Demo Swarm Cluster Setup

1. Initialize the swarm
2. Join worker nodes
3. Promote and demote managers
4. Drain and reactivate nodes
5. Remove nodes from the cluster

​

System Overview

​

1. Initialize the Swarm on managerone

1. Confirm Swarm is inactive:
   docker system info | grep -i swarm
   # Swarm: inactive
   
2. Initialize with the manager’s advertise address:
   docker swarm init --advertise-addr 172.31.42.232
   
   After initialization, note the worker join command output:
   docker swarm join --token <SWARM_WORKER_TOKEN> 172.31.42.232:2377
   

​

2. Add the First Worker (workerone)

docker swarm join \
  --token SWMTKN-1-3fdj9fgrjcrrj5t0pekb42n45tj96zgwxodtwd4ujv4qnhl-cop40spyhgc1tmzythfss49xn \
  172.31.42.232:2377
# This node joined a swarm as a worker.

docker node ls
# ID       HOSTNAME    STATUS  AVAILABILITY  MANAGER STATUS  ENGINE VERSION
# ...      managerone  Ready   Active        Leader          19.03.8
# ...      workerone   Ready   Active                        19.03.8

​

3. Add the Second Worker (workertwo)

​

3.1. Prepare workertwo

cat /etc/centos-release
nproc
free -m
docker --version
systemctl status docker
ping -c3 managerone
ping -c3 workerone

​

3.2. Retrieve and Run the Join Token

docker swarm join-token worker

docker node ls
# ... workertwo   Ready   Active   19.03.8

​

4. Promote a Worker to Manager

1. List current nodes:
   docker node ls
   
2. Promote workerone:
   docker node promote workerone
   # Node workerone promoted to a manager in the swarm.
   
3. Verify the new manager status:
   docker node ls
   # workerone now shows MANAGER STATUS: Reachable
   
4. (Optional) Inspect workerone in detail:
   docker node inspect workerone --pretty
   

​

5. Demote a Manager back to Worker

docker node demote workerone
docker node ls
# workerone returns to no MANAGER STATUS

​

6. Drain and Reactivate a Node

docker node update --availability drain workerone
docker node ls
# AVAILABILITY: Drain

docker node update --availability active workerone
docker node ls
# AVAILABILITY: Active

​

7. Remove a Node from the Swarm

1. Drain workerone:
   docker node update --availability drain workerone
   
2. Attempt removal (will fail if the node is still up):
   docker node rm workerone
   # Error: node is not down and can't be removed
   
3. On workerone, leave the swarm:
   docker swarm leave
   # Node left the swarm.
   
4. Back on managerone, remove the node:
   docker node rm workerone
   docker node ls
   # Only managerone and workertwo remain.
   

​

References and Further Reading

• Docker Swarm Overview
• Get Started with Swarm Mode
• Docker Engine Installation Guide