Cannot add new node to swarm cluster using docker_swarm

[kimdre]

SUMMARY

I'm trying to add an additional worker node to my swarm cluster but the ansible playbook when triying to join the second worker node.

ISSUE TYPE

• Bug Report

COMPONENT NAME

ANSIBLE VERSION

- hosts: server_type_worker
  tasks:
    - name: add workers to swarm
      docker_swarm:
        state: join
        advertise_addr: "{{private_ipv4}}"
        remote_addrs: ["{{hostvars['manager-1'].private_ipv4}}"]
        join_token: "{{hostvars['manager-1'].swarm_init_result.swarm_facts.JoinTokens.Worker}}"

COLLECTION VERSION

# /home/kim/.ansible/collections/ansible_collections
Collection       Version
---------------- -------
community.docker 3.10.3 

# /usr/lib/python3.12/site-packages/ansible_collections
Collection       Version
---------------- -------
community.docker 3.9.0  

CONFIGURATION

CONFIG_FILE() = /home/kim/Git/Magicline/magicline-server-deployment/ansible/ansible.cfg
DEFAULT_HOST_LIST(/home/kim/Git/Magicline/magicline-server-deployment/ansible/ansible.cfg) = ['/home/kim/Git/Magicline/magicline-server-deployment/ansible/inventory/prod.hcloud.yaml']
DEFAULT_REMOTE_USER(/home/kim/Git/Magicline/magicline-server-deployment/ansible/ansible.cfg) = root
EDITOR(env: EDITOR) = code --wait
HOST_KEY_CHECKING(/home/kim/Git/Magicline/magicline-server-deployment/ansible/ansible.cfg) = False
INVENTORY_ENABLED(/home/kim/Git/Magicline/magicline-server-deployment/ansible/ansible.cfg) = ['hcloud']
PAGER(env: PAGER) = most

OS / ENVIRONMENT

Fedora 40 and also Github Actions on Linux Runner
Worker Node has Ubuntu 24.04 LTS

STEPS TO REPRODUCE

...
- hosts: server_type_worker
  tasks:
    - name: add workers to swarm
      docker_swarm:
        state: join
        advertise_addr: "{{private_ipv4}}"
        remote_addrs: ["{{hostvars['manager-1'].private_ipv4}}"]
        join_token: "{{hostvars['manager-1'].swarm_init_result.swarm_facts.JoinTokens.Worker}}"

EXPECTED RESULTS

ACTUAL RESULTS

worker-1 already exists, worker-2 is the new node.

TASK [add workers to swarm] ********************************************************************************************************************************************************************************************************
ok: [worker-1]
fatal: [worker-2]: FAILED! => {"changed": false, "msg": "Error connecting: Error while fetching server API version: HTTPConnection.request() got an unexpected keyword argument 'chunked'"}

[felixfontein]

You probably want to upgrade Docker SDK for Python on the target where the module is executed.

This bug was tracked in docker/docker-py#3113 and got fixed in Docker SDK for Python 7.0.0. You can alternatively downgrade urllib3 to a version < 2.

[kimdre]

Thank you! After installing docker on the new fresh node, the python docker SDK was still on 5.0.3 instead of >7.0.0 so I had to install the latest version manually. This fixed it for me. :)