What Is VMFS in VMware: Architecture, Features, and VMFS File System Explained

VMFS (Virtual Machine File System) is VMware's clustered filesystem for storing virtual machine files on block storage. Multiple ESXi hosts can mount a shared VMFS datastore, and filesystem locks coordinate access to its files.

A VMFS datastore holds VMDK virtual disks, VMX configuration files, snapshot files, and logs. Its storage can be local to one host or presented to several hosts through a storage network.

This overview explains how VMFS organizes VM data, how shared access works, and what to check when a datastore becomes inaccessible.

What is VMFS in VMware?

VMware VMFS is a clustered filesystem that ESXi uses to store and manage virtual machine files.
Several ESXi hosts can mount the same VMFS datastore when they can access its underlying block storage.
A datastore contains the VMDK disks, VMX configurations, snapshots, and logs required by its virtual machines.
Shared VMFS storage can support vMotion, HA, and clustered deployments. These VMware features can also use other supported storage types.

How VMFS works: architecture and structure

How the VMFS shared filesystem coordinates hosts

Multiple ESXi hosts can read and write files on a shared datastore. Shared access supports clusters without requiring a separate copy of each VM on every host.
VMFS distributed locking coordinates file access so hosts do not make conflicting changes to the same files.
Locking and filesystem metadata help maintain consistency during concurrent access. They cannot prevent every type of corruption or protect against failed storage hardware.

What files are stored on VMFS?

VMFS stores each virtual machine's components alongside filesystem metadata:

VMDK files describe or contain a virtual disk that holds guest operating system, application, and user data.
VMX configuration files record virtual hardware settings, including CPUs, memory, network adapters, and storage controllers.
Snapshot files include disk changes and snapshot metadata, such as VMSD files. A snapshot depends on the rest of its disk chain.
Log files record VM activity and errors for troubleshooting.
VMFS metadata tracks datastore organization, allocated space, and file locations.

Keeping these files together lets administrators manage a VM's configuration and disks during normal operation, migration, backup, and recovery.

Block allocation and datastore capacity

VMFS allocates storage in blocks and uses sub-block allocation for small files. Allocation details depend on the VMFS version.
Thin-provisioned VMDKs allocate datastore space as data is written instead of reserving their full virtual capacity when created.
Sub-block allocation reduces unused space in blocks occupied by small files.
VMFS-6 supports datastores up to 64 TB. Expanding a datastore must remain within the supported volume limit and storage configuration.

VMFS features and benefits

The clustered filesystem lets several ESXi hosts use a common datastore, with locking to coordinate access.
VMFS supports large datastores and can grow through supported expansion operations, up to the volume capacity limit.
Supported configurations allow virtual disks up to 62 TB. Check the ESXi release and virtual hardware requirements before creating very large disks.
File locks coordinate access between hosts. Locking protects consistency during normal operation but does not replace backups.
Thin provisioning and sub-block allocation help use datastore capacity efficiently, but administrators still need to monitor free space.

Feature	VMFS Capability
Multi-host access	Yes
Max file size	Multi-terabyte support
Thin provisioning	Supported
Snapshot handling	Native support

VMFS versions

VMFS-5 and VMFS-6 compared

VMFS-6 adds automatic space reclamation through UNMAP on supported thin-provisioned storage. Reclamation depends on the capabilities and settings of the storage array.
VMFS-6 changes metadata and allocation behavior. Both VMFS-5 and VMFS-6 support a maximum datastore capacity of 64 TB.
Choose a VMFS version supported by your ESXi release and storage hardware. VMFS is used on block storage; vSAN and NFS use different datastore formats.

VMFS-6 retains shared access and distributed locking while adding features such as automatic reclamation. The table below provides a broad comparison; use the configuration limits for your ESXi release when planning capacity.

Feature	VMFS-5	VMFS-6
Space reclamation	Manual	Automatic
Max disk size	Up to 62 TB (ESXi 5.5 or later)	Up to 62 TB
Performance	High	Optimized

VMFS and general-purpose filesystems

Differences in shared access

VMFS manages virtual machine files on ESXi datastores.

It stores virtual disks, VM configurations, snapshots, and logs in a filesystem designed for ESXi.
Multiple hosts can access a shared datastore, supporting VM movement and restart across hosts.
Distributed locking and metadata coordinate access to files used by those hosts.

NTFS and ext4 normally have one host managing a mounted filesystem.

A single operating system ordinarily controls filesystem updates on an NTFS or ext4 volume.
Writing to the same block volume from several independent hosts requires a suitable cluster storage design.
Those filesystems can store virtual machine images, but they do not provide VMFS's ESXi datastore management and multi-host locking model.

The VMFS filesystem combines VM file storage with coordinated multi-host access. A general-purpose filesystem can still serve as VM storage through a supported file server or another storage layer.

Feature	VMFS	NTFS / EXT4
Multi-host access	Yes	No
Virtual machine optimization	Yes	No
Cluster support	Native	External tools required

VMFS datastores in ESXi clusters

For vMotion with shared VMFS storage, the source and destination hosts can access the VM's files on the same datastore. Other vMotion storage workflows are also available.
With HA configured, another host can restart a failed host's VMs when their files remain accessible. Restarting a VM still causes an interruption.
A shared datastore gives administrators a common location for VM files, snapshots, and logs.
DRS can move VMs between suitable ESXi hosts using vMotion. Shared VMFS storage is one way to support that movement.

Common reasons a VMFS datastore becomes inaccessible

Corrupted metadata or storage inconsistencies can prevent ESXi from locating files or mounting the datastore.
Disk, controller, path, or SAN failures can interrupt access to a VMFS volume. RAID redundancy does not cover every failure.
Deleting a VM or its VMDK files removes data needed to run it. Restore from a separate backup when available; otherwise, recovery depends on what remains on storage.
Missing or damaged files in a snapshot chain can prevent a virtual disk from opening or a VM from starting.
Damage to allocation metadata can make files inaccessible even when some data blocks remain readable.

A failed datastore can stop several VMs at once. Keep independent backups and investigate the storage fault before attempting recovery.

Why VMFS data recovery requires specialized tools

Recovery software must understand VMFS metadata to locate files and reconstruct their layout. Ordinary NTFS or ext4 recovery tools do not interpret it.
File-lock problems can also prevent access. Diagnose a stale lock or an active owner before treating the issue as lost data.
VM disks, configuration files, and snapshot metadata can depend on one another. The files that need recovery depend on the failure and the intended restore.

Typical VMFS recovery work

Recover VMDK virtual disks that contain the guest operating system and application data.
Recover VMX configuration files or rebuild a VM definition to attach usable virtual disks.
Scan damaged metadata to locate file data and reconstruct the datastore's file layout where possible.

Recovering individual files may be enough to extract data. Starting the original VM may also require a usable configuration and a complete virtual disk or snapshot chain.

Recovering virtual machines from VMFS datastores

Using DiskInternals VMFS Recovery™

DiskInternals VMFS Recovery™ provides tools to examine inaccessible VMFS storage and recover readable VM data:

Scan damaged or formatted VMFS volumes to locate files and surviving datastore structures.
Search for deleted or damaged VMDK virtual disks and recover data that has not been overwritten.
Recover available VMX configuration files to help rebuild VM definitions.
Extract files from virtual machines whose disks cannot be opened in the original VMware environment.
Use recovered files as part of a tested disaster recovery process.

Recovery results depend on the damage and the data still present. Save recovered files to separate storage so the scan does not overwrite the source.

Ready to get your data back?

To start recovering your data, documents, databases, images, videos, and other files, press the FREE DOWNLOAD button below to get the latest version of DiskInternals VMFS Recovery® and begin the step-by-step recovery process. You can preview all recovered files absolutely for FREE. To check the current prices, please press the Get Prices button. If you need any assistance, please feel free to contact Technical Support. The team is here to help you get your data back!

FREE DOWNLOADVer 4.28, Win BUY NOWFrom $699

Managing VMFS storage

Choose RAID and storage redundancy for the workload. RAID 10 and RAID 6 involve different capacity, performance, and failure-tolerance tradeoffs.
Monitor latency, throughput, available space, and storage errors with vSphere and storage-system tools.
Monitor snapshot growth and consolidate or remove snapshots when they are no longer needed. Confirm that enough space is available for the operation.
Back up VMs regularly to separate storage and test restores. Snapshots on the same datastore do not provide an independent backup.
Check redundancy across disks, controllers, and paths. Verify that multipathing and failover work for the actual storage configuration.

Capacity monitoring, tested backups, and working storage failover help keep VMFS datastores usable and reduce recovery time.

VMFS explained: what to remember

VMFS stores virtual machine files and coordinates access from ESXi hosts.
A shared VMFS datastore is one supported storage option for vMotion, HA, and DRS; these features are not exclusive to VMFS.
Thin provisioning and block allocation help manage storage capacity within the limits of the VMFS and ESXi versions in use.
When storage fails, recovery tools must understand VMFS and the files required by each VM.

Before recovering a VMFS datastore, determine whether the problem is a storage path, a file lock, damaged metadata, or missing VM files. That diagnosis determines the next step.