How do you configure Docker to use a different container runtime?

Below, I will explain step by step how to configure Docker to use different container runtimes, such as CRI-O or containerd, and how to apply these configurations in real-world scenarios.

Step 1: Install the Required Container Runtime

First, you need to install the container runtime you want to use on your system. For example, with CRI-O, you can install it using the following command (for Ubuntu):

bash
sudo apt-get update
sudo apt-get install -y software-properties-common
sudo add-apt-repository -y ppa:projectatomic/ppa
sudo apt-get update
sudo apt-get install -y cri-o-1.18

For containerd, the installation process may be as follows:

bash
sudo apt-get update
sudo apt-get install -y containerd

Step 2: Configure Docker to Use the New Container Runtime

After installation, you need to configure the Docker daemon to use the new container runtime. This typically involves modifying or creating the Docker configuration file daemon.json, which is usually located in the /etc/docker/ directory.

Example Configuration for containerd:

Edit the /etc/docker/daemon.json file and add the following content:

json
{
  "exec-opts": ["native.cgroupdriver=systemd"],
  "log-driver": "json-file",
  "storage-driver": "overlay2",
  "containerd": "/run/containerd/containerd.sock"
}

After saving and closing the file, restart the Docker service to apply these changes:

bash
sudo systemctl restart docker

Step 3: Verify the Configuration

After configuration, you can confirm that Docker is using the new container runtime by running test containers. You can also check the current runtime using the docker info command:

bash
docker info | grep Runtime

Real-World Example

In my previous work, we migrated the development environment from Docker's default runc runtime to containerd primarily due to containerd's superior resource control and security features. By following these steps, we successfully implemented this migration across several production environments without service interruptions. We also configured automation scripts to manage runtime settings for new virtual machines, ensuring consistent and predictable deployments.