↓Skip to main content
  1. Posts/

Docker Note

·8 mins·
Table of Contents

IMG-20250121235045541.webp

Installation and Configuration #

1. Install Docker #

Step 1: Update System Packages #

Ensure system packages are up to date:

sudo apt update
sudo apt upgrade -y

Step 2: Install Docker #

Run the following command to install Docker:

sudo apt install docker.io -y

Step 3 (Optional): Start and Enable Docker Service #

Start Docker and set it to start on boot:

sudo systemctl start docker
sudo systemctl enable docker

Step 4: Verify Docker Installation #

Run the following command to check Docker version:

docker --version

Output similar to Docker version xx.xx.xx indicates successful installation.

Configure Docker Proxy #

Create configuration file and add proxy settings:

  1. Create or edit Docker service configuration file:

    sudo mkdir -p /etc/systemd/system/docker.service.d 
sudo gedit /etc/systemd/system/docker.service.d/http-proxy.conf

  2. Add the following content to http-proxy.conf file:

    [Service] 
Environment="HTTP_PROXY=http://127.0.0.1:7890/" 
Environment="HTTPS_PROXY=http://127.0.0.1:7890/" 
Environment="NO_PROXY=localhost,127.0.0.1"

  3. Reload system daemon configuration and restart Docker service:

    sudo systemctl daemon-reload 
sudo systemctl restart docker

Add User Group #

Avoid needing sudo every time you use docker.

Post-installation steps | Docker Docs

Username seems to require lowercase.

alexavier@alexavier:~$ sudo usermod -aG docker Alexavier
usermod: user 'Alexavier' does not exist
alexavier@alexavier:~$ sudo usermod -aG docker alexavier

After configuration, it seems you need to restart/re-login for VSCode plugins to work.

Common Operations #

Pull Image #

docker pull osrf/ros:noetic-desktop-full

View Local Images #

docker images

View Running Containers #

docker ps

View All Containers (Including Stopped) #

docker ps -a

Start Container #

If multiple containers share the same image, the read-only part of the image will be shared, not occupying duplicated space.

docker run -it my_noetic:7_26 /bin/bash

docker run -it --network host my_noetic:7_26 /bin/bash

docker run -it --network host -v ~/SumResearch_2024/Repo/Host/ROS1_ws:/root/ROS1_ws my_noetic:latest /bin/bash

docker run -it --network host -v ~/SumResearch_2024/Repo/Host/ROS1_ws:/root/ROS1_ws -v ~/SumResearch_2024/Repo/RDKx5:/root/RDKx5 arm_noetic:latest

docker run -it --network host -v ~/SumResearch_2024/Repo/Host/ROS1_ws:/root/ROS1_ws -v ~/SumResearch_2024/Repo/RDKx5:/root/RDKx5 --privileged --device /dev/ttyUSB0 arm_noetic:latest

Start New Shell Session #

Find the running container ID or name, suppose it is my_noetic_container.

docker exec -it my_noetic_container /bin/bash

Stop Container #

docker kill <container_name_or_id>

Save Container as Image #

docker commit youthful_khayyam airbot_img:latest

View Container Space Usage #

docker system df

View Image Info #

docker inspect b56b435576e8

Local Save and Load Docker Images #

Save a Docker image as a file, then load this file elsewhere.

Save Image #

Use docker save command to save the image as a tar file:

docker save -o my-ubuntu-image.tar my-ubuntu-image
Load Image #
docker load -i my-ubuntu-image.tar

Delete Image and Container #

  • Delete Container:

    docker rm <container_id_or_name>

  • Delete Image:

    docker rmi <image_id_or_repository:tag>

File Operations #

Copy to Container #

docker cp ~/SumResearch_2024/Repo/Host/ROS1_ws youthful_khayyam:/home/

GUI Configuration #

Host #

Install x11

sudo apt update
sudo apt install xserver-xorg xauth

Verify function

xeyes

Grant permissions

xhost +local:

Container #

Recommend using compose to open container, specific content can refer to later sections.

Install x11

apt update
apt install -y x11-apps

Test

xeyes

Dockerfile #

File Example #

Dockerfile is a text file containing a series of instructions to define how to build a Docker image. Suppose you want to create a Docker image containing a basic Ubuntu system, you can create a simple Dockerfile:

# Use official Ubuntu image as base image
FROM ubuntu:20.04

# Maintainer info
LABEL maintainer="your_email@example.com"

# Update package list and install some basic tools
RUN apt-get update && apt-get install -y \
    curl \
    wget \
    vim \
    git \
    && apt-get clean

# Set working directory
WORKDIR /root

# Copy your system configuration file (Optional)
# COPY my_config_file /root/my_config_file

# Install other software or do other configurations (As needed)
# RUN ...

# Default command
CMD ["bash"]

You can add more instructions in Dockerfile according to your needs, such as installing other software, copying files, etc.

Build Image using Docker #

In the directory where Dockerfile is located, open terminal and run the following command to build Docker image:

docker build -t my-ubuntu-image .

Where my-ubuntu-image is the name you specify for the image, and . indicates Dockerfile is in the current directory.

 

Docker Compose #

Docker Compose uses a YAML file to define and manage multiple containers, thereby simplifying command line operations.

Old Version Content (Skip for ubuntu 24.04 and later) #

Installation #

Docker Compose Official Docs

Docker Compose | Runoob Tutorial

Usage #

Create a docker-compose.yml file in the project root directory and add the following content:

version: '3.8'

services:
  arm_noetic:
    image: arm_noetic:latest
    container_name: arm_noetic_container
    network_mode: host
    volumes:
      - ~/SumResearch_2024/Repo/Host/ROS1_ws:/root/ROS1_ws
      - ~/SumResearch_2024/Repo/RDKx5:/root/RDKx5
    privileged: true
    devices:
      - "/dev/ttyUSB0:/dev/ttyUSB0"
    stdin_open: true
    tty: true

In this file:

  • version specifies the Docker Compose file version.
  • services section defines a service (i.e., container), here named arm_noetic.
  • image specifies the Docker image to use.
  • container_name specifies a name for the container.
  • network_mode set to host, same as --network host parameter.
  • volumes defines volumes to mount.
  • privileged set to true, giving the container privileges.
  • devices section specifies devices to share.
  • stdin_open and tty set to true to enable interactive mode.

In the directory where docker-compose.yml is located, use the following command to start the container in background mode:

docker-compose up -d

Installation #

When using 24.04, docker-compose cannot be used because docker-compose 1.x version depends on distutils, and distutils was removed in Python 3.12. The new version of docker-compose has switched to Go implementation (docker-compose-plugin) and no longer depends on Python. So use the new version ‘docker compose plugin’

sudo apt remove docker-compose
sudo apt update 
sudo apt install -y ca-certificates curl gnupg lsb-release

# Add Docker official GPG key
sudo install -m 0755 -d /etc/apt/keyrings
curl -fsSL https://download.docker.com/linux/ubuntu/gpg | sudo gpg --dearmor -o /etc/apt/keyrings/docker.gpg 
sudo chmod a+r /etc/apt/keyrings/docker.gpg 

# Add Docker software source
echo "deb [arch=$(dpkg --print-architecture) signed-by=/etc/apt/keyrings/docker.gpg] https://download.docker.com/linux/ubuntu $(lsb_release -cs) stable" | sudo tee /etc/apt/sources.list.d/docker.list > /dev/null 

# Update & Verify
sudo apt update
sudo apt install docker-compose-plugin
docker compose version

The new version command has no hyphen in the middle, it is docker compose, not docker-compose.

Correspondingly, the specific content of compose seems to need slight modification (filename unchanged).

version: '3.8'

services:
	graspnet:
		image: nvidia/cuda:12.1.0-devel-ubuntu22.04 # Use the image you downloaded
		container_name: graspnet-container
		network_mode: host
		environment:
			- NVIDIA_VISIBLE_DEVICES=all # Show all GPUs (Can specify GPU)
			- NVIDIA_DRIVER_CAPABILITIES=all
			- DISPLAY=${DISPLAY} # Support visualization
			- QT_X11_NO_MITSHM=1 # Prevent Qt error
			- HTTP_PROXY=http://127.0.0.1:7890 # Clash Proxy Port
			- HTTPS_PROXY=http://127.0.0.1:7890
		volumes:
			- ~/graspnet:/workspace # Mount local ~/graspnet to container /workspace
			- /media/user/disk/dataset/graspnet:/workspace/dataset
			- /tmp/.X11-unix:/tmp/.X11-unix # Share X11 Unix socket, support GUI
		privileged: true
		working_dir: /workspace # Default working directory inside container
		stdin_open: true # Support interactive session
		tty: true # Enable terminal
		deploy:
			resources:
				reservations:
					devices:
						- capabilities: [gpu]

Commands #

The new version commands also have slight changes, missing the hyphen. Execute in the directory where docker-compose.yml is located:

docker compose up -d

Note, this method will delete existing containers with the same name and recreate the container using the image. If you just want to start a container that already exists but is in stopped state, use the following command:

docker compose start

# Check if container is alive:
docker compose ps
# View real-time logs:
docker compose logs -f
# Stop service:
docker compose down

Use Graphics Card in Container #

To support GPU, NVIDIA Container Toolkit needs to be installed.

Set NVIDIA Docker Repository #

Add NVIDIA’s GPG key and repository:

distribution=$(. /etc/os-release;echo $ID$VERSION_ID) \
    && curl -s -L https://nvidia.github.io/nvidia-docker/gpgkey | sudo apt-key add - \
    && curl -s -L https://nvidia.github.io/nvidia-docker/$distribution/nvidia-docker.list | sudo tee /etc/apt/sources.list.d/nvidia-docker.list

Install NVIDIA Container Toolkit #

Update package list and install:

sudo apt update
sudo apt install -y nvidia-docker2
# If download is too slow, can also use proxy
sudo apt -o Acquire::http::Proxy="http://127.0.0.1:7890" \
         -o Acquire::https::Proxy="http://127.0.0.1:7890" install -y nvidia-docker2

Restart Docker Service #

sudo systemctl restart docker

Verify NVIDIA Docker Availability #

Run the following command to check if GPU can be accessed by Docker:

docker run --rm --gpus all nvidia/cuda:11.8-base nvidia-smi

# Monitor GPU usage
sudo apt install gpustat
gpustat --watch

If GPU info is displayed, installation is successful.

Ubuntu 24.04 Version #

Currently (2024.12.15) NVIDIA Container Toolkit does not support ubuntu 24.04 yet, but it can be downloaded, refer to:

https://www.server-world.info/en/note?os=Ubuntu_24.04&p=nvidia&f=3

Summary:

# Update dependencies and system
sudo apt update
sudo apt install -y curl apt-transport-https ca-certificates software-properties-common

# Add NVIDIA GPG public key
curl -fsSL https://nvidia.github.io/libnvidia-container/gpgkey | sudo gpg --dearmor -o /usr/share/keyrings/nvidia-toolkit.gpg

# Add NVIDIA container toolkit source
curl -fsSL https://nvidia.github.io/libnvidia-container/stable/deb/nvidia-container-toolkit.list | \
sudo tee /etc/apt/sources.list.d/nvidia-toolkit.list
sudo sed -i -e "s/^deb/deb [signed-by=\/usr\/share\/keyrings\/nvidia-toolkit.gpg]/g" /etc/apt/sources.list.d/nvidia-toolkit.list

# Update package index and install toolkit
sudo apt update
sudo apt install -y nvidia-container-toolkit

# Restart Docker and test
sudo systemctl restart docker
sudo docker run --rm --gpus all nvidia/cuda:12.3.0-runtime-ubuntu22.04 nvidia-smi

Pull CUDA 11.8 Image #

https://hub.docker.com/r/nvidia/cuda/tags

Select suitable image and run:

sudo docker pull nvidia/cuda:11.8.0-base-ubuntu22.04

Run Container #

Start container using CUDA image and verify if CUDA is available:

docker run --rm --gpus all -it nvidia/cuda:11.8-base-ubuntu22.04 bash

Or use compose

version: '3.8'

services:
  graspnet:
    image: pytorch/pytorch:2.2.0-cuda11.8-cudnn8-devel # Use the image you downloaded
    container_name: graspnet-container
    network_mode: host
    runtime: nvidia # Enable GPU support (Need to install NVIDIA Container Toolkit)
    environment:
      - NVIDIA_VISIBLE_DEVICES=all # Show all GPUs (Can specify GPU)
      - NVIDIA_DRIVER_CAPABILITIES=all
      - DISPLAY=${DISPLAY}         # Support visualization
      - QT_X11_NO_MITSHM=1         # Prevent Qt error
      - HTTP_PROXY=http://127.0.0.1:7890  # Clash Proxy Port
      - HTTPS_PROXY=http://127.0.0.1:7890
    volumes:
      - ~/graspnet:/workspace      # Mount local ~/graspnet to container /workspace
      - /media/alexavier/8018420DD8B69548/dataset/graspnet:/workspace/dataset 
      - /tmp/.X11-unix:/tmp/.X11-unix # Share X11 Unix socket, support GUI
    privileged: true  
    working_dir: /workspace        # Default working directory inside container
    stdin_open: true               # Support interactive session
    tty: true                      # Enable terminal

Verify Environment: #

Run the following command in container:

nvidia-smi

Since nvcc is a tool only available in devel version, nvcc -V cannot be used. If you don’t need to develop cuda programs and only use it when running programs, you can just use base.

Other Issues #

Git Permission Issues Caused by Shared Volumes (Possible) #

Something like this

❯ git pull
error: unable to create file RDKx5/hand/servo_limit.json: Permission denied

Ensure user owns the ownership of the entire Git repository files and directories

sudo chown -R $USER:$USER ~/SumResearch_2024/Repo

Container Configure Host Proxy #

Long-term configuration can be added in compose

export HTTP_PROXY="http://127.0.0.1:7890"
export HTTPS_PROXY="http://127.0.0.1:7890"
# test
curl -I https://www.google.com

Vscode Development #

VSCode Remote - Containers Plugin #

Modify Shared Directory Permissions #

Unable to write file “/home/semitia/SumResearch_2024/Repo/Host/ROS1_ws/src/hand_control/src/hand_ctrl.py”(NoPermissions (FileSystemError): Error: EACCES: permission denied, open ‘/home/semitia/SumResearch_2024/Repo/Host/ROS1_ws/src/hand_control/src/hand_ctrl.py’)

Modify permissions of the mounted directory on the host machine to make it writable for all users:

chmod -R 777 ~/SumResearch_2024/Repo/Host/ROS1_ws