Question

我对 docker 非常陌生，我一直遇到“权限被拒绝”的问题。我能够使用以下 docker 文件构建图像：

FROM tensorflow/tensorflow:latest

ENV DEBIAN_FRONTEND=noninteractive

# Python

RUN apt-get update -y && apt-get upgrade -y &&\
    apt-get install python-opencv -y &&\
    pip install -U pip setuptools

RUN pip install keras jupyter && \
    pip install ipdb pytest pytest-cov python-coveralls coverage==3.7.1 pytest-xdist pep8 pytest-pep8 pydot_ng jupyter && \
    pip install Pillow scikit-learn notebook matplotlib nose pyyaml six h5py pandas scikit-image python-resize-image glob2 && \
    pip install opencv-python && \
    pip install --upgrade tensorflow-gpu tensorflow-probability

# COPY ST2inJupyter.js ~/.jupyter/custom/custom.js #this does not work.  put explicit abs path on image (abs path for files in directories outside Dockerfile do not work currently)
# Set up our notebook config.

COPY jupyter_notebook_config.py /root/.jupyter/
ADD startup /src/startup

ENV PYTHONPATH='/src/:$PYTHONPATH'

WORKDIR /src

#Configure Jupyter notebook port exposure
EXPOSE 8888

# For CUDA profiling, TensorFlow requires CUPTI.
ENV LD_LIBRARY_PATH \usr\local\cuda\extras\CUPTI\lib64:$LD_LIBRARY_PATH

# TensorBoard (only needed for visualizing lower level TF models.  in keras, not really too helpful, but just keep in)
EXPOSE 6006

# Define startup bash script in external file (loaded via ADD or COPY above)
CMD /src/startup

当我运行 docker 镜像：docker run -v ${PWD}:/src/hostpwd -it -p 8888:8888 [image_name] 时，我一直收到“权限被拒绝”错误。这是它的样子：

/bin/sh: 1: /src/startup: Permission denied

谁能帮我解决这个问题？我有一个 Mac 操作系统，我有新 M1 芯片的 docker 预览版。

Answer 1

来自评论，工作代码如下

FROM tensorflow/tensorflow:latest

ENV DEBIAN_FRONTEND=noninteractive

# Python

RUN apt-get update -y && apt-get upgrade -y &&\
    apt-get install python-opencv -y &&\
    pip install -U pip setuptools

RUN pip install keras jupyter && \
    pip install ipdb pytest pytest-cov python-coveralls coverage==3.7.1 pytest-xdist pep8 pytest-pep8 pydot_ng jupyter && \
    pip install Pillow scikit-learn notebook matplotlib nose pyyaml six h5py pandas scikit-image python-resize-image glob2 && \
    pip install opencv-python && \
    pip install --upgrade tensorflow-gpu tensorflow-probability

# COPY ST2inJupyter.js ~/.jupyter/custom/custom.js #this does not work.  put explicit abs path on image (abs path for files in directories outside Dockerfile do not work currently)
# Set up our notebook config.

COPY jupyter_notebook_config.py /root/.jupyter/
ADD startup /src/startup

ENV PYTHONPATH='/src/:$PYTHONPATH'

WORKDIR /src

#Configure Jupyter notebook port exposure
EXPOSE 8888

# For CUDA profiling, TensorFlow requires CUPTI.
ENV LD_LIBRARY_PATH /usr/local/cuda/extras/CUPTI/lib64:$LD_LIBRARY_PATH

# TensorBoard (only needed for visualizing lower level TF models.  in keras, not really too helpful, but just keep in)
EXPOSE 6006

# Define startup bash script in external file (loaded via ADD or COPY above)
CMD /src/startup

（转自 chepner）

权限被拒绝 docker Mac Os M1

1 个答案: