I'm seeing a new (to me) behavior. My PVS Target devices are thick provisioned with 30GB cache disks. VM Objects have 12 GB of ram. When the VMs boot, about 12GB of write cache appears to be unavailable, consumed by the amount of RAM the VM is set to. If a VM is shut down and the amount of RAM is changed, the size of the cache disk changes to reflect the new size of the RAM (If I drop the VM to having 8GB of ram, about 8GB of cache is unavailable.) Win10 and 11 vdisks are set to standard image, cache in device ram with overflow on hard disk, 1024MB.
I just built new images for Windows 11 desktops and Server 2022 App Servers, hypervisor is vmware, PVS is version 2203 LTSR CU3 (2203.0.17). Our Win10 VMs that have been around for years do not exhibit this behavior.
My page file is set like the Win10 boxes I'm trying to get rid of. The screen shot shows 12GB or ram, and the roughly 30GB that should be cache disk being reduced to 17GB.