Skip to content

Latest commit

 

History

History
178 lines (108 loc) · 4.42 KB

notebook.md

File metadata and controls

178 lines (108 loc) · 4.42 KB

Hypervisor-based virtual machines-->compute,memory capacity is fixed cannot run more workloads on same machines as compared to docker

1.containers are loosely held isolated environments, light weight because they dont need the extra load of hypervisor they are portable--independent of machine can run more no of workload on same hardware bcoz of light weight

2.containers runs directly on the host machine's kernel instead of running on HOST machine hardware as in VMs

3.Docker containers can run on host machine even if the host machine is a VM

4.Docker engine:client-server model

docker-client("docker" CLI)<--->REST API<--->server[daemon is a long running process]("dockerd" cmd)

  1. Dockerfile is used to define docker image

DOCKER ARCHITECTURE:

docker daemon docker client docker registeries docker objects: images, container, services

docker daemon(dockerd): Listens API and manages docker image creation

docker-machine ip

docker ps -a

docker run -p 4000:80 -d hello-node:1.0
mapping pod 4000 of local host with port 80 of container 
-d: run container in background

docker run -p 4001:80 -d --name container2 $IMG_NAME

docker log container1

docker rmi -f image-name

docker image inspect <image-name>

docker ps -q (Running containers id)

docker rm -f $(docker ps -a -q)


docker inspect $INSTANCE_ID # return a low level description of docker object

Getting the instance IP addr

$docker inspect --format='{{range .NetworkSettings.Networks}}{{.IPAddress}}{{end}}' $INSTANCE_ID

List all the binding ports

$ docker inspect --format='{{range $p, $conf := .NetworkSettings.Ports}} {{$p}} -> {{(index $conf 0).HostPort}} {{end}}' $INSTANCE_ID

"registery collection of repository,repository is collection of built images"

dockerhub.io is a registory gcr.io username/repo:tag

SERVICES:

  • In distributed application, different pieces of app are called as services
  • Services are nothing but "containers in production"
  • For creating the services in docker we create "docker-compose.yaml"
  • services contains container's which are of same image
  • Automatic load balancing,scaling
  • Services allows to deploy Multi-container application

before deploying the stack you have to init a swarm stack-->services

Stack allows to deply application with one or more services interdependent with other

docker swarm init 
docker stack deploy -c docker-compose.yml getstartedlab<stack name>
docker stack ls
docker stack rm getstarted<stackname>
docker stack logs <stackname> -->Getting the logs for a particular stack

docker service ls
docker service logs serviceName<getstarted_web>
docker service ps <service-name> -->shows the task for one or more service

docker swarm leave --force

for deploying stack docker-compose.yaml file name is must

DOCKER SWARM:

  • a group of machine running docker, joined into a cluster(a grp of machines/nodes)
  • In swarm, we have one MANAGER and other node as WORKER Only manager-node is authorized to run the cmds while the worker-nodes are just to provide capacity.
  • The node which run docker swarm init becomes manager and workers are joined using the command docker join --token
  • Swarm participates an ingress routing mesh so that all containers port point at same port
docker-machine create --driver virtualbox myvm1
docker-machine ls


docker swarm init 
docker node ls
docker stack deploy -c docker-compose.yaml stack1
docker stack ls
docker swarm leave --force

DOCKER STACK

  • Stack is a group of inter-related services which share some dependencies with each other for proper functionality of entire application

  • In production environment we use single stack which can be orchested and scaled together

      	  service1-----		
  	 /			  |
  	/			  |

stack ---service2---------|----->SINGLE STACK ENTIRE APPLICATION \ | \ | service3----|

service1:a hello-flask container om port 80i.e http(displaying container id)

service2:Visualizer showing the different services running on different host running on port 8080

service3:redis db keeping the count for the number of visitor and storing on persistently on host machine's file system

80(http),443(https),22(ssh)

  • Functionality of entire application can be defined into a single stack
docker swarm init

docker stack deploy -c docker-compose.yaml stack_name
docker stack ls
docker stack logs <stack_name>
docker service ls

docker stack rm <stack_name>