isaac-sim / IsaacGymEnvs

Isaac Gym Reinforcement Learning Environments
Other
2.05k stars 429 forks source link

How to run the IsaacGym simulation environment in an Ubuntu container and display the visual training interface? #199

Open Lihongjie-123 opened 10 months ago

Lihongjie-123 commented 10 months ago

Hello, sir. I would like to consult an issue regarding running the IsaacGym simulation environment in a Docker container. My goal is to run the IsaacGym simulation environment in a Docker container and display the visual training results through a VNC desktop during training.

The environment I am using is as follows:

  1. Ubuntu 20.04
  2. Docker 20.10.8
  3. GPU: Tesla T4 15360MiB
  4. GPU Driver Version: 535.129.03
  5. CUDA Version: 12.2
  6. VNC: TigerVNC

When executing the joint_monkey.py example, I encountered the following error: root@bdd460d9bfa1:~/isaacgym/python/examples# python joint_monkey.py Importing module 'gym_38' (/root/isaacgym/python/isaacgym/_bindings/linux-x86_64/gym_38.so) Setting GYM_USD_PLUG_INFO_PATH to /root/isaacgym/python/isaacgym/_bindings/linux-x86_64/usd/plugInfo.json pygame 2.5.2 (SDL 2.28.2, Python 3.8.10) Hello from the pygame community. https://www.pygame.org/contribute.html WARNING: Forcing CPU pipeline. Not connected to PVD +++ Using GPU PhysX Physics Engine: PhysX Physics Device: cuda:0 GPU Pipeline: disabled WARNING: lavapipe is not a conformant vulkan implementation, testing use only. Creating 36 environments XIO: fatal IO error 11 (Resource temporarily unavailable) on X server ":1" after 385 requests (385 known processed) with 0 events remaining.

Tracking the code, I found that the error occurs when the program reaches the following line: gym.draw_viewer(viewer, sim, True)

Could you please provide insights into the reason for this error? Any advice or solutions would be greatly appreciated!

Thank you very much!

Lihongjie-123 commented 10 months ago

I am running the 'joint_monkey.py' example code on a physical machine, and the displayed visualization interface looks like this: image

Lihongjie-123 commented 10 months ago

However, when running the 'joint_monkey.py' example code in a container, the displayed visualization interface looks like this, with the interface being very blurry, and the characters on top are not visible. 4e1b313c133af8d46339e0d4dd1ce89

Lihongjie-123 commented 10 months ago

Why is this happening? I am using a Tesla T4 GPU, and even after updating the GPU driver to version 510.108.03, I still encounter the same issue. The problem persists with GPU driver version 535.129.03 as well. Can you provide some advice or guidance?

Odin233 commented 1 month ago

Same problem. Have you solved this problem? ;_;

zhichenghe2008 commented 3 days ago

is there any issue if using --headless mode to run isaacgym in docker container env?

zhichenghe2008 commented 3 days ago

I am also try to use isaacgym in docker conainer env , but when I run a simple isaacgym demo as flowws: python from isaacgym import gymapi from isaacgym import gymutil gym = gymapi.acquire_gym() args = gymutil.parse_arguments(description="Example") sim_params = gymapi.SimParams() sim_params.up_axis = gymapi.UP_AXIS_Z sim_params.dt = 1.0 / 60.0 sim = gym.create_sim(0, 0, gymapi.SIM_PHYSX, sim_params)

there still have issue as fllow: Segmentation fault (core dumped)