Describe the bug
When loading a simple playbook in the Ansible kernel (from example docs), full copies of the user's home directory are created in /tmp/ and eventually the system runs out of disk space.
To Reproduce
Install Jupyter Hub and JupyterLab
Install ansible-jupyter (tried both from git and Pypl)
Create a new notebook with Ansible kernel
Create cells from the README and attempt to run them, for example
#inventory
[all]
ahost ansible_connection=local
anotherhost examplevar=val
---
#play
name: Hello World
hosts: all
gather_facts: false
---
#task
debug:
Expected behavior
Clean return of cell output
Desktop (please complete the following information):
Python3.8.10 on Ubuntu 20.04
Jupyter Lab 2.2.2 (upgraded, also tested with unknown older version)
Jupyter Hub 3.4.0 (upgraded, also tested with unknown older version)
Additional context
This appears to be (possibly expected) behaviour of copy_files() in ansible-jupyter-kernel/ansible_kernel/kernel.py:242, but I'm not sure why it needs all of my home directory and not restricted to a subset / subdirectory.
The directories in /tmp do not get cleaned up, so successive kernel restarts and re-attempts will certainly lead to an out-of-space condition, eventually.
Describe the bug When loading a simple playbook in the Ansible kernel (from example docs), full copies of the user's home directory are created in /tmp/ and eventually the system runs out of disk space.
To Reproduce
Expected behavior Clean return of cell output
Desktop (please complete the following information):
Additional context This appears to be (possibly expected) behaviour of copy_files() in
ansible-jupyter-kernel/ansible_kernel/kernel.py:242
, but I'm not sure why it needs all of my home directory and not restricted to a subset / subdirectory.