Hi,
Our vSphere hosts were upgraded to include the newest host driver version 14.0, and this apparently caused instability issues on vGPU enabled Linux guests running on earlier 13-series guest drivers. VMs started to crashing intermittently and this was something we had not seen before. Naturally the first thought remedy was upgrading the Linux drivers as well. However when applying Linux upgrade from the .deb package we got into a situation where Xorg doesn’t start anymore. Here is a relevant snippet from a Xorg.log. The Xorg session is started by VMware Horizon Agent 2111.1 so I cannot tell whether this is a VMware issue or Nvidia issue and which party to contact but anyhow it began with the driver upgrade and Horizon Agent reinstallation on top of it didn’t help. The GPUs are P100s, hosts are ESXi 7.0U3c and tested distributions are Ubuntu LTS 18.04 and 20.04.
242280.950] Current Operating System: Linux vdi-.. 5.4.0-100-generic #113~18.04.1-Ubuntu SMP Mon Feb 7 15:02:59 UTC 2022 x86_64
..clip
[242280.950] xorg-server-hwe-18.04 2:1.20.8-2ubuntu2.2~18.04.6 (For technical support please see http://www.ubuntu.com/support)
..
[242280.950] (++) Using config file: "/usr/lib/vmware/viewagent/resources/X11/xorg.conf"
..
[242280.960] (II) LoadModule: "nvidia"
[242280.960] (II) Loading /usr/lib/xorg/modules/drivers/nvidia_drv.so
[242280.961] (II) Module nvidia: vendor="NVIDIA Corporation"
[242280.961] compiled for 1.6.99.901, module version = 1.0.0
[242280.961] Module class: X.Org Video Driver
[242280.961] (II) NVIDIA dlloader X Driver 510.47.03 Mon Jan 24 23:02:31 UTC 2022
[242280.961] (II) NVIDIA Unified Driver for all Supported NVIDIA GPUs
..
[242280.966] (II) NVIDIA GLX Module 510.47.03 Mon Jan 24 22:57:16 UTC 2022
[242280.966] (II) NVIDIA: The X server supports PRIME Render Offload.
[242280.967] (--) NVIDIA(0): Valid display device(s) on GPU-0 at PCI:2:2:0
[242280.967] (--) NVIDIA(0): DFP-0
[242280.967] (--) NVIDIA(0): DFP-1
[242280.967] (--) NVIDIA(0): DFP-2
[242280.967] (--) NVIDIA(0): DFP-3
[242280.967] (**) NVIDIA(0): Using ConnectedMonitor string "DFP-0, DFP-1, DFP-2, DFP-3".
[242280.968] (II) NVIDIA(0): NVIDIA GPU GRID P100-4Q (GP100GL-A) at PCI:2:2:0 (GPU-0)
[242280.968] (--) NVIDIA(0): Memory: 4194304 kBytes
..
[242280.969] (II) NVIDIA(0): Validated MetaModes:
[242280.969] (II) NVIDIA(0): "DVI-D-0:640x480"
[242280.969] (II) NVIDIA(0): Virtual screen size determined to be 640 x 480
[242280.969] (--) NVIDIA(0): DPI set to (25, 30); computed from "UseEdidDpi" X config
[242280.969] (--) NVIDIA(0): option
[242280.969] (II) NVIDIA: Reserving 24576.00 MB of virtual memory for indirect memory
[242280.969] (II) NVIDIA: access.
[242280.974] (EE) NVIDIA(0): Failed to find a suitable copy object
[242280.974] (EE) NVIDIA(0): *** Aborting ***
[242280.974] (EE) NVIDIA(0): Failed to allocate 2D objects
[242280.974] (EE) NVIDIA(0): *** Aborting ***
[242280.976] (EE)
Fatal server error:
[242280.976] (EE) AddScreen/ScreenInit failed for driver 0