This question goes out to all my Nvidia GPU users for VDIs. We have a weird new issue that has shown up since we have moved to VCenter 7.0.
In the past we are used to once we set a memory reservation (All Lock) and attached the NVidia Profile to the VM...VCenter will show that 100% memory is being utilized in VCenter. This was annoying because there were alerts that would have to be set to be ignored because VCenter would show 100% of the memory was being used on the VDI. My Nvidia Grid users will know what I’m talking about. Now, for some reason most....but not all of our VDIs show 75%. So 16GB shows 12GB active, 32GB shows 24GB active, etc, etc. Has anyone else experienced this? The other weird part is if I run a stress test on a VDI with lets says 32Gb of memory.....Windows will show all 32GB being used while VCenter stays at 75%. Not going above 24GB. Is this VCenter reporting back the wrong information or do we have something set wrong. We have had the same setup for years without any power on issues of the VDIs. All the hosts have plenty of memory to spare. Below are snips of what we are used to seeing compared to what we are seeing now
Accepted Solutions