docker_swarm module does not actually apply node_id parameter

[jwitko]

SUMMARY

the docker_swarm ansible module contains a parameter to specify node_id in the case of state=remove. In its current state the module does not apply this parameter regardless if its specified or not.

ISSUE TYPE

• Bug Report

COMPONENT NAME

ANSIBLE VERSION

ansible 2.7.7
  config file = /home/admin.witkowski/throtle-ansible/ansible.cfg
  configured module search path = [u'/home/admin.witkowski/throtle-ansible/library']
  ansible python module location = /home/admin.witkowski/throtle-ansible-venv/lib/python2.7/site-packages/ansible
  executable location = /home/admin.witkowski/throtle-ansible-venv/bin/ansible
  python version = 2.7.5 (default, Oct 30 2018, 23:45:53) [GCC 4.8.5 20150623 (Red Hat 4.8.5-36)]

CONFIGURATION

ANSIBLE_PIPELINING(/home/admin.witkowski/throtle-ansible/ansible.cfg) = True
ANSIBLE_SSH_ARGS(/home/admin.witkowski/throtle-ansible/ansible.cfg) = -o UserKnownHostsFile=/dev/null -o StrictHostKeyChecking=no -o ForwardAgent=
DEFAULT_CALLBACK_PLUGIN_PATH(/home/admin.witkowski/throtle-ansible/ansible.cfg) = [u'/home/admin.witkowski/throtle-ansible/library/callback']
DEFAULT_FILTER_PLUGIN_PATH(/home/admin.witkowski/throtle-ansible/ansible.cfg) = [u'/usr/share/ansible_plugins/filter_plugins', u'/home/admin.witko
DEFAULT_FORKS(/home/admin.witkowski/throtle-ansible/ansible.cfg) = 15
DEFAULT_GATHERING(/home/admin.witkowski/throtle-ansible/ansible.cfg) = smart
DEFAULT_GATHER_SUBSET(/home/admin.witkowski/throtle-ansible/ansible.cfg) = !hardware,!ohai,!facter
DEFAULT_HOST_LIST(/home/admin.witkowski/throtle-ansible/ansible.cfg) = [u'/home/admin.witkowski/throtle-ansible/inventory']
DEFAULT_MANAGED_STR(/home/admin.witkowski/throtle-ansible/ansible.cfg) = Created by Ansible
DEFAULT_MODULE_PATH(/home/admin.witkowski/throtle-ansible/ansible.cfg) = [u'/home/admin.witkowski/throtle-ansible/library']
DEFAULT_PRIVATE_KEY_FILE(/home/admin.witkowski/throtle-ansible/ansible.cfg) = /home/admin.witkowski/.ssh/throtle-ansible.key
DEFAULT_ROLES_PATH(/home/admin.witkowski/throtle-ansible/ansible.cfg) = [u'/home/admin.witkowski/throtle-ansible/roles']
DEFAULT_SCP_IF_SSH(/home/admin.witkowski/throtle-ansible/ansible.cfg) = True
DEFAULT_VAULT_PASSWORD_FILE(/home/admin.witkowski/throtle-ansible/ansible.cfg) = /home/admin.witkowski/throtle-ansible/.vault
DISPLAY_ARGS_TO_STDOUT(/home/admin.witkowski/throtle-ansible/ansible.cfg) = False
HOST_KEY_CHECKING(/home/admin.witkowski/throtle-ansible/ansible.cfg) = False
RETRY_FILES_SAVE_PATH(/home/admin.witkowski/throtle-ansible/ansible.cfg) = /home/admin.witkowski/.ansible/retry-files

OS / ENVIRONMENT

CentOS 7
Docker-CE 18.09.2, 18.09.3
Swarm mode enabled

STEPS TO REPRODUCE

    - name: Remove node from swarm
      docker_swarm:
        docker_host: 'tcp://{{ swarm_master_ip }}:2375'
        state: remove
        node_id: "{{ node_id }}" # Server must be demoted from manager in swarm

EXPECTED RESULTS

Node removed from docker swarm cluster.

ACTUAL RESULTS

The removal attempts to run on the host which is receiving the API call. Error reports about the node being a manager since it has not been demoted.

[felixfontein]

!component =lib/ansible/modules/cloud/docker/docker_swarm.py

[felixfontein]

Are you sure this already happens in 2.7.7? I think it just happens in devel, and in current stable-2.7 (which hasn't been released yet).

[felixfontein]

Anyway, a very good catch, since this is unfortunately one of the things the tests won't catch (since that requires more than one node...), and neither can I locally (with also just a one-node cluster).

[felixfontein]

Ah, sorry, it does happen in 2.7.7 as well (and probably all earlier versions). I thought I've thrown node_id out from TaskParameters (but forgot to adjust the places where it is used), but apparently it has never been in there...