DevOps Classroom Series – 03/Apr/2021

Docker Compose

• This is a tool for defining an running multi container Docker applications
• All the modern applications will have the some/all of the following components
• When all of these components work together we have a successful application
• Lets understand microservices
• To bringup our application we need to run multiple containers in a specific order.
• Docker compose is one approach to bring up multiple container of your application on a single docker host platform.
• Docker compose is widely used in Developer environments
• Docker stack is one more approach which helps to run multiple containers of your application on multi nodes.

Microservices

• This is an architectural pattern that structures and application as a collection of services that are
  • Highly maintable and testable
  • Loosely coupled
  • Independently deployable
  • Organized around business capabilities
• Simple microservice architecture
• Now we can deploy each service in a container/vm Now this gives us the following benifits
  • We can have parallel developments for each service
  • Reusability of services can be done across applications
  • We can scale the individual services when there is load.
• The best way to run individual services is in containers and to manage the whole application deploymets we use orchestration platforms such as
  • Kubernetes
  • Docker Swarm
  • OpenShift

How to work with Docker compose

• In Docker compose we create a file called docker-compose.yaml and we specify the containers to be created, their network, their volumes as specified in docker compose specifications Refer Here

• In with the code repository docker-compose.yaml is commited to make it easier to bring up multiple containers to run an application

• Docker-Compose is pre-installed when you install docker on mac and windows 10

• On Linux machines we need to install docker-compose Refer Here

• To learn about json and yaml Refer Here

• Hello-world:

  • In this lets try to create a httpd container and nginx container
  • Refer Here for the docker-compose.yaml

• Building images and running containers

  • Refer Here for the docker-compose.yaml
  • Execute the commands

  docker-compose build
  docker-compose up -d
  docker container ls
  docker image ls
  

  

• Creating a docker-compose file with a network and volume

  • Refer Here for the docker-compose file
  • Execute-commands

  docker-compose build
  docker-compose up -d
  docker container ls
  docker image ls
  docker volume ls
  docker network ls
  

  

Run container in Production Environments

• We generally run Production environment on multiple servers
• So if we run our application which is microservices in the production environment, We have to deal with
  • Multiple containers
  • Ensure that the containers are running
  • We need to scale the containers depending on demand
  • We need to persist the volumes
• A distributed systems cluster is a group of machines that are virtually or geographically separted and they work together to provide the same service or applications to clients
• A distributed system can help
  • There is no single point of failure (SPOF)
  • This design supports scalability
  • The data is shared b/w all machines
  • The design is readable
• Consensus is the task of getting all the systems ina group agree on some specific values based on votes. Docker swarm uses raft consensus algorithm Refer Here
• The configuration or state of the system will be stored in the database on all managers. In the case of docker swarm this database is etcd
• We need to manage networks for multiple nodes and we already know about overlay
• Docker Swarm supports loadbalancing to interact with services externally as well as internally
• How can we deploy our application i.e. new microservice version in container with out downtime?
• Docker has a swarm mode which is running docker containers on multiple nodes with cluster features
  • Manager & Worker nodes
  • Multi node networking
  • Loadbalancing
  • Deployments
  • Scalability
• To be very frank industry has majorly adopted kubernetes as the orchestration technology for running contaniers in production.
• To run multiple applications on docker swarm which is
  • Service => Task
  • Docker has developed a docker stacks or you can use docker command line

Running Docker Containers

• With CPU Limits
• With memory limits
• Refer Here for limiting cpu, ram for docker container
• Docker container allows us to change few configurations while the container is running