Docker

Cacao’s Robot contains 2 processor unit that is Intel NUC and a Jetson Xavier NX developer kit. Two processors have different objectives first Jetson Xavier NX developer kit using for running the machine learning models and Intel NUC using for Navigation. Both processors do not have the same python environment, therefore, Cacao’s team is building the docker images for executing software in the same environment as Cacao’s processor

Jetson Xavier Docker Image

docker repository: https://hub.docker.com/repository/docker/thanachart29/jetson_athome

Jetson Xavier package

you can find all package at https://github.com/MBSE-2022-1/Software-Team/blob/docker/jetson_python_library.txt

requirement package included:

  1. System

    • OS: ubuntu focal fossa 20.04.5 LTS

    • Architecture: arm64/v8

    • CUDA: 11.4

    • Python 3.8.10

    • Default work directory : /usr/app/src/cacao_ws

  2. Library

    • ROS2-foxy-desktop

    • PyTorch v1.11.0

    • torchvision v0.12.0

    • torchaudio v0.11.0

    • TensorFlow 2.9.1

    • Protobuf 3.19.4

    • Numpy 1.23.1

    • OpenCV 4.6.0.66

    • PyYAML 5.3.1

    • Pillow 9.2.0

    • Request 2.28.1

    • Tqdm 4.64.1

    • Tensorboard 2.9.1

    • Pandas 1.5.1

    • Mediapipe 0.8.11

    • SpeechRecognition 3.8.1

    • pocketsphinx 0.1.15

    • Pyaudio 0.2.11

    • Pyttsx3 0.29

    • Pybind11 2.10.0


Enabling Jetson Containers on an x86 workstation (using qemu) [1]

$ sudo apt-get install qemu binfmt-support qemu-user-static
# Check if the entries look good.
$ sudo cat /proc/sys/fs/binfmt_misc/status
>> enabled
# See if /usr/bin/qemu-aarch64-static exists as one of the interpreters.
$ cat /proc/sys/fs/binfmt_misc/qemu-aarch64
>> enabled
>> interpreter /usr/bin/qemu-aarch64-static
>> flags: OCF
>> offset 0
>> magic 7f454c460201010000000000000000000200b700
>> mask ffffffffffffff00fffffffffffffffffeffffff

Note

Unfortunately you won’t be able to run any binary that calls into the NVIDIA driver on the x86 host.


Inter NUC Docker Image

docker repository: https://hub.docker.com/repository/docker/thanachart29/nuc_athome

Inter NUC package

you can find all package at https://github.com/MBSE-2022-1/Software-Team/blob/docker/nuc_python_library.txt

requirement package included:

  1. System

    • OS: ubuntu focal fossa 20.04.5 LTS

    • Architecture: x86_64

    • CUDA: 11.2

    • Python 3.8.10

    • Default work directory : /usr/app/src/cacao_ws

  2. Library

    • ROS2-foxy-desktop

    • PyTorch v1.11.0

    • torchvision v0.12.0

    • torchaudio v0.11.0

    • TensorFlow 2.9.1

    • Protobuf 3.19.6

    • Numpy 1.23.1

    • OpenCV 4.6.0.66

    • PyYAML 5.3.1

    • Pillow 9.2.0

    • Request 2.28.1

    • Tqdm 4.64.1

    • Tensorboard 2.9.1

    • Pandas 1.5.1

    • Mediapipe 0.8.11

    • SpeechRecognition 3.8.1

    • pocketsphinx 0.1.15

    • Pyaudio 0.2.11

    • Pyttsx3 0.29

    • Pybind11 2.10.0


Using Docker Instruction

  • first of all, install Docker following the Docker install manuals

  • Pull docker images

    • Jetson Xavier Images

      docker pull thanachart29/jetson_athome:latest

    • Intel NUC Images

      docker pull thanachart29/nuc_athome:latest

  • Run a command in a new containers

    docker run [OPTIONS] IMAGE [COMMAND] [ARG...]

    • Enable display from host: [OPTIONS] = –privileged –env DISPLAY=unix$DISPLAY -v /tmp/.X11-unix:/tmp/.X11-unix

    • Enable GPU from host: [OPTIONS] = –gpus all

    • Enable audio from host: [OPTIONS] = –group-add audio

    For example

    docker run --rm -it --privileged \
                    --env DISPLAY=unix$DISPLAY \
                    -v /tmp/.X11-unix:/tmp/.X11-unix \
                    –gpus all –group-add audio <image name>

  • Run with bash command

    docker run --rm -it <image name>

    Note

    using –rm option for remove the container when it exits

  • Execute scripts from local hosts to run inside container

    docker run -i <image name> <python3, e.g.> < <path to scripts>

    OR

    docker run -itd <image name>
docker exec -i <container ID> <python3, e.g.> <  <path to scripts>

  • Basically, run ros package in docker

    docker run -itd <image name>
docker cp <host file/directory> <container ID>:<container directory>
docker exec -itd <container ID> /bin/bash

    Note

    This command will copy your directory to container and you can use terminal as superuser


Using Nvidia-container runtime (for using gpu on docker) [2]

  1. Setting up docker

    curl https://get.docker.com | sh \
    && sudo systemctl --now enable docker

  2. Setting up NVIDIA Container Toolkit

    distribution=$(. /etc/os-release;echo $ID$VERSION_ID) \
    && curl -fsSL https://nvidia.github.io/libnvidia-container/gpgkey | sudo gpg --dearmor -o /usr/share/keyrings/nvidia-container-toolkit-keyring.gpg \
    && curl -s -L https://nvidia.github.io/libnvidia-container/$distribution/libnvidia-container.list | \
    sed 's#deb https://#deb [signed-by=/usr/share/keyrings/nvidia-container-toolkit-keyring.gpg] https://#g' | \
    sudo tee /etc/apt/sources.list.d/nvidia-container-toolkit.list

  3. Install the nvidia-docker2

    sudo apt-get update
sudo apt-get install -y nvidia-docker2

  4. Restart the Docker daemon after .. docker_rationale.rst .. jetson_docker.rst .. nuc_docker.rst .. docker_instruction.rst .. docker_reference.rst setting the default runtime:

    sudo systemctl restart docker

  5. (optional) tested by running a base CUDA container

    sudo docker run --rm --gpus all nvidia/cuda:11.0.3-base-ubuntu20.04 nvidia-smi

Reference

[1] : https://github.com/NVIDIA/nvidia-docker/wiki/NVIDIA-Container-Runtime-on-Jetson

[2] : https://docs.nvidia.com/datacenter/cloud-native/container-toolkit/install-guide.html