Hands-on Docker-08 : Docker Swarm Basic Operations

Purpose of the this hands-on training is to give students the understanding to the Docker Swarm basic operations.

Learning Outcomes

At the end of the this hands-on training, students will be able to;

• Explain the what Docker Swarm cluster is.

• Set up a Docker Swarm cluster.

• Deploy an application as service on Docker Swarm.

• Do basic service operations within the Swarm cluster.

• Manage services within the Swarm.

Outline

• Part 1 - Launch Docker Machine Instances and Connect with SSH

• Part 2 - Set up a Swarm Cluster with Manager and Worker Nodes

• Part 3 - Managing Docker Swarm Services

• Part 4 - Updating and Rolling Back in Docker Swarm

Part 1 - Launch Docker Machine Instances and Connect with SSH

• Launch five Compose enabled Docker machines on Amazon Linux 2 with security group allowing SSH connections using the of Clarusway Docker Swarm Cloudformation Template.

• Connect to your instances with SSH.

ssh -i .ssh/call-training.pem ec2-user@ec2-3-133-106-98.us-east-2.compute.amazonaws.com

Part 2 - Set up a Swarm Cluster with Manager and Worker Nodes

• Prerequisites (Those prerequisites are satisfied within cloudformation template in Part 1)

  • Five EC2 instances on Amazon Linux 2 with Docker and Docker Compose installed.

  • Set these ingress rules on your EC2 security groups:

    • HTTP port 80 from 0.0.0.0\0

    • TCP port 2377 from 0.0.0.0\0

    • TCP port 8080 from 0.0.0.0\0

    • SSH port 22 from 0.0.0.0\0 (for increased security replace this with your own IP)

• Change the hostname of the nodes, so we can discern the roles of each nodes. For example, you can name the nodes (instances) like master, worker-1 and worker-2.

sudo hostnamectl set-hostname <node-name-manager-or-worker>

• Check if the docker is active or not from the list of docker info (should be inactive at first).

docker info

• Get the internal IP addresses of docker machines, you find out the Private IPs either on the EC2 dashboard, or from the command line by running the following command. Response should be something like 172.31.42.71.

hostname -i

• Initialize docker swarm with Private IP and assign your first docker machine as manager:

docker swarm init
# or
docker swarm init --advertise-addr <Private IPs>

Output

Swarm initialized: current node (6gpbx6r9vxwp37hjcdi41bhuc) is now a manager.

To add a worker to this swarm, run the following command:

    docker swarm join --token SWMTKN-1-4yfmqmk2xvly7mczdi21cuuhmem34v4ytrnvvkhzfajlptp4rk-dkfxxv4bq48rxomjjsjoxo5f1 172.31.9.149:2377

To add a manager to this swarm, run 'docker swarm join-token manager' and follow the instructions.

• Check if the docker swarm is active or not and explain the swarm part of docker info.

docker info

Output - Swarm part of docker info.

 Swarm: active
  NodeID: kdo4jnfu5rzzxpqej7mnqnv3m
  Is Manager: true
  ClusterID: rbs71ftk1nnl91bo9k6cyotwa
  Managers: 1
  Nodes: 1
  Default Address Pool: 10.0.0.0/8
  SubnetSize: 24
  Data Path Port: 4789
  Orchestration:
   Task History Retention Limit: 5
  Raft:
   Snapshot Interval: 10000
   Number of Old Snapshots to Retain: 0
   Heartbeat Tick: 1
   Election Tick: 10
  Dispatcher:
   Heartbeat Period: 5 seconds
  CA Configuration:
   Expiry Duration: 3 months
   Force Rotate: 0
  Autolock Managers: false
  Root Rotation In Progress: false
  Node Address: 172.31.42.71
  Manager Addresses:
   172.31.42.71:2377

• Add 2 more nodes as manager to improve fault-tolerance. It is recommended to create clusters with an odd number of managers in Swarm, because a majority vote is needed between managers to agree on proposed management tasks according to Raft Algorithm. An odd—rather than even—number is strongly recommended to have a tie-breaking consensus. Having two managers is actually worse than having one.

• Get the manager token with docker swarm join-token manager command.

docker swarm join-token manager

Output

To add a manager to this swarm, run the following command:

  docker swarm join --token SWMTKN-1-4yfmqmk2xvly7mczdi21cuuhmem34v4ytrnvvkhzfajlptp4rk-3zzm33eyq2gsj1sm0qnaygf8a 172.31.9.149:2377

• Add second and third Docker Machine instances as manager nodes, by connecting with SSH and running the given command above.

docker swarm join --token <manager_token> <manager_ip>:2377

• Add fourth and fifth Docker Machine instances as worker nodes. (Run docker swarm join-token worker command to get join-token for worker, if needed)

docker swarm join --token <worker_token> <manager_ip>:2377

• List the connected nodes in Swarm and explain difference between leader and other managers.

docker node ls

Part 3 - Managing Docker Swarm Services

Warning: If you have problem with docker swarm installation, you can use following link for lesson. https://labs.play-with-docker.com

• Create a service for visualization.

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

• Learn which node is running service viz and check if the visualizer is running by entering http://<ec2-host-name of this node>:8080 in a browser. (Nothing should be running)

docker service ps viz

• Start a nginx service with 5 replicas and show the replicas running on visualizer.

docker service create --name webserver --replicas=5 -p 80:80 -d nginx

• List services, explain what service is and the difference between container and service.

docker service ls

• Check if the Nginx Web Server is running by entering http://<ec2-host-name-of-any-node> in a browser.

• Display detailed information on service.

docker service inspect --pretty webserver

• List the tasks of service and explain what task is.

docker service ps webserver

• Fetch the logs of the service or a task.

docker service logs webserver

• Reboot a worker node and explain the last situation on visualizer app.

sudo reboot -f

• Scale up services and show the changes on visualizer app.

docker service scale webserver=10

• Scale down services and show the changes on visualizer app.

docker service scale webserver=3

• Remove service and show the changes on visualizer app.

docker service rm webserver

• Create service in global mod, show the changes on visualizer app and explain global mod.

docker service create --name glbserver --mode=global -p 80:80 nginx

• Remove a container and show that swarm creates a new task immediately.

docker container rm -f <containerid>
docker service ps glbserver

• Leave worker nodes from swarm and show the changes on visualizer app.

docker swarm leave

• Remove the glbserver service.

docker service rm glbserver

• Leave manager nodes from swarm

docker swarm leave --force

• Join manager and worker nodes again.

docker swarm join --token SWMTKN-1-1nkizkzhhqk4i7blzwww3znwhd0lqfsu3nlh9gl1pe7mco84up-5468yx80p9nfowv4eck0dqrvd 172.31.38.249:2377

Part 4 - Updating and Rolling Back in Docker Swarm

• Create a new service of clarusways/container-info:1.0 with 10 replicas.

docker service create --name clarusweb -p 80:80 --replicas=10 clarusways/container-info:1.0

• Explain docker service update command.

docker service update --help

• Update clarusways/container-info:1.0 image with clarusways/container-info:2.0 image and explain the changes.

docker service update --detach --update-delay 5s --update-parallelism 2 --image clarusways/container-info:2.0 clarusweb
watch docker service ps clarusweb

• Revert back to the earlier state of clarusweb service and monitor the changes.

docker service rollback --detach clarusweb
watch docker service ps clarusweb

• Remove the service.

docker service rm clarusweb