Current Situation
I spent my first 2 years at Pure Storage (now Everpure) traveling the US and evangelizing the value and majesty of vVols as part of a virtualization and data services architecture. At the time, this was a brilliant marriage of capabilities and outcomes. Check out this blog post that profiled the WHAT and the WHY – https://tristantodd.com/?p=443 I thought the future was bright for vVols and couldn’t imagine a world where smart oganizations wouldn’t pivot to using vVols as a key service component in their virtualization stack.
HOWEVER – Broadcom officially deprecated vVols starting with VMware Cloud Foundation (VCF) 9.0, with plans for full removal in version 9.1. Consequently, NFS is now being recommended as a high-potential alternative for customers migrating off vVols, as it offers a similar balance of simplicity and scalability without the management complexities of LUNs.
So why did Broadcom “kill” vVols?
Well – this is primarily due to low industry adoption over the past decade, a desire to streamline the VMware Cloud Foundation (VCF) stack, and a strategic pivot to drive customers toward vSAN. Broadcom indicated vVols lacked a consistent operational model across hybrid cloud environments. Most Customers and Partners see this as part of a strategy to drive increased VCF and vSAN stickiness and lock-in. The benefits and beauty of SPBM (storage policy based management) but only if you use vSAN.
How and why should I move my vVols based virtual machines to NFS?
The primary reason to move your virtual machines (VMs) from vVols to NFS is the official deprecation of vVols announced by Broadcom, while the transition itself is primarily achieved through Storage vMotion.
Why You Should Move to NFS
The shift is largely driven by changes in VMware’s product roadmap and the maturation of NFS as a high-performance alternative:
vVols Deprecation and End of Life: Broadcom announced that vVols will be deprecated starting with VMware Cloud Foundation (VCF) 9.0 and fully removed in version 9.1. While support for vSphere 8.x continues until October 11, 2027, VCF 9.0 will not allow vVols as principal or supplemental storage, and Broadcom will not provide fixes for vVols bugs in that version.
- Performance Parity: Modern NFS has narrowed the technical gap with block storage. Advancements in VAAI-NAS allow NFS to offload intensive tasks like full file cloning and space reservation to the storage hardware, similar to how vVols operates.
- Maintaining VM Granularity: One of the main draws of vVols was VM-centric management. Modern NFS implementations can mimic this through features like Auto-Directory, which treats each VM as its own managed directory on the array, enabling per-VM statistics and monitoring directly on the storage backend.
- Operational Simplicity: NFS is often favored for its management simplicity, as it eliminates the need to manage individual LUNs, relying instead on a single large datastore.
How to Move to NFS
The migration process is straightforward but requires careful planning to maintain your existing data service levels:
Use Storage vMotion: The actual movement of data from a vVol datastore to an NFS datastore is performed using standard VMware Storage vMotion.
Align Storage Policies: Before moving, review your current Storage Policy-Based Management (SPBM) profiles. Because NFS typically manages services at the datastore level, you must ensure the destination NFS datastore has a snapshot, replication, and Quality of Service (QoS) profile that matches the VM’s original vVol policy.
Install Necessary Plugins: To maintain vVol-like performance and visibility, ensure the VAAI-NAS plugin is installed on your ESXi hosts. For Pure Storage users, the vSphere Plugin can be used to manage NFS datastores, monitor performance, and handle point-in-time recoveries.
Review Specialized Workflows: If you use vVols for database refresh operations (where individual vDisks are snapshotted and overwritten), you will need to update your scripts. Alternatives include mounting NFS snapshots as temporary datastores or using Raw Device Mappings (RDMs) if you require simpler backend-led refresh workflows for clustered databases.
Set Up Granular Monitoring: To replace the per-vDisk telemetry provided by vVols, you can leverage tools like Pure1 VM Analytics (which uses an OVA collector) to see metrics down to the individual virtual disk level on NFS.
When MUST you move off of vVols?
End of support for vVols is currently in the year 2030 when the introduction of VCF 9.3.