Post Imaging Error - "Why Did My PC Restart"?

[dhc2pilot]

We're using VMWare Workstation Pro 16.x to build Windows 10 images.  For 2+ years we haven't had a problem imaging IDE, SSD, or NVME equipped laptops.  I'd say within the last 3 months I've recently started seeing more and more devices go into a continual reboot loop at first startup after imaging.  Windows 10 will display an error "Why did my PC Restart" and give me a "Next" button.  It doesn't matter if I press Next or just wait about 30 seconds, it tries to do something (Load a driver I presume), then reboots immediately and the loop starts again.  This seems to be happening to laptops which have an SSD or NVME drive in them.  I have not experienced it on devices with a standard mechanical drive in them.  Originally I thought that the laptops, due to them having SSD's or NVME drives, were running the SetupComplete.bat file much faster and not giving Windows enough time to load the NVME drivers.  I put a 90 second pause at the start of the batch file, and this resolved the issue for about 2 weeks.  Now the problem is back again.

I want to start eliminating possible causes - I'd like to know what drive type setting I should be using on my VM for a windows 10 image that I plan to deploy to both SSD and NVME equipped laptops.  I'm currently using NVME as the drive type in my VM.  We are using FOG to capture/deploy our images.  Our FOG server is hosted on Ubuntu 20.x.  We currently are deploying images onto Lenovo T470, T480, T490, T14, L14, E14, P15 series laptops.  (The T470 and T480's have been refitted with SSD drives.) 

Should I change the drive type of my VM to SSD, SCSI, or IDE?  I'm not worried about the VM's performance at all - I only care about the resulting imaged devices not going into a reboot loop at first run.  Any ideas or suggestions appreciated.

NOTE:  Latest drivers installed via Lenovo Vantage, no driver errors in HW Manager.  Sysprep went flawlessly.  I've noticed that every time the device restarts a new user profile is created.  Sometimes after 15-20 restarts the device will finally boot normally, but it didn't join the domain or get renamed successfully.

Thanks in advance!

[continuum]

just my 2 cents ... in my experience NVME-virtual disks are less stable than SCSI-virtual disks.
Upto 2012 or so we would even recommend virtual IDE disks for really heavy usage...

In other words - I would not use NVME virtual adapters inside a VM at this early stage.
Upto 2012 or so we would recommend IDE-disks for VMs with heavy use. Virtual NVME controllers are still in their infancy state.
None of the VMs I created with NVME controller was long term stable or ready for heavy usage.
Dont think that just because your host uses NVME your VMs should use NVME too.

[louyo]

FWIIW, another .02

PMJI: I have not used that imaging system, so I could be out in left field. I fail to see how the drive type of your Ubuntu VM could make any difference. The image is the same no matter where it is hosted and dispensed. I would be more inclined to blame drivers for different models or drives within the image. As far as domain connection, I would wonder about the script reusing machine names and MAC addresses. Perhaps there are log entries. 

Uli, I am hosting Windows, Linux and  ESXi systems on a 1 TB NVME drive. No "known" problems.

Perhaps dumb luck? Like the optimistic skydiver whose parachute doesn't open. As he passes 2,000 feet: "so far, so good"

Lou

[continuum]

> Uli, I am hosting Windows, Linux and ESXi systems on a 1 TB NVME drive. No "known" problems.

That then would be physical NVME drives - right ?
I only say that virtual NVME controllers are "still experimental"
Ulli

[louyo]

Ah, yes, my bad. Physical. Thank you for setting me straight. 

I jinxed myself though, this morning my system wouldn't boot, puked when trying to mount that 1 TB nvme drive. Had to edit fstab and then do an fsck to "fix" it. We shall see. 

Lou