Brief theoretic introduction can be found in Instructions.md

This repository contains all Dockerfiles which are used for software containerization that implies easier integration and development.

In order to install docker engine follow this instructions for installation on ubuntu.

You will also need nvidia-docker which can be installed by executing following commands:

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

sudo apt-get update

sudo apt-get install -y nvidia-docker2

In Figure 1. it is shown how to create docker container from Dockerfile

In Figure 2. it's possible to see how different commands for docker container are used:

Enter corresponding directory and run following command to build docker image:

docker build -t <image_name>:<tag_name> <dockerfile_path> 

For concrete example (ros-kinetic with gazebo) run following:

docker build -t ros_gazebo_img:melodic_11

In order to create container from Dockerfile run following commmand:

docker run -it --network host --privileged -v /dev:/dev --name <container_name> <img_name>:<tag_name> 

Run docker container with GUI support

docker run -it --network host --gpus all --privileged -e DISPLAY=$DISPLAY -v /dev:/dev -v /tmp/.X11-unix:/tmp/.X11-unix \
       --name <container_name> <img_name>:<tag_name> 

Also, before using param --gpus make sure you've installed nvidia-container-toolkit as follows:

sudo apt-get update
sudo apt-get install -y nvidia-container-toolkit

Make sure you have latest drivers for your GPU.

docker start -i <container_name> 

docker exec -it <container_name> bash 

Kill container:

docker kill  <container_name> 

Stop container:

docker  stop <container_name> 

Remove container:

docker rm <container_name> 

Attach existing detached cotainer:

docker attach <container_name> 

Clean dangling images and stopped containers (bear in mind that this deletes every container that's not currently running, not recommended):

docker system prune 

In order to run multiple docker containers which are build from docker images in certain configuration use docker-compose

To be able to properly use available GPU's it's necessary to follow instructions related to setting up docker container with proper arguments depending on GPU vendor. There are different commands and prerequisites (drivers) for almost every GPU vendor (NVIDIA,ATI,Intel). Rest of this instructions related to graphical rendering are written for NVIDIA GPUs, but here you can find everything else necessary to achieve same behaviour for different GPUs (Intel/ATI).

Explanation for using ROS + Docker for GUI applications is available here

Explanation for creating OpenGL image for CUDA applications can be found here.

Available prebuilt nvidia/openGL dockers can be found here

There's example of using build-args for installing IDE in container in following Dockerfile

docker build --build-arg <arg_name>=<arg_value> -t <image_name>:<image_tag> .

Info about CLion IDE setup can be found here

One of the main problems that occurs when using Docker files for robotic applications is expiration of GPG keys for ROS. More about this incident can be found here

In order to fix that, in case that some sudo apt-get update or sudo apt-get install fails, please replace GPG keys to newer ones. Error for expiration of GPG keys looks like following:

An error occurred during the signature verification. The repository is not updated and the previous index files will be used. GPG error: http://packages.ros.org/ros/ubuntu bionic InRelease: 
The following signatures were invalid: EXPKEYSIG F42ED6FBAB17C654 Open Robotics <[email protected]>

You can add keys by following 1.3 section in Installation instructions for corresponding ROS distribution.

If you already have docker image built, and you get following error message when trying to install new ROS package:

E: Failed to fetch http://packages.ros.org/ros/ubuntu/pool/main/r/ros-melodic-sound-play/ros-melodic-sound-play_0.3.11-1bionic.20210414.224641_amd64.deb  404  Not Found [IP: 140.211.166.134 80]
E: Unable to fetch some archives, maybe run apt-get update or try with --fix-missing?

You can fix it by typing following command:

curl -s https://raw.githubusercontent.com/ros/rosdistro/master/ros.asc | sudo apt-key add -

• Create example docker files with different hardware accelerations
• Create base images for ROS/Gazebo combinations (Kinetic/Gazebo9, Kinetic/Gazebo11, Melodic/Gazebo9, Melodic/Gazebo11)
• Refactor images to be multistage
• Add BUILD_ARGS -> ide for now
• Add image for decoupled NVIDIA + pytorch support for OpenPose
• Add entrypoint script for every Docker containing ROS (can be used from darknet ros)
• Build image with GAZEBO gym
• Build image with AirSim
• Build base blender image
• Add cloning of Github repo with ssh keys
• Create intial compose
• Check integration of Github actions for CI/CD