使用 --u=$(id -u $USER):$(id -g $USER) 运行时 Docker [Errno 13] Permission denied 错误

Question

I try to run a docker container with a tkinter gui app in python and x11 forwarding for a machine learning application. When I run the image with

sudo docker run --runtime=nvidia --gpus='all' --net=host --env="DISPLAY" --volume="$HOME/.Xauthority:/root/.Xauthority:rw" -i ganspace

I got the following display connection error, but the app seams to run

Downloading https://drive.google.com/uc?export=download&id=1FJRwzAkV-XWbxgTwxEmEACvuqF5DsBiV
Not cached
[19.07 19:59] Computing stylegan2-ffhq_style_ipca_c80_n1000000_w.npz
Reusing InstrumentedModel instance
Using W latent space
Feature shape: torch.Size([1, 512])
B=10000, N=1000000, dims=512, N/dims=1953.1
Sampling latents: 100%|██████████| 101/101 [00:08<00:00, 11.80it/s]
Fitting batches (NB=10000): 100%|##########| 100/100 [00:19<00:00,  5.11it/s]Authorization required, but no authorization protocol specified
Authorization required, but no authorization protocol specified

Total time: 0:00:28.239383
Loaded components for ffhq from /ganspace/cache/components/stylegan2-ffhq_style_ipca_c80_n1000000_w.npz
Traceback (most recent call last):
  File "interactive.py", line 645, in <module>
    setup_ui()
  File "interactive.py", line 214, in setup_ui
    root = tk.Tk()
  File "/opt/conda/envs/ganspace/lib/python3.7/tkinter/__init__.py", line 2023, in __init__
    self.tk = _tkinter.create(screenName, baseName, className, interactive, wantobjects, useTk, sync, use)
_tkinter.TclError: couldn't connect to display ":1"

If I run the image in the following way with

sudo docker run -u=$(id -u $USER):$(id -g $USER) -e DISPLAY=$DISPLAY -v /tmp/.X11-unix:/tmp/.X11-unix:rw --rm -it --runtime=nvidia --gpus='all' ganspace

the display connection is working (I tested it with another tkinter image), but I got the following permission error due to the user I think

Traceback (most recent call last):
  File "interactive.py", line 644, in <module>
    setup_model()
  File "interactive.py", line 143, in setup_model
    inst = get_instrumented_model(model_name, class_name, layer_name, torch.device('cuda'), use_w=args.use_w)
  File "/opt/conda/envs/ganspace/lib/python3.7/functools.py", line 840, in wrapper
    return dispatch(args[0].__class__)(*args, **kw)
  File "/ganspace/models/wrappers.py", line 695, in get_instrumented_model
    model = get_model(name, output_class, device, **kwargs)
  File "/opt/conda/envs/ganspace/lib/python3.7/functools.py", line 840, in wrapper
    return dispatch(args[0].__class__)(*args, **kw)
  File "/ganspace/models/wrappers.py", line 680, in get_model
    model = StyleGAN2(device, class_name=output_class)
  File "/ganspace/models/wrappers.py", line 125, in __init__
    self.load_model()
  File "/ganspace/models/wrappers.py", line 160, in load_model
    os.makedirs(checkpoint.parent, exist_ok=True)
  File "/opt/conda/envs/ganspace/lib/python3.7/os.py", line 213, in makedirs
    makedirs(head, exist_ok=exist_ok)
  File "/opt/conda/envs/ganspace/lib/python3.7/os.py", line 223, in makedirs
    mkdir(name, mode)
PermissionError: [Errno 13] Permission denied: '/ganspace/models/checkpoints'
ERROR conda.cli.main_run:execute(49): `conda run python interactive.py --model=StyleGAN2 --class=ffhq --layer=style --use_w -n=1_000_000 -b=10_000` failed. (See above for error)

Any ideas how to fix this or another solution to have the right permission and get the connection to the display correctly? thanks in advance!

Answer 1

It says permission issue because current user is not having the permission for

/ganspace/models/checkpoints directory

Add the below line in your Dockerfile to give permission for that path

 RUN chmod -R 777 /ganspace/models/checkpoints

Note: 777 gives full permission. change it accordingly based on what all permission you require. check permissions in linux