached devices, and heavy write activity can complicate or lengthen the move.
How to verify: Review snapshots, virtual devices, guest activity, and recent task or event messages.

Migration time depends on virtual disk size, the rate at which the guest changes disk blocks during the operation, datastore performance, and the available network or storage configuration. A running database VM with high write activity can take substantially longer than a lightly used file server of the same nominal size.

Storage vMotion Workflow in the vSphere Client

  1. Select the target virtual machine in the vSphere inventory.
  2. Open the Migrate action.
  3. Choose the migration type that changes storage or datastore placement. Exact labels vary by vSphere version and client interface.
  4. Select a destination datastore, datastore cluster, or other compatible storage resource.
  5. Choose a virtual disk format if the interface presents that option.
  6. Review per-disk placement options if you need to separate an operating system disk and a data disk.
  7. Use vSphere compatibility checks and placement recommendations, including recommendations from a datastore cluster or Storage DRS where configured.
  8. Review the migration summary carefully, including destination paths, policies, capacity, and disk formats.
  9. Submit the task.
  10. Monitor Recent Tasks, events, and the VM status until the operation completes.

PowerCLI alternative

PowerCLI can initiate a storage relocation with Move-VM. Validate syntax, authentication handling, storage policy parameters, and supported options against the installed PowerCLI and vSphere versions.

Move-VM -VM <VMName> -Datastore <DestinationDatastore>

Virtual Disk Provisioning Formats

Provisioning controls how a virtual disk consumes and initializes datastore capacity. Format conversion can be useful during a datastore move, but the choices shown depend on datastore capabilities, policies, VM compatibility, and environment configuration.

Virtual Disk Provisioning Format Comparison

Format: Thin-provisioned
Capacity allocation behavior: Datastore space is consumed as guest data is written, subject to file growth and metadata overhead.
Initialization behavior: Blocks are allocated as needed.
Typical consideration: Efficient initial capacity use, but free space and growth must be monitored carefully.

Format: Thick lazy zeroed
Capacity allocation behavior: The configured capacity is reserved initially.
Initialization behavior: Blocks are zeroed when they are first written.
Typical consideration: Predictable reservation with potentially more initialization work on first writes.

Format: Thick eager zeroed
Capacity allocation behavior: The configured capacity is reserved initially.
Initialization behavior: Disk blocks are initialized before the disk is used.
Typical consideration: Up-front creation time and space reservation; may be required by particular workloads or features.

For a capacity-constrained datastore, converting a disk to thin provisioning can reduce initial consumption, but it does not remove the need to plan for the disk's possible growth. For a workload requiring reserved capacity, selecting a thick format may be appropriate; confirm that the destination has enough space for the reservation and any other VM files.

Snapshots, Devices, and File Placement

A snapshot records a point-in-time VM disk state and creates dependent delta files. Snapshot chains can increase the amount of data and metadata involved in a migration. Unnecessary snapshots should generally be removed and consolidated before a move, using an operationally safe procedure and confirming that consolidation completes.

Review mounted CD or DVD media, attached devices, raw mappings, encryption settings, and other VM configuration restrictions before migration. Storage policies are sets of placement and capability requirements assigned to VM objects. The destination must satisfy those requirements, or the migration may be blocked or leave the VM noncompliant.

When supported, place the operating system disk and data disks on different datastores. This can distribute performance or protection requirements, but it also means that the VM's files will not all appear in one location. A datastore cluster is a group of datastores managed as a storage resource pool, often with Storage DRS; its recommendations should be reviewed rather than accepted without checking capacity and policy requirements.

Operational Considerations

  • Storage vMotion generates storage reads and writes and may generate network traffic. Monitor datastore latency, host resources, storage paths, and guest I/O.
  • For high-I/O production systems, schedule the move during a lower-demand period or apply the environment's available migration and resource controls.
  • Confirm that encryption requirements are preserved. Encrypted VMs and disks can have additional compatibility or policy constraints.
  • Shared datastores are commonly visible to multiple hosts. Local datastores may be visible only to particular hosts, which can restrict later vMotion or failover operations.
  • Do not assume that nominal virtual disk size equals actual source capacity consumption, especially for thin disks and snapshot chains.

Verification After Migration

Post-Migration Validation Checklist

Validation item: Task completion
Expected result: The migration task reports success without unresolved warnings.
Where to check: Recent Tasks, events, and task details.

Validation item: Datastore assignment
Expected result: The VM home and intended virtual disks are on the expected destination datastore or datastores.
Where to check: VM summary, storage settings, and the datastore browser.

Validation item: Power and guest health
Expected result: The VM remains powered on and VMware Tools or equivalent guest services remain healthy where applicable.
Where to check: VM status, guest operating system, and monitoring system.

Validation item: Disk availability
Expected result: All expected disks are online and accessible to the guest.
Where to check: Guest operating system and VM hardware settings.

Validation item: Application function
Expected result: Production services respond normally and application latency is acceptable.
Where to check: Application checks, logs, and performance metrics.

Validation item: Source capacity
Expected result: Capacity is reclaimed only after every relevant VM file has moved and old files are no longer referenced.
Where to check: Source datastore browser and capacity view.

If the VM still appears associated with the original datastore, inspect each virtual disk and the VM home location. Only selected disks may have moved, or VM files may have been intentionally distributed across datastores. Refresh inventory views before drawing conclusions.

Relationship to Other Migration Types

Migration type: vMotion
VM power state: Powered on
Changes host: Yes
Changes datastore: Normally no
Primary use case: Move compute execution between ESXi hosts while retaining storage placement.

Migration type: Storage vMotion
VM power state: Powered on
Changes host: Normally no
Changes datastore: Yes
Primary use case: Relocate VM storage while the guest remains online.

Migration type: Enhanced or shared-nothing migration
VM power state: Often powered on when supported
Changes host: Yes or optionally yes
Changes datastore: Yes or optionally yes
Primary use case: Change both compute and storage placement in one broader migration scenario.

Migration type: Cold migration
VM power state: Powered off
Changes host: Yes or optionally yes
Changes datastore: Yes or optionally yes
Primary use case: Move a VM when live migration is unavailable or not appropriate.

Use enhanced vMotion concepts when a supported workflow must change both host and storage. A cold migration is different from live migration because the guest is powered off for the operation.

Troubleshooting Storage vMotion

Destination datastore is unavailable

The current ESXi host may not have access to the datastore, the datastore may conflict with the VM's policy or encryption state, or the account may lack datastore allocation privileges. Verify storage presentation and connectivity, compatibility results, storage policies, encryption constraints, and permissions.

Compatibility check fails

Read the specific compatibility message. Common causes include insufficient free space, unsupported disks or devices, policy conflicts, snapshots, encryption constraints, and datastore capability limitations. Correct the listed constraint or select another compatible destination.

Migration progresses slowly

Large disks, high guest write rates, congested storage paths, datastore latency, and snapshot chains can all increase migration time. Review host, datastore, and guest I/O metrics, reduce competing activity where safe, assess snapshot state, and schedule the operation during a lower-demand window.

Destination runs out of space

Capacity planning may have omitted snapshots, metadata, overhead, disk provisioning choices, or concurrent guest growth. Thin-provisioned source disks can have significant actual consumption. Provide more headroom, choose another datastore, reassess the format, and include all VM files in the calculation.

Source datastore still contains VM files

Only selected disks may have moved, or the VM home and disks may have been intentionally placed separately. Inspect storage settings and datastore file listings for every disk, snapshot, swap file, and VM home file. Refresh the inventory view and do not delete source files until the VM no longer references them.

Practical Migration Examples

  • Datastore maintenance: Move a running application VM from an aging shared datastore to a new datastore, then verify that the source no longer contains referenced VM files.
  • Capacity reclamation: Relocate a VM from an almost-full datastore to one with sufficient free capacity. Select thin provisioning only when it matches the organization's capacity-management policy.
  • Storage tier change: Move database disks to a higher-performance datastore, keep the guest available, and validate application latency after completion.
  • Disk format conversion: Select a thick format during the move for a workload requiring reserved capacity, then confirm destination consumption matches expectations.
  • Separate VM files: Place the operating system disk and data disk on different compatible datastores when the design and wizard options support per-disk placement.

Related Learning

Review browsing VMFS datastores for file-location checks, consolidating snapshots before a move, and assigning permissions for migration access. For storage policy and automated placement planning, study DRS automation levels and the related datastore-cluster features in your environment.