Docker Intro for not beginners.
-
Go to http://play-with-docker.com sign in with your user:pass (create account if needed).
-
Click on the wrench and select either "3 Managers and 2 Workers" or "5 managers and no workers"
-
Unless instructed run all the commands on the first node.
-
Make sure to clone the repo in the swarm nodes (PWD)
git clone https://github.com/jmarcos-cano/compose-swarm.git
cd compose-swarmSlides to this repo can be found here
#create a local network, so that containers can see each other
docker network create mynetwork
# create the backend Redis container and attach it to the network
docker run --name redis -d --network mynetwork redis:alpine
# create the app container, expose it in a different port
docker run -p 5500:5000 -it --network mynetwork -e "REDIS_HOST=redis" mcano/docker:intro
# OR if you prefer local environment development supported by Docker
docker run -p 5500:5000 -it --network mynetwork -e "REDIS_HOST=redis" -v $(pwd):/code mcano/docker:intro shto start using this project make sure you follow up the .env.example fill it with your own variables
cp .env.example .env# build and up
docker-compose -f docker-compose.yml -f docker-compose-build.yml up --build
# only build
docker-compose -f docker-compose.yml -f docker-compose-build.yml build
# push it
docker-compose -f docker-compose.yml -f docker-compose-build.yml pushdocker-compose upThis section will give you the necessary to go full to production with Docker swarm mode.
docker service create \
--name=viz \
--publish=8080:8080/tcp \
--constraint=node.role==manager \
--mount=type=bind,src=/var/run/docker.sock,dst=/var/run/docker.sock \
dockersamples/visualizer
# wait until it says "service converged"go to your visualizer (click in your upper link port 8080) and see how the services are spread.
# Create a swarm service from a Nginx docker image
docker service create --name nginx-ws -p 80:80 nginx
# List the current services
docker service ls
⚠️ Go to your visualizer (click in your upper link port 8080) and see how the services are spread.
⚠️ Click also on Port 80 (Nginx) - it should say "Welcome to Nginx"
Scale the service
docker service update --replicas 3 nginx-wsGo to your visualizer (click in your upper link port 8080) and see how the services are spread.
Checking service logs
docker service logs nginx-wsDelete the service
docker service rm nginx-ws
# check for running services
docker service ls# inspect the stack file and try to understand it
cat docker-compose.simple.yml
# deploy it
docker stack deploy -c docker-compose.simple.yml --resolve-image=always --with-registry-auth intro
# list current services
docker service ls
⚠️ Go to your app (click in your upper link port 500) and see how the app looks like. !
Go to your visualizer (click in your upper link port 8080) and see how the services are spread.
Show current status
docker service ps docker_intro_web💡 This will give you a small intro to how you can manage configuration per environment (dev,qa,stage,production)
# inspect the stack file and try to find the directive "FOO=${FOO:-BAR}"
cat docker-compose.simple.yml
# inject the new value
export FOO="Hola Edmundo"
# deploy it and see it update automatically
docker stack deploy -c docker-compose.simple.yml --resolve-image=always docker_intro🥇 Dare you to put your own Text there, see how sometimes the application becomes unaccessible?
Want to handle more traffic? Want to be more resilient? Want High Availability?
Swarm got you covered
docker service scale docker_intro_web=4
⚠️ Go to your app (click in your upper link port 500) and see how which task/container responds
Go to your visualizer (click in your upper link port 8080) and see how the services are spread.
Instead of scaling your service everytime, why don't we declare it?
# Inspect the .replicas file and find "deploy: " section
cat docker-compose.replicas.yml
# Deploy new update for the stack
docker stack deploy -c docker-compose.replicas.yml --resolve-image=always docker_introGo to your visualizer (click in your upper link port 8080) and see how the services are spread.
🥇 I dare you to set more than 3 replicas for docker_intro_web, how?
Rolling updates let you update your app with zero-downtime.
💡 v1 has been running for 2 weeks now and you are ready to ship your new and hottest feature on v2, with rolling updates you can easily ship v2 let it coexist with v1 until v1 gets fully drain (removed) and v2 gets out.
# inspect .rolling file and find the "update_config:" section, try to understand it
less docker-compose.rolling.yml
# Deploy/update this new configuration for your stack
docker stack deploy -c docker-compose.rolling.yml --resolve-image=always docker_intro
press 'q' to exit from 'less'
Do this how many times you need in order to see it working.
# graceful full restart of your app
docker service update --force docker_intro_webGo to your visualizer (click in your upper link port 8080) and see how the services are spread.
One can prevent memory starvation or CPU consumption of your app by adding "resources:" section
# inspect .resources file and find the "resources:" section, try to understand it
less docker-compose.resources.yml
# Deploy/update this new configuration for your stack
docker stack deploy -c docker-compose.resources.yml --resolve-image=always docker_intro
Auto restarts and health-check can also be possible by adding "healthcheck: "
# Run docker ps first to see there's no (healthy)
docker ps
# inspect .health file and find the "healthcheck:" section, try to understand it
less docker-compose.health.yml
# Deploy/update this new configuration for your stack
docker stack deploy -c docker-compose.health.yml --resolve-image=always docker_intro
# after a few seconds run
docker psDo a: docker service ps docker_intro_web, Identify the placement of a container (identify on which node is running).
Jump into that node and run docker ps find the container and its ID (first column), kill it and see how it self heals
docker kill <container ID>Go to your visualizer (click in your upper link port 8080) and see how the services are spread and self healed.
🥇 I dare you to do a rolling update with healthcheck included and see what happens, can you predict what will happen ahead of 'time'?
Can you explain why this times do not match up?
docker stack deploy -c docker-compose.rolling.yml --resolve-image=always docker_intro
time docker service update --force docker_intro_web
docker stack deploy -c docker-compose.health.yml --resolve-image=always docker_intro
time docker service update --force docker_intro_web
React
Typescript
Django
Laravel
Facebook
Microsoft
Google
Alibaba
D3
Tencent