Instant clones slow provisioning

[LukaszDziwisz]

Hello,

I was wondering if maybe someone came across similar issue and might offer an advise on how to troubleshoot it and resolve.  We are using Windows 1809 LTSC, Horizon 7.7, AppVolumes 2.16 and UEM 9.7 and vCenter 6.7 U2. So pretty much all the newest. Our storage is all flash based Pure, however it appears that the Instant clone takes roughly 5  minutes before it becomes available. It is in provisioning state for about 1 min 30 sec and then 3 mins 30 sec in customization state and then becomes available.

I opened a ticket with VMware on it and here is what we are seeing from the log:

2019-05-22 16:07:14 (3444) info CSvmgaService - (svgaservice.cpp, 397 ) shutting down the service

2019-05-22 16:11:54 (4268) debug Svmga:: setservicestatus - (svmga.cpp , 522) service stats set to 2

As you can see there is a 4  minute gap that would account for the long time before machine becomes available.

Next update from Vmware engineer is as follows:

Furthermore i discussed the case with my technical lead and he mentioned that it either could be some script or some program on VM itself which is causing it to reboot. As we are using clone prep so machine should not reboot however in your case its doing exactly the opposite.

The process C:\Program Files (x86)\Common Files\VMware\View Composer Guest Agent\vmware-svi-ga.exe (TESTIMRW5) has initiated the shutdown of computer TESTIMRW5 on behalf of user NT AUTHORITY\SYSTEM for the following reason: Operating System: Recovery (Planned)

Reason Code: 0x80020002

So the next suggestion is to build brand new image with plain windows, install vmtools and horizon agent only and provision pool. I did that and bam the instant clone only takes 30 seconds before it becomes available.

I don't really want to redo my images as it will be very time consuming and pretty sure will break other things so I was wondering if anyone might have any idea on how to resolve it?

[LukaszDziwisz]

lainalex​,

I'm fully aware of that problem and it has been resolved with 6.7 U3 which is the version I'm on now.

[elgwhoppo]

VMware R&D has concluded that with the current workflows, it takes longer to create a vGPU enabled Instant Clone, as compared to a regular Instant Clone. The delays that we are seeing here are expected and shouldn't be concerning.

I mean 5 minutes isn't anywhere near "instant" so it would seem like someone missed the mark when making vGPU supported GA feature for instant clones.

[Jubish-Jose]

We are also seeing this issue. The instant clone goes to through an additional reboot to plug in the GPU.

We are on ESX 6.7 P02 (16075168) with Horizon 7.9 and Windows 10 1903.

Is there any resolution to this? This is clearly a bug and instant clones with GPU is useless with this much delay.

[Ezrahite]

I will second your concern.

We are in a similar situation.

Has anyone found a solution to this issue?

[sjesse]

I do not work for VMware, but I don't consider this a bug exactly, just a new technology not quite there yet. 5 mins is not the 20-30 mins that linked clones was taking per vm. I'm completely guessing but based on what I've read about the issues with the vms freezng on during vmotions, I wonder if the instant clone process problems mentioned in

vGPUs and vMotion, why the long stun times? | Yellow Bricks

is similar here, where there isn't an efficient way to copy the frame buffer over and the extra reboot solves that problem. I suggest anyone who is effect by this to open up tickets if you haven't just to make sure that VMware is tracking the importance of this issue.

[Ezrahite]

Thank you for that artical sjesse​, it does make a lot of sense.  And I would agree with you that it is maybe not a bug, so much a limitation of how the process works.

I ran a small tested in our environment.

With vGPU instant clones - time to create a VM is about 2-5 minutes.

Without vGPU instant clones - time to create a VM is about 10-30 seconds.

I am still highly interested in hearing about future updates that allow a vGPU Instant clone to be created in that 10-30 second mark as we run a large number CAD programs and require the GPU.

[jordanht]

Was this issue resolved for anyone who might be following this thread? We are also experiencing the additional reboot of instant clones with nvidia GPU.

[Jubish-Jose]

From 7.13 onwards, they have something called Smart Provisioning and in the case of VMs with GPUs, there will be no cp-parent VMs. Desktops will be cloned directly from cp-replica VMs. However, a power on is required to attach the GPU, but there is no reboot.

[Viewhelpdesk]

.

[jordanht]

Thank you very much for this. We are at version 8 but the pools were not being created with Smart Provisioning. I read that Smart Provisioning defaults to off when a vGPU is used (VMware Horizon version 8 (2006) - Smart Provisioning - YouTube) so I manually changed the pool to use Smart Provisioning (How to switch provisioning scheme for an Instant Clone Pool or Farm (81026) (vmware.com)) and now the clones work without a parent VM! 

[JustSomeGuyDude]

I think you are on to something with vMotion and switching between physical GPUs. The DRS in my clone cluster is set to be fairly conservative to prevent clones from moving around for slight imbalances between the hosts.

The only other thing I could think of that caused problems for me on 1809 was the AutoEndTasks registry key. This is a good GPO to set to keep other applications from holding up the image on shutdown even without this cloning problem.

https://www.tenforums.com/tutorials/97821-turn-autoendtasks-restart-shut-down-sign-out-windows-10-a....