docker_container_exec module generates warning on exit when using stdin and managed node has python 3.13

[btmc]

SUMMARY

When running docker_container_exec module on Debian 13 managed nodes I get a non-critical warning, while the task I run in docker exec is actually being done.

There is no warning when I apply it to Debian 12 nodes, which have python 3.11 (the task itself and the controller environment are exactly the same).

And this only happens when I use stdin option to pass the command to sh, when passing it in args there is no warning.

The problem seems to be in this arm of code. Here exec_socket (a raw socket from Response object) is exposed and returned, and closed further on in a finally block.

In other arms the close() method of the Response itself is called within client.post_json_to_stream, with no warnings in that case.

The exec_socket actually has the whole Response object as its attr, so this Response object's close() method is invoked on exit, but flush() on already closed socket generates this exception.

These errors were previously ignored silently.

If I replace socket's close() with its response's close() here, the exception goes away.

                    exec_socket._response.close()

ISSUE TYPE

• Bug Report

COMPONENT NAME

docker_container_exec module

ANSIBLE VERSION

ansible [core 2.18.12]
  config file = .ansible.cfg
  configured module search path = ['.ansible/plugins/modules', '/usr/share/ansible/plugins/modules']
  ansible python module location = /usr/local/lib/python3.14/site-packages/ansible
  ansible collection location = .ansible/collections:/usr/share/ansible/collections
  executable location = /usr/local/bin/ansible
  python version = 3.14.3 (main, Feb  4 2026, 20:07:43) [GCC 15.2.0] (/usr/local/bin/python3.14)
  jinja version = 3.1.6
  libyaml = True

COLLECTION VERSION

Collection       Version
---------------- -------
community.docker 4.8.5

CONFIGURATION

CONFIG_FILE() = .ansible.cfg
DEFAULT_HOST_LIST(.ansible.cfg) = ['inventory']
DEFAULT_JINJA2_EXTENSIONS(.ansible.cfg) = jinja2.ext.do
DUPLICATE_YAML_DICT_KEY(.ansible.cfg) = ignore
HOST_KEY_CHECKING(.ansible.cfg) = False
INTERPRETER_PYTHON(.ansible.cfg) = auto_silent
PAGER(env: PAGER) = less

OS / ENVIRONMENT

Managed node:

Debian GNU/Linux 13 (trixie)

Python 3.13.5

STEPS TO REPRODUCE

Case 1: using stdin

- hosts:
    managed-node
  
  tasks:
  - community.docker.docker_container_exec:
      container:
        some-container
      command:
        sh
      stdin: |
        echo 123

Case 2: not using stdin

- hosts:
    managed-node
  
  tasks:
  - community.docker.docker_container_exec:
      container:
        some-container
      command:
        sh -c 'echo 123'

EXPECTED RESULTS

In both cases:

TASK [community.docker.docker_container_exec] ************************************************************************************************************************************************************************
changed: [managed-node]

ACTUAL RESULTS

Case 1:

TASK [community.docker.docker_container_exec] ************************************************************************************************************************************************************************
[WARNING]: Module invocation had junk after the JSON data: Exception ignored in: <http.client.HTTPResponse object at 0x7fa4ce6bb880> Traceback (most recent call last):   File "/usr/lib/python3.13/http/client.py",
line 432, in close     super().close() # set "closed" flag   File "/usr/lib/python3.13/http/client.py", line 445, in flush     self.fp.flush() ValueError: I/O operation on closed file.
changed: [managed-node]

Case 2:

TASK [community.docker.docker_container_exec] ************************************************************************************************************************************************************************
changed: [managed-node]

[felixfontein]

I tried to reproduce this with Python 3.14, but I don't get this error. How do you connect to the Docker daemon?

[felixfontein]

Ping @btmc