Docker Swarm continuous delivery pipeline demo

Deploying the cluster

• Launch the cluster

• Tunneling port 8080 from localhost to the master

username=<username>
masterfqdn=<masterfqdn>
agentsfqdn=<agentsfqdn>
ssh -L 8080:localhost:8080 $username@$masterfqdn

Visualizing the state of the cluster

• Create a container with the visualizer in the master node (it is not a task of the cluster)

$> docker run -d -p 8080:8080 -v /var/run/docker.sock:/var/run/docker.sock --name visualizer dockersamples/visualizer

• Access the visualizer at localhost:8080

Manual deployment of the application

• Steps to manually deploy both apps in a new network:

docker network create --driver overlay gallifrey
docker service create --name riversong --publish 8888:8888 --network gallifrey ciberado/riversong
docker service create --name thedoctor --publish 80:80 --env INTERNAL_SERVICE_NAME=riversong  --network gallifrey ciberado/thedoctor

• Check the edge microservice with curl http://$agentsfqdn:80
• Undeploy the application with:

docker service rm thedoctor
docker service rm riversong
docker network rm gallifrey

Automatic deployment with Docker Compose

• Create the new stack:

wget -O docker-compose.yml https://raw.githubusercontent.com/capside/docker-swarm-mode-workshop/master/docker-compose.yml
docker stack deploy --with-registry-auth --compose-file docker-compose.yml test

• Check the edge microservice with curl http://$agentsfqdn:80
• Scaledown the inner service to 1

docker service scale test_riversong=1