In VMware vSphere environment, NFS datastore may report 100% utilization in vCenter even when the sum of virtual machine files appears significantly lower. This discrepancy often leads to "No space left on device" errors and VM instability.
2026-07-01T19:12:22.1562 In (14) vobd [2097767] :[vmfsCorrelator] 4062499818842us: [vob.vmfs.nfs.volume.no.space] NFS_Datastore1: No space on NFS volume.2026-07-01T19:12:22.157Z In (14) vobd[2097767] :[vmfsCorrelator] 4062526723580us: [esx.problem.vmfs.nfs.volume.no. space] NFS_Datastore1: No space on NFS volume.2026-07-01T19:12:22.157Z In (14) vobd [2097767] :The event ([esx.problem.vmfs.nfs.volume.no.space]NFS_Datastore1: No space on NFS volume.) was sent immediately to hostd;VMware ESXi 7.x
VMware ESXi 8.x
The issue occurs because ESXi does not calculate its own file system usage for NFS. Instead, it relies on the FSSTAT response from the NFS server.
If the backend storage array is consuming space for snapshots, replication metadata, or thick-provisioned allocations, the server reports these blocks as "used" to the ESXi host. vCenter categorizes this as "Other" because these blocks do not correspond to files registered in the vCenter inventory.
NFS datastores do not support host-side space reclamation via SCSI UNMAP.
To resolve the discrepancy:
Log in to the ESXi CLI and run df -h to confirm the physical consumption reported by the storage server.
Run du -sh /vmfs/volumes/<datastore_name> to verify the actual size of the VMDK files.
Coordinate with the storage vendor (e.g., Nutanix, NetApp, Dell) to identify and prune backend snapshots or active replication tasks.