It would be a disservice to you, reader, for us to talk about containers and not mention Docker. Docker is not the only way to implement containers, but it is a popular container software; perhaps the most popular one. Docker has almost become synonymous with the term container. Docker, Inc., the product maker, follows a freemium model, offering both a free and a premium version. Docker was released to the public in 2013 at the PyCon Conference.
As container software, Docker can package an application with its libraries, configuration files, and dependencies. Docker can be installed in Linux environments as well as on Windows. The virtual containers that Docker enables allow applications to run in isolation without affecting any other processes running on the same physical machine.
Docker is often used by both developers and system administrators, making it an essential tool for many DevOps teams. Developers like using it because it allows them to focus on writing code without worrying about the system’s implementation and deployment details where it will eventually be deployed. They can be assured that the characteristics of the environment will be identical regardless of the physical machine. Developers can also leverage many of the programs and extensions that come bundled with Docker. System administrators use Docker frequently because it gives them flexibility and its light footprint allows them to reduce the number of servers needed to deploy applications at scale.
The complete Docker documentation, installation instructions, and a download link for the Community edition of Docker can be found here: https://docs.docker.com/ .
Docker does not have a monolithic architecture. It has a set of well-defined components, each in charge of an individual function and fully dedicated to performing only that function. The following architecture shows the major Docker components.
As shown in the preceding diagram, the Docker system operates on a client-server model where the Docker client communicates with the Docker daemon. The Docker daemon handles the complex tasks associated with building, running, and administering Docker containers. The daemon has the flexibility to run on the same host as the client or establish a connection with a remote host. The Docker client and daemon can run on a variety of OSes, including Windows and Linux. Let’s go through the Docker components in detail to increase our understanding of Docker.
Every Docker container needs to have a Dockerfile. A Dockerfile is a plain old text file containing instructions showing how the Docker image will be built. Don’t worry; we’ll cover Docker images in a second.
Some of the instructions that a Dockerfile will contain are the following:
Let’s now move on to the Docker image component.
After creating the Dockerfile, the next step is creating an image. The Docker build utility can take a Dockerfile and create an image. The Docker build utility’s purpose is to create ready-for-de- ployment container images. The Dockerfile contains instructions that specify how the Docker image will be built.
The Docker image is portable across environments and instance types, and that’s one of the reasons for Docker’s popularity. You can deploy the same image in a Linux or Windows environment, and Docker will handle the details to ensure that the deployment functions correctly in both environments. One recommended best practice is to ensure that any external dependencies specified in the Dockerfile have the version of the dependency explicitly declared. If this is not done, inconsistencies may result from the same Dockerfile because a different library version may be picked up.
Docker run is a utility where commands can be issued to launch containers. In this context, a container is an image instance. Containers are designed to be transient and temporary. The Docker run utility can restart, stop, or start containers. The utility can launch several instances of the same image, and those instances can run simultaneously to support additional traffic. For example, if you have ten similar instances taking traffic and the traffic increases, you can use the Docker run utility to launch an additional instance.
When you build a container, you can configure it from scratch, creating your own Dockerfile and configuring it yourself. However, many times, it is not necessary to reinvent the wheel. If you want to leverage the work that others have done already, you can use Docker Hub. Docker Hub is a collection of previously created containers shared by Docker users. In Docker Hub, you will find Docker images created by Docker and by other vendors who sometimes support those containers. Also, other Docker users publish versions of the containers they have created that they have found useful. You can also share your containers with the public if you choose to do so. However, if you choose, you can also upload containers into a local Docker registry, keep them private, and only share them with select groups and individual
Docker Engine is the heart of Docker. When someone says they are using Docker, it is shorthand for saying “Docker Engine.” Docker Engine instantiates and runs containers. The company offers two versions of Docker Engine: the open-source version, dubbed Docker Engine Community Edition, and Docker Engine Enterprise Edition.
Docker launched Docker Engine Enterprise Edition in 2017. However, as with many companies that use the freemium model, the original open-source version is still available and maintained. It is now called Docker Engine Community Edition. The Enterprise Edition has added advanced features, such as vulnerability monitoring, cluster management, and image management.
Docker Compose is another Docker tool that can be used to configure and instantiate multi-container Docker applications. In order to configure it, a YAML file is used. Once the YAML configuration is defined, the service can be started with one command. Some of the advantages of using Docker Compose are as follows:
A common use case for Docker Compose is setting up development, testing, and UAT environments on one host.
Docker Swarm groups VMs or physical machines that are running Docker Engine and are configured to run as a cluster. Once the machines have been clustered, you can run regular Docker commands, and those commands will be executed on the cluster rather than on individual services. The controller for a swarm is called the swarm manager. The individual instances in the cluster are referred to as nodes.
The process of managing nodes in a cluster in unison is called orchestration.
Operating instances as a cluster or a swarm increases application availability and reliability. Docker swarms consist of multiple worker nodes and at least one manager node. The worker nodes perform the application logic and handle the application traffic, and the manager oversees the management of the worker nodes, thus managing resources efficiently. Let’s look into the AWS-managed container service for hosting Docker in the cloud.