VMs configured with mks 3d GL renderer fail to power on
Article ID: 414725
Updated On:
Products
Issue/Introduction
- VMs with NVIDIA vGPU cannot start and it fails to power 'On' the virtual machine with error "Disconnected from virtual machine. Remote connection failure. Failed to establish transport connection."
- Under vmware.log, you see below backtrace when powerON event fails
2025-09-29T01:36:07Z[+0.000] In(05) mks - Backtrace:
2025-09-29T01:36:07Z[+0.000] In(05) mks - Backtrace[0] 000000245efc3d50 rip=000000241ad8ae04 rbx=000000245efc4fb8 rbp=000000245efc3f40 r12=000000000000000b r13=0000000000000000 r14=000000241b2d4520 r15=000000245efc4fb8
2025-09-29T01:36:07Z[+0.000] In(05) mks - Backtrace[1] 000000245efc3f50 rip=000000241ad8b661 rbx=000000241b955aa8 rbp=000000245efc3f90 r12=000000245efc3fa0 r13=000000000000000b r14=000000245efc5c98 r15=000000000000000b
2025-09-29T01:36:07Z[+0.000] In(05) mks - Backtrace[2] 000000245efc3fa0 rip=00000000002d3002 rbx=000000241bf4e870 rbp=000000245efc4ea0 r12=000000245efc4020 r13=000000241b8a74e0 r14=000000241b638f40 r15=000000245efc5088
2025-09-29T01:36:07Z[+0.000] In(05) mks - SymBacktrace[0] 000000245efc3d50 rip=000000241ad8ae04 in function (null) in object /bin/vmx loaded at 000000241a258000
2025-09-29T01:36:07Z[+0.000] In(05) mks - SymBacktrace[1] 000000245efc3f50 rip=000000241ad8b661 in function (null) in object /bin/vmx loaded at 000000241a258000
2025-09-29T01:36:07Z[+0.000] In(05) mks - SymBacktrace[2] 000000245efc3fa0 rip=00000000002d3002
2025-09-29T01:36:07Z[+0.000] Cr(01) mks - PANIC: Unexpected signal: 11.
2025-09-29T01:36:07Z[+0.316] Wa(03) mks - A core file is available in "/var/core/vmx-zdump.000"
2025-09-29T01:36:07Z[+0.316] In(05) mks - Backtrace:
2025-09-29T01:36:07Z[+0.316] In(05) mks - Backtrace[0] 000000245efc3840 rip=000000241a4ef0f0 rbx=0000000000000000 rbp=000000245efc3d40 r12=000000241b9924a8 r13=000000245efc3860 r14=000000241b2d4520 r15=000000245efc4fb8
2025-09-29T01:36:07Z[+0.316] In(05) mks - Backtrace[1] 000000245efc3d50 rip=000000241ad8ad04 rbx=000000245efc4fb8 rbp=000000245efc3f40 r12=000000000000000b r13=000000245efc3e70 r14=000000241b2d4520 r15=000000245efc4fb8
2025-09-29T01:36:07Z[+0.316] In(05) mks - Backtrace[2] 000000245efc3f50 rip=000000241ad8b661 rbx=000000241b955aa8 rbp=000000245efc3f90 r12=000000245efc3fa0 r13=000000000000000b r14=000000245efc5c98 r15=000000000000000b
2025-09-29T01:36:07Z[+0.316] In(05) mks - Backtrace[3] 000000245efc3fa0 rip=00000000002d3002 rbx=000000241bf4e870 rbp=000000245efc4ea0 r12=000000245efc4020 r13=000000241b8a74e0 r14=000000241b638f40 r15=000000245efc5088
- /var/run/log/Xorg.log in ESXi shows:
4075 2025-09-27T14:15:57.797Z In(14) Xorg[2103326]: unix:3 (EE) NVIDIA(GPU-0): Failed to initialize the NVIDIA GPU at PCI:xx:0:0. Please
4076 2025-09-27T14:15:57.797Z In(14) Xorg[2103326]: unix:3 (EE) NVIDIA(GPU-0): check your system's kernel log for additional error
4077 2025-09-27T14:15:57.797Z In(14) Xorg[2103326]: unix:3 (EE) NVIDIA(GPU-0): messages and refer to Chapter 3: Common Problems in the
4078 2025-09-27T14:15:57.797Z In(14) Xorg[2103326]: unix:3 (EE) NVIDIA(GPU-0): README for additional information.
4079 2025-09-27T14:15:57.797Z In(14) Xorg[2103326]: unix:3 (EE) NVIDIA(GPU-0): Failed to initialize the NVIDIA graphics device!Environment
vSphere 8.x
Cause
VM has vSGA 3D graphics enabled in addition to the vGPU graphics that may be contributing an issue with NVIDIA's Xorg servicing leading to RmInitAdapter failure.
Resolution
For vSGA-configured systems, the overall software stack's functionality relies directly on the proper operation of all GPU devices and their drivers. The above noted failure, occurring during device initialization, points to a hardware-level problem that requires investigation by the hardware vendor, NVIDIA.
Workaround:
Option 1:
To disable vSGA 3D, uncheck the "3D" graphics checkbox in the VM properties, or ensure that the "mks.enable3d=TRUE" line is removed from the VMX file.
Option 2:
Another workaround that can be tried is to enable Passthrough for the GPU that's crashing.
Put the Host into Maintenance Mode.
From the Host Client
1) Navigate to Host -> Manage -> Hardware -> PCI Devices
2) Find the impacted device Address
3) Click the checkbox to the left of the device and click "TOGGLE PASSTHROUGH"
From the vSphere Client
1) Navigate to the Host -> Configure -> Hardware -> PCI Devices
2) Click "ALL PCI DEVICES" and find the problematic ID 0000:xx:00.0
3) Click the checkbox to the left of the device and click "TOGGLE PASSTHROUGH"
Feedback
