VcXsrv WSL X server closes unexpectedly. Do I need to disable GPU?

Question

I am trying to run some code with pybullet. I am on windows 10, have the latest vscode, and I am using WSL remote on vscode with ubuntu 18.04 LTS. I have a GTX 2070 graphics card. I just want to see this work, I've been trying to fix it for the last 3 hours.

First, here is the code I am trying to run in WSL:

import numpy as np
import pybullet as pb 
physicsClient = pb.connect (pb.GUI)
#load plane
import pybullet_data
pb.setAdditionalSearchPath(pybullet_data.getDataPath())
planeId = pb.loadURDF('plane.urdf')
#load visual shape
visualShapeId = pb.createVisualShape(
    shapeType=pb.GEOM_MESH,
    fileName='random_urdfs/000/000.obj',
    rgbaColor=None,
    meshScale=[0.1, 0.1, 0.1])
collisionShapeId = pb.createCollisionShape(
    shapeType=pb.GEOM_MESH,
    fileName='random_urdfs/000/000_coll.obj',
    meshScale=[0.1, 0.1, 0.1])
multiBodyId = pb.createMultiBody(
    baseMass=1.0,
    baseCollisionShapeIndex=collisionShapeId, 
    baseVisualShapeIndex=visualShapeId,
    basePosition=[0, 0, 1],
    baseOrientation=pb.getQuaternionFromEuler([0, 0, 0]))

I get no errors, but the X server window will pop up (black) and close immediately. I read that you need to disable your GPU with WSL, but I am scared of messing up my PC. I would only want to disable it for when I need to see graphics / use the X server, not for all WSL applications.

Here is what shows in my bash script:

user@DESKTOP-######:~/program$ python3 openAI.py
pybullet build time: Sep 22 2020 00:54:31
startThreads creating 1 threads.
starting thread 0
started thread 0 
argc=2
argv[0] = --unused
argv[1] = --start_demo_name=Physics Server
ExampleBrowserThreadFunc started
X11 functions dynamically loaded using dlopen/dlsym OK!
X11 functions dynamically loaded using dlopen/dlsym OK!
Creating context
Failed to create GL 3.3 context ... using old-style GLX context
Indirect GLX rendering context obtained
Making context current
GL_VENDOR=NVIDIA Corporation
GL_RENDERER=GeForce RTX 2070 SUPER/PCIe/SSE2
GL_VERSION=1.4 (4.6.0 NVIDIA 451.67)
GL_SHADING_LANGUAGE_VERSION=(null)
pthread_getconcurrency()=0
Version = 1.4 (4.6.0 NVIDIA 451.67)
Vendor = NVIDIA Corporation
Renderer = GeForce RTX 2070 SUPER/PCIe/SSE2
Segmentation fault (core dumped)
user@DESKTOP-######:~/program$

Answer 1

@Emilio, I have got this working without any changes to the GPU, using the following process:

I used the VcXsrv application set up in the same way as this tutorial : https://jack-kawell.com/2020/06/12/ros-wsl2/ where crucially Native openGL is unchecked.

Export your ip address as in the tutorial, however instead of 'export DISPLAY={your_ip_address}:0.0', go to the VcXsrv window (which should be blank at this point) and replace :0.0 with whatever number of display is given. So for Display DESKTOP-1234AB:1.0 you would enter 'export DISPLAY={your_ip_address}:1.0'

In the linux terminal enter: export LIBGL_ALWAYS_INDIRECT=0

You can check that this has made an effect by entering: glxinfo Which should print out: direct rendering: yes

When you run your python program it should open up in the VcXsrv window. For me there was no cursor visible but I could still interact with the object as if I did have a cursor.