Vmkfstools -d Jun 2026

To avoid errors, use absolute paths (e.g., /vmfs/volumes/Datastore01/VM01/VM01.vmdk ). Conclusion

For a database workload requiring high I/O throughput (IOPS), the latency introduced by Zeroed Thick or Thin provisioning during the "first write" phase can lead to application timeouts. vmkfstools -d

-d thin

vmkfstools -d eagerzeroedthick /vmfs/volumes/datastore1/myVM/myVM.vmdk To avoid errors, use absolute paths (e

Check your storage array's compatibility—some modern All-Flash Arrays (AFAs) handle thin provisioning at the hardware level, making the choice of -d format even more strategic. The power of vmkfstools -d comes with important warnings

The power of vmkfstools -d comes with important warnings. First, the operation is ; converting a 2TB thick disk to eagerzeroedthick can take hours, as every block must be explicitly zeroed. Second, during conversion, the original VMDK is locked, preventing the VM from powering on. Therefore, conversions are typically performed during maintenance windows. Third, -d works on block-level , not file-level; it cannot convert a raw device mapping (RDM) or a virtual disk that is part of a snapshot chain without collapsing the snapshots first. Finally, an often-overlooked limitation: vmkfstools -d does not shrink a disk's virtual size (the capacity the guest OS sees); it only changes how that capacity is allocated on the datastore.