Docker Intro

Before Docker :

Previously if we want to run a Java application in the server, then in traditional way we need to install JDK in the server. Like JDK for java, each language has their own dependencies or libraries needs to be installed. If we want to deploy the same application in another server/VM, then again we need to install the respective libraries, which will become overhead and difficult to scale.

Also running different applications in a same server becomes overhead and leads to chaos on maintainability and scalability.

Docker :

Docker introduced lightweight, isolated environments (containers) within a server/VM, in which our application will be running respectively.

In a single server/VM, we can run multiple containers, if we have java, python applications, then each application can be deployed in separate container which will have their respective libraries, so in this case, dependencies of java and python won’t be installed directly in the host OS.

Docker Flow :

1. First we need to install the Docker application in the server/VM.
2. Need to create a DockerFile, which will have the details of,
   • Container’s OS and libraries
   • Working directory location
   • Our app code location
   • In which port app will run
   • App run command
3. Need to build this DockerFile which will create the Image.
4. Need to run the Image, which will creates and runs the container, which eventually starts our application.

Docker Architecture :

Docker Client :

Docker client is nothing but the host OS and via the below ways we can access the docker components like Images, Containers,

• Docker Desktop
  • It’s an User Interface through we can access the docker objects.
• Docker CLI
  • Through command line we can access docker objects.

Docker Host :

Docker Host is nothing but a Docker Engine which internally runs in Linux kernel. Though we have Windows/Mac OS as a host OS, using WSL/HyperKit we can run the docker, which helps to provide the Linux kernel environment internally.

• If the host OS is Windows, then WSL (Windows SubSytem for Linux) / Hyper-V is needed via which Docker will be running.

• If the host OS is Mac, then using HyperKit, docker will be running.

Docker Engine Components :

• Docker REST API
  • If it’s either Docker Desktop or Docker CLI, when a command executed, internally it will hit the Docker REST API, which will process and communicates the Docker Daemon (dockerd) and returns the response to the client.
• Docker Daemon
  • Docker Damon is the central process responsible for managing Docker Objects like containers, images, volumes and networks
  • Some of the tasks will be handled by dockerd itself where some will be delegated,
    • Tasks like pulling the Images or managing storage, dockerd handles it directly.
    • For container lifecycle, like create, start, stop of the container, dockerd will delegates to containerd which will further delates to runc.
• Docker Images
  • Docker Image is nothing but a dockerFile which contains the spec details of a container, like,
    • Container’s OS and libraries
    • Working directory location
    • Our app code location
    • In which port app will run
    • App run command
• Docker Containers
  • After creation of an Image, while running, it will create a container, which eventually starts our application.
  • We can run multiple containers, where each can have their respective libraries.
• Docker Volumes
  • This helps to store the container data like app logs etc.
  • Example like, if we are going to remove a container, then we can move the logs/files from that container to Docker Volumes where we can access post the deletion the container also.
• Docker Networks
  • While running multiple containers, this helps to communicate between them, those configurations can be maintained/handled here.

Registry :

Registry is something linked to external where Images can be distributed across networks, vice versa we can download the distributes Images from the registry.

This will be helpful in an organization, which avoids creating multiple Images with same spec, instead everyone can use the same distributed Image.

Docker Hub is one of the commonly used distributed registry, among the multiple registry available.