kentar0kozai/omniverse-isaac-docker-proxy-gui

omniverse-isaac-docker-proxy-gui

This repository summarizes how to run omniverse-launcher and isaac-sim in a docker environment with GUI display.
It also supports proxy environments, so you can easily run isaac-sim on your company PC.
*Not tested in a WSL2 environment, so someone please try it.

Usage

1. Setting host PC

1. Grant access privileges to root.

xhost +local:root

2. Add to .bashrc.

echo 'xhost +local:root > nul >> ~/.bashrc

3. Kill processes.

fuser -k 33480/tcp
fuser -k 34080/tcp

4. Use the command line to log in to NGC before pulling the Isaac Sim container.

docker login nvcr.io

5. Clone this repository and change directory.

git clone https://github.com/kentar0kozai/omniverse-isaac-docker-proxy-gui.git
cd omniverse-isaac-docker-proxy-gui

6. Modify line 8 of the Dockerfile to your own proxy and port.

e.g.
ENV PROXY=***.***.***.***:**
￬
ENV PROXY=192.0.2.1:8080

7. Run docker compose.
   *Download and installation takes a long time.

docker compose up

8. Success if Container isaac-sim-enhanced Created / Attaching to isaac-sim-enhanced is displayed.　　

9. Run the command in another terminal.

docker exec -it isaac-sim-enhanced bash

2. How to execute Omniverse-launcher

1. Run the command the docker terminal.

ov-launcher

2. Create an NVIDIA account and log in.

3. Select the NUCLEUS tab and register the local server.

4. Click 'Connect to a Server' to connect to the localhost server.
   *Log in with ID: admin, Password: admin if you cannot log in.

5. Select the EXCHANGE tab and install 'Cache'.

6. Install Isaac-Sim.

7. Select the Library tab and run Isaac-Sim.

8. A select tab will appear and you can start with any setting.

9. After loading, try running one of the Isaac Examples.

10. Press “load” on the pop-up tab to load the environment.
    (The initial loading takes a while, so be patient; from the second time on, the environment is saved in Cache, which is faster.)

11. The isaac-sim directory also contains Python scripts and sample code, allowing you to start the Python tutorial right away.