Getting Started with Docker

Conquering Docker!!!!!!!

References: [Book] Troubleshooting Docker

Container and Docker

• Docker is an Open source project receiving tremendous attention

• It provides functionality to package applications as lightweight containers for deployment and operation

• Docker Container provides functionality to deliver applications in a consistent and standardized form like containers used in actual cargo ships!

What is a Container?

VM vs Container

• Container is technology that encapsulates applications to provide independent operating environments

  • Can be used as an alternative to Virtual Machines!

• Container is built on Linux Container, which is a User Space Interface for Linux Kernel's isolation features!

  • By providing powerful APIs and simple tools, Linux users can create and manage applications as containers

• Container differs from traditional Hypervisor in that OS running on Host machine share Linux kernel!

  • Hypervisor means?

    • Software that creates and operates Virtual Machines

  • In other words, even when running multiple containers on the same machine, all use the same Kernel

    • Therefore, it's faster than VM and has almost no performance overhead!

OS Container

• OS Container is similar to VM

  • but, it differs in that it provides user space isolation functionality while sharing the Kernel of Host Machine OS

• Like VM, you can allocate dedicated resources to Container and install, configure, and operate applications and libraries

• OS Container is useful for scalability testing!

  • You can easily deploy multiple containers composed of various distributions, much lighter than VM

• Container is created through images or templates that define structure and contents

  • Therefore, when building development environments, it's easy to configure to have the same execution environment, version, and package environment as other Containers!

• Currently, various container technologies are available, among which suitable ones for OS Container include LXC, OpenVZ, Docker, BSD Jail, etc.

Application Container

• While OS Container consists of multiple processes, Application Container packages and operates one service

  • Application container operates only one process

    • The process that operates is the application process, contrasting with OS Container running multiple services on one OS!

• Docker container uses a layer approach, which can minimize duplication and increase reusability

  • First operate a container for base image commonly used by all components,

  • Then add desired elements as separate layers to the file system on top of it

• Using layer-based file system allows easy rollback to previous layers at any desired time!

• Every time a RUN command specified in Dockerfile executes, a new layer is created in the container

• The main purpose of Application Container is to make various components constituting the application into separate container-type packages

  • Application components bundled in container units operate applications through interaction via APIs or services

    • This distributed multi-component system deployment approach is core to implementing microservice architecture!

      • This way

        • Developers can package applications in their desired container form,

        • IT teams can take full responsibility for deploying containers to various platforms for horizontal or vertical scaling

OS Container vs Application Container

OS Container	Application Container
Operate multiple services in one OS Container	Operate one service in one Container
Does not use layered file system	Based on layered file system
ex) LXC, OpenVZ, BSD Jail	ex) Docker, Rocket

Learning Docker in Detail

• Docker combines various Linux Kernel features with file systems to compose images in module form

• Through composed images

  • You can freely configure application virtual environments,

  • Realize WORA (Write-Once-Run-Anywhere) principle

• Since applications can be simplified to the level of operating one process

• You can easily build distributed systems where multiple processes collaborate,

• And support high scalability

• Docker provides four important characteristics for application development

  1. Autonomy

  2. Decentralization

  3. Parallelism

  4. Isolation

• Docker containers can reproduce identical behavior anywhere - development machines, bare metal servers, virtual machines, data centers, etc.

  • From application designer's perspective, all operational matters can be left to DevOps team, allowing full concentration on development..!

    • This enables modularizing team workflow, increasing productivity and efficiency

Docker and VM

• Docker provides application-level isolation and security features for applications operating in Container by sharing OS

  • Despite providing strong isolation and security features through perfect abstraction of OS layer, Docker's resource usage is considerably less than VM, making performance and efficiency excellent!

Benefits of Docker

Using Docker Container in Microservice Architecture has the following advantages

• Fast application deployment

  • Containers contain only application-related parts, so they are small and can be quickly deployed with minimal runtime

• Portability

  • Application and operating environment (dependency information including libraries used by application) can be bundled into one Docker Container, and such containers are independent of OS version or deployment model

  • Docker Containers can be transmitted to machines capable of operating docker and executed immediately without worrying about compatibility issues

• Easy sharing

  • Pre-made container images can be uploaded to public repos or internal private repos for easy sharing

• Low resource usage

  • Docker images are small and use fewer resources when deploying new applications by utilizing other containers

• Reusability

  • It's easy to continue versioning Docker containers and easy to roll back to previous versions anytime

  • Since components in existing layers can be reused, tremendous lightweight is possible

Lifecycle of Docker Container

image-20200809005844880

1. Build

• Write all commands needed to package Container in Dockerfile and use it to build Docker image

  docker build

• To tag an image, use -t option

  docker build -t username/image name

• When Dockerfile exists in a different path than current directory, specify Dockerfile path with -f option

  docker build -t username/image name -f /path/Dockerfile

2. Run

• After creating image, execute docker run command to deploy Container

  docker run

• To check the status of running container, execute docker ps command

  docker ps

  • This command shows the list of currently active containers

• To pause all processes, execute docker pause command

  docker pause

  • This command uses cgroups freezer to pause all processes currently running in the container

    • Internally sends SIGSTOP signal

  • Using this command, you can pause and resume processes anytime

• To execute one or more containers in stop state, use docker start command

  docker start

3. Stop / Kill

• When done using Docker, stop or kill it

  docker stop