Introduction to Containers & Diving a little deeper into the benefits of Containerisation
Download as PPTX, PDF0 likes822 views
The document outlines a comprehensive training agenda on container orchestration using Docker and Kubernetes, covering key topics such as container architecture, Docker architecture, and application modernization. It discusses the benefits and challenges of application packaging, the differences between containers and virtual machines, and provides hands-on labs for installing Docker on different systems. Additionally, it introduces orchestration platforms including Docker Swarm and Kubernetes, emphasizing their features for managing containerized applications.
![ManageImplementAdviseEducate
Container & Images
• Containers are running instances of image
• Image is a read-only filesystem made of “layers”
• Each “layer” is an image which may contain a specific application, library or runtime.
• Container has an extra “Writable” layer.
• The bottom-most image is “base” image.
Host Kernel
Debian Image [Read Only]
BusyBox Image [Read Only]
Apache2 HTTP Image [Read Only]
MySQL DB Image [Read Only]
Writable Layer
Writable Layer](https://image.slidesharecdn.com/mspptfor-containerorchestration-part1-200218104152/85/Introduction-to-Containers-Diving-a-little-deeper-into-the-benefits-of-Containerisation-11-320.jpg)
![ManageImplementAdviseEducate
Docker-compose reference
Section Description
Services: Define one or more service [YAML syntax]
Build: Provide relative path to directory which contains “Dockerfile”
Image: Provide image name using syntax: [repository-name]/[image-name]:[tag]
For public images on docker-hub [repository-name] is dropped.
Volumes: Define volumes to be mounted
Networks: Define a private network for application.
Labels: Additional meta-data](https://image.slidesharecdn.com/mspptfor-containerorchestration-part1-200218104152/85/Introduction-to-Containers-Diving-a-little-deeper-into-the-benefits-of-Containerisation-31-320.jpg)
![ManageImplementAdviseEducate
Hands On
• HOL 2] Deploying front-end application (ASP.NET core) and Back-end (MS
SQL) using single docker-compose.
• Use docker-compose to define entire application
• Use existing VM created in HOL-1](https://image.slidesharecdn.com/mspptfor-containerorchestration-part1-200218104152/85/Introduction-to-Containers-Diving-a-little-deeper-into-the-benefits-of-Containerisation-33-320.jpg)
Introduction to Containers & Diving a little deeper into the benefits of Containerisation
- 1. ManageImplementAdviseEducate Container Orchestration using Docker & Kubernetes By Mahendra Shinde
- 2. ManageImplementAdviseEducate Agenda Module 1: Introduction to Containers Module 2: Container Architecture Module 3: Docker Architecture HOL 1: Installing Docker CE on Windows & Linus Host Module 4: Application modernization using docker Demo : Deploying TWO TIER application Using Docker compose HOL 2: Deploying TWO TIER application (DotNet Core + MS SQL) using Docker compose Module 5: Understanding Orchestration
- 3. ManageImplementAdviseEducate Module 1 Introduction to Containers
- 4. ManageImplementAdviseEducate Application packaging strategies Every Application has certain dependencies. Libraries provided by Operating System Libraries provided by runtime environment like Java, Python & Dot Net runtime Server Runtime like Tomcat for Java, IIS for Asp.net, Apache httpd Third Party Libraries Change in Dependencies affects Application.
- 5. ManageImplementAdviseEducate Application packaging strategies /challenges Application goes through following phases: Development Testing Staging Production Managing dependencies across all these environments could be challenging. Creating compatible dev-test environment may take considerable time. Question : Challenges in moving application between environments Question: List Existing solutions based on Virtualization
- 6. ManageImplementAdviseEducate Containers as alternative • A Typical container would have application & all its dependencies packed together. • This makes container “decoupled” from targeted environment. • Only restriction being Windows or Linux platform. Container Base Image Runtime (JRE / Python / Node) Third Party Libs Application Image Writable Layer
- 7. ManageImplementAdviseEducate Benefits of Containers • Isolate application from underlying OS and Hardware. • Lightweight, provides higher density than Virtual machines. • Dev-Ops ready. Many CI/CD platforms support container deployments/build. • Every container executes in a separate user space.
- 8. ManageImplementAdviseEducate Container vs Virtual machines • Virtual machines provides hardware level virtualization. • Virtual machines contains Operating system. • Virtual machines use guest OS kernel. • Host OS and Guest OS could be completely different. • Booting up VM is equivalent of booting a physical machine. • Containers provide software level isolation, no hardware virtualization. • Containers use “base image” which provide basic OS level libraries but not the OS. • Containers use host OS kernel. • Due to dependence on Host OS Kernel, containers must be deployed on compatible host OS. • Launching container is equivalent of launching an application.
- 9. ManageImplementAdviseEducate Container platforms • Docker from Docker Inc • LXC from Linux • LXD from Ubuntu • RKT from CoreOS
- 11. ManageImplementAdviseEducate Container & Images • Containers are running instances of image • Image is a read-only filesystem made of “layers” • Each “layer” is an image which may contain a specific application, library or runtime. • Container has an extra “Writable” layer. • The bottom-most image is “base” image. Host Kernel Debian Image [Read Only] BusyBox Image [Read Only] Apache2 HTTP Image [Read Only] MySQL DB Image [Read Only] Writable Layer Writable Layer
- 12. ManageImplementAdviseEducate Read-Only vs Writable layers • Read-only layers are shared by multiple containers / images. • Writable layer allows read and write operations • Every container has ONE writable layer. • Every layer has unique ID and meta-data.
- 13. ManageImplementAdviseEducate Image repositories • An Image repository is persistent storage for container images • A private repository can be accessed by selected users. • A public repository can be accessed by any one. • A Developer may “Push” application as container image. • An operations team may “Pull” the same image for deployment. • Many cloud platforms do provide “Hosted” container repositories. • Host repositories are managed by vendor.
- 14. ManageImplementAdviseEducate Container Networking • Containers can communicate through a container network. • Container platform like docker have Multiple network models. Host Kernel Container Network Container 1 Container 2
- 16. ManageImplementAdviseEducate Docker As a Container platform • An open platform for developing, shipping, and running applications. • Started it’s life as a PaaS provider by name dotCloud. • Since its inception, working with linux container. • The container platform was named “Docker” • In 2013, got “rebranded” as Docker Inc. • A container platform based on “runc” and “libcontainer”
- 17. ManageImplementAdviseEducate Docker Editions • Community edition (Docker CE) • Ideal for developers and small teams experimenting with containers. • Enterprise Edition (Docker EE) • Enterprise development & IT team for business critical applications at production scale.
- 18. ManageImplementAdviseEducate Docker components • Docker engine (docker daemon) • Docker container images • Docker containers • Docker CLI (client) • Docker (REST) API • Docker Registry
- 19. ManageImplementAdviseEducate Docker components • Batteries included but “removable” (“replaceable”) • Built in security • Linux hosts can use AppArmor or SELinux • Namespaces to restrict access to host system resources.
- 20. ManageImplementAdviseEducate Docker on Windows & Linux hosts • Docker being only container platform supporting both Windows and Linux. • Windows based container supported only on Windows 10 PRO and Windows Server 2016. • Docker installation on windows uses different components than linux host. • Images built for linux cannot be deployed on windows hosts. • Windows do have a trick: docker-machine and moby project to run linux containers through a lightweight VM.
- 21. ManageImplementAdviseEducate Hands On LAB • HOL 1: Installing Docker CE on Windows Server 2016 • Pre-requisite: • Azure subscription • Basic Windows administration w/ powershell. • Windows server 2016 VM on Azure • Remote Desktop client (built in windows) for accessing windows VM.
- 22. ManageImplementAdviseEducate Hands On LAB • HOL 2: Installing Docker CE on Ubuntu Server 18.04 LTS • Pre-requisite: • Azure subscription • Basic Linux administration using bash shell. • Ubuntu Server 18.04 LTS VM on Azure • SSH Client (PuTTY or Git Bash or Ubuntu Bash)
- 24. ManageImplementAdviseEducate Application types • A multi-layered monolith application • A distributed application / SOA • Micro-services • Containers do support all of them! • Monolith = Single large Container for App + DB Container • Distributed = Container for each Service / Module
- 25. ManageImplementAdviseEducate A typical TWO TIER Application Front End App Container Back End App Container Back End App Container Back End App Container Front End App Container
- 26. ManageImplementAdviseEducate Dockerfile • Automates the “Build” process of a container. • Integrates with all popular CI/CD tools. • Simple language to define entire build workflow • Basic system administration commands needed. • Windows : prefer powershell • Linux: prefer bash • Can be placed in SCM
- 27. ManageImplementAdviseEducate Docker compose • A Tool for defining and deploying multi-container application. • Compose is a THREE STEP process • Define the application environment using Dockerfile • Define the services that make up your application using docker-compose.yml file • Run docker-compose up command to quickly run entire application in isolated environment. • Provides re-producible deployment artefacts. • Only recreates container that have changed. • Upon “restart” containers are spawned from “cached” versions if no change detected. • Supports environment variables • Supports “secrets” • Traditionally used for Single host deployment.
- 29. ManageImplementAdviseEducate Docker Compose vs Dockerfile
- 30. ManageImplementAdviseEducate Dockerfile reference Command Description FROM Initializes a new build and define “base” image to be used for building. RUN A Step to be executed in build process. Can be used for installing necessary package. CMD Provide DEFAULT “startup” action for container. Must be used only ONCE. LABEL Provides additional META-DATA like Author name EXPOSE A Hint for port used by container, can be overridden at run-time. ENV Define environment variables. COPY Copy files from host system to container. NOTE : optional feature --chown for linux containers
- 31. ManageImplementAdviseEducate Docker-compose reference Section Description Services: Define one or more service [YAML syntax] Build: Provide relative path to directory which contains “Dockerfile” Image: Provide image name using syntax: [repository-name]/[image-name]:[tag] For public images on docker-hub [repository-name] is dropped. Volumes: Define volumes to be mounted Networks: Define a private network for application. Labels: Additional meta-data
- 32. ManageImplementAdviseEducate Demos • Demo 1: Deploying an application back-end and front-end as single application. • Demo 1-A Using MySQL with Wordpress • Demo 1-B Using MS-SQL with Asp.net core
- 33. ManageImplementAdviseEducate Hands On • HOL 2] Deploying front-end application (ASP.NET core) and Back-end (MS SQL) using single docker-compose. • Use docker-compose to define entire application • Use existing VM created in HOL-1
- 35. ManageImplementAdviseEducate Orchestration • Need for Orchestration • Ephemeral computing • Desired State • Cluster management • Container Orchestration platforms • Docker Swarm (from Docker Inc) • Kubernetes (from Google) • DC/OS (based on Apache Mesos) • Nomad (from HashiCorp)
- 36. ManageImplementAdviseEducate Docker Swarm • Cluster management integrated with Docker engine • Decentralized design (for cluster management) • Declarative service model (ref: docker-compose) • Scaling • Desired state reconciliation • Multi-host networking • Service discovery • Load balancing
- 37. ManageImplementAdviseEducate Kubernetes • Cluster managed from google with 10+ years of internal use by google. • Open source, run anywhere • Service discovery • Load balancing • Self healing • Storage orchestration • Scaling
- 38. ManageImplementAdviseEducate DC/OS • Distributed OS based on Apache Mesos • Treat cluster as a single machine • High resource utilization • Container orchestration using “Marathon” • Cloud agnostic installation