vRO vMotion VM to SDRS Datastore cluster and inflate disk

The following can be used to vMotion a machine to an SDRS Datastore Cluster, inflating the disk to Lazy Zero at the same time (done by VcVirtualMachineRelocateTransformation.flat)

Create two input parameters. vMotionVM of type VC:VirtualMachine & a parameter name SDRSCL of type VC:StoragePod

The output parameter will be myVcTask of type VC:Task

var storageSpec = new VcStoragePlacementSpec();
var StorageResourceManager = vMotionVM.sdkConnection.storageResourceManager;
var myStoragePlacementResult = VcStoragePlacementResult();
var myKey = new Array();
var podSelectionSpec = new VcStorageDrsPodSelectionSpec();
var relocateSpec = new VcVirtualMachineRelocateSpec();
relocateSpec.transform = VcVirtualMachineRelocateTransformation.flat;

podSelectionSpec.storagePod = SDRSCL;

storageSpec.type = "relocate";
storageSpec.vm = vMotionVM;
storageSpec.podSelectionSpec = podSelectionSpec;
storageSpec.relocateSpec = relocateSpec

myStoragePlacementResult = StorageResourceManager.recommendDatastores(storageSpec);

myKey[0] = myStoragePlacementResult.recommendations[0].key
myVcTask = StorageResourceManager.applyStorageDrsRecommendation_Task(myKey);

Horizon 6.1 NVIDIA GRID vGPU

VMware Horizon 6.1 now supports NVIDIA GRID vGPU for their virtual desktop users, but before jumping on the rich graphic bandwagon there are some considerations to be made, for instance NVIDIA GRID VGPU doesnt support vMotion or HA because of the direct-path access to the hardware.

in addition to that there is no support for vGPU in a Linux guest as NVidia drivers are not availble


A handy comparison of vSGA, NVDIA GRID vGPU and vDGA is shown below


Before using the hardware-based (GPU) virtual Shared Graphics Acceleration (vSGA) feature in vSphere, you must install and configure the NVIDIA GPU VIB (vSphere Installation Bundle) on ESXi.

More information can be found here


I’m sure there will be some use cases for GRID, but I’d imagine the lack of vMotion & HA support will be a deal breaker for most.