How to fix Error response from daemon: This node is already part of a swarm in Docker

First, verify that the node is indeed part of a swarm and understand its current role:

docker info | grep -A 5 "Swarm"

This will show output like:

Swarm: active
  NodeID: abc123xyz
  Is Manager: true
  ClusterID: def456uvw
  Managers: 1
  Nodes: 1

You can also check the node's role directly:

docker node ls

If you're a manager, this shows all nodes in the swarm. If you see "Error response from daemon: This node is not a swarm manager", the node is a worker (not a manager).

If the node is a worker (not a manager), you can simply leave the swarm:

docker swarm leave

This gracefully removes the node from the swarm. The output will confirm:

Node left the swarm.

After leaving, verify the swarm status:

docker info | grep "Swarm"

Should now show:

Swarm: inactive

You can now join a new swarm or initialize a fresh one.

If the node is a manager, Docker requires the --force flag to leave:

docker swarm leave --force

Warning: If this is the last manager in the swarm, the entire swarm state (services, networks, secrets) will be lost. Make sure you understand the implications before proceeding.

For a safer approach with multiple managers:
1. First demote the manager to a worker:

# Run from another manager node
docker node demote <node-id>

2. Then leave normally:

# Run from the demoted node
docker swarm leave

3. Remove the node from the swarm's node list (from a remaining manager):

docker node rm <node-id>

After leaving the previous swarm, you can now create a new one or join an existing one.

To initialize a new swarm:

docker swarm init

Or specify an advertise address for multi-network hosts:

docker swarm init --advertise-addr <IP-ADDRESS>

To join an existing swarm as a worker:

docker swarm join --token <WORKER-TOKEN> <MANAGER-IP>:2377

Get the join token from a manager node:

docker swarm join-token worker

To join as a manager:

docker swarm join --token <MANAGER-TOKEN> <MANAGER-IP>:2377

Get the manager token from an existing manager:

docker swarm join-token manager

If you're trying to recover a swarm after losing quorum (majority of managers), use the --force-new-cluster flag:

docker swarm init --force-new-cluster

This reinitializes the swarm on the current node while preserving existing services, networks, and other swarm objects. Use this when:
- Other manager nodes are permanently lost
- The swarm lost quorum and can't be recovered normally
- You need to rebuild the swarm from a single remaining manager

Important considerations:
- Run this only on a node that was previously a manager
- All other nodes will need to rejoin the swarm
- Worker nodes will be marked as "down" until they rejoin
- This is a recovery operation, not for routine use