'How to activate python virtual environment on remote machine with ansible?
I'm learning Vagrant and Ansible, I'm trying to setup a local development environment for a basic flask app in ubuntu20.04 with Nginx.
my vagrantfile looks like this:
Vagrant.configure("2") do |config|
config.vm.define :ubuntuserver do | ubuntuserver |
ubuntuserver.vm.box = "bento/ubuntu-20.04"
ubuntuserver.vm.hostname = "ubuntuserver"
ubuntuserver.vm.provision :ansible do | ansible |
ansible.playbook = "development.yml"
end
ubuntuserver.vm.network "private_network", ip:"10.11.1.105"
ubuntuserver.vm.network "forwarded_port", guest: 80, host: 8080
ubuntuserver.vm.network "public_network", bridge: "en1: Wi-Fi (AirPort)"
ubuntuserver.vm.provider :virtualbox do |vb|
vb.memory = "1024"
end
ubuntuserver.vm.synced_folder "./shared", "/var/www"
end
end
my ansible-playbook like so:
-
name: local env
hosts: ubuntuserver
tasks:
- name: update and upgrade apt packages
become: yes
apt:
upgrade: yes
update_cache: yes
- name: install software properties common
apt:
name: software-properties-common
state: present
- name: install nginx
become: yes
apt:
name: nginx
state: present
update_cache: yes
- name: ufw allow http
become: yes
community.general.ufw:
rule: allow
name: "Nginx HTTP"
- name: installing packages for python env
become: yes
apt:
name:
- python3-pip
- python3-dev
- build-essential
- libssl-dev
- libffi-dev
- python3-setuptools
- python3-venv
update_cache: yes
- name: Create app directory if it does not exist
ansible.builtin.file:
path: /var/www/app
state: directory
mode: '0774'
- name: Install virtualenv via pip
become: yes
pip:
name: virtualenv
executable: pip3
- name: Set python virual env
command:
cmd: virtualenv /var/www/app/ -p python3
creates: "/var/www/app/"
- name: Install requirements
pip:
requirements: /var/www/requirements.txt
virtualenv: /var/www/app/appenv
virtualenv_python: python3
My playbook fails at the next task with error:
- name: Activate /var/www/app/appenv
become: yes
command: source /var/www/app/appenv/bin/activate
fatal: [ubuntuserver]: FAILED! => {"changed": false, "cmd": "source /var/www/app/appenv/bin/activate", "msg": "[Errno 2] No such file or directory: b'source'", "rc": 2}
Rest of the playbook
- name: ufw allow 5000
become: yes
community.general.ufw:
rule: allow
to_port: 5000
- name: Run app
command: python3 /var/www/app/appenv/app.py
From what I understand from this thread, The "source" command must be used from inside the vagrant machine. (I tried solutions from the thread but couldn't get it to work) If I ssh into the vagrant machine and execute the three last commands of my playbook manually:
source /var/www/app/appenv/bin/activate
sudo ufw allow 5000
python3 /var/www/app/appenv/app.py
my basic flask app is running on port 5000 at the IP set in the vagrantfile 10.11.1.105
My questions are:
How can I get the playbook to work and not have to ssh into the machine to accomplish the same?
Is my approach even correct, knowing that my end goal is to replicate in the vagrant machine a similar environment to what would be the production environment and develop the flask app from my local machine in the synced folder?
to give a maximum of information, if one wants to reproduce this. I also have a shared/app/appenv/app.py file containing the basic flask app
from flask import Flask
app = Flask(__name__)
@app.route("/")
def hello():
return "<h1 style='color:blue'>Hello There!</h1>"
if __name__ == "__main__":
app.run(host='0.0.0.0')
and shared/requirements.txt file
wheel
uwsgi
flask
Solution 1:[1]
This question was coming from a misconception about python venv. I thought that in order to install packages inside the virtual env it had to be activated. for example:
source env/bin/activate
pip install package_name
I understood later that I can install packages in the venv without activating it by doing:
env/bin/pip install package_name
So the solution with ansible is not to activate the venv to install packages but instead
- name: "Install python packages with the local instance of pip"
shell: "{{virtualenv_path}}/bin/pip3 install package_name"
become: no
or better even with the pip module and the packages in a requirements.txt file:
- name: Install project requirements in venv
pip:
requirements: '{{project_path}}/requirements.txt'
virtualenv: '{{virtualenv_path}}'
virtualenv_python: python3
Solution 2:[2]
I wrote this article for a Linux computer with Python 3.x. In this scenario, this is your Ansible development machine. First, verify the installed Python version and path:
# check Python version
$ python3 -V
Python 3.6.8
$ which python3
/usr/bin/python3
I recommend setting up a directory for the virtual environment:
$ mkdir python-venv
$ cd !$
Create a new virtual environment
$ python3 -m venv ansible2.9
$ ls
ansible2.9
activate python venv
$ source ansible2.9/bin/activate
(ansible2.9)$ python3 -V
Python 3.6.8
upgrade pip
(ansible2.9)$ python3 -m pip install --upgrade pip
Install Ansible in a virtual environment
(ansible2.9)$ python3 -m pip install ansible==2.9
(ansible2.9)$ which ansible
~/python-venv/ansible2.9/bin/ansible
Verify your new installation:
(ansible2.9)$ ansible --version
ansible 2.9.0
config file = /etc/ansible/ansible.cfg
configured module search path = ['/home/devops/.ansible/plugins/modules', '/usr/share/ansible/plugins/modules']
ansible python module location = /home/devops/python-venv/ansible2.9/lib64/python3.6/site-packages/ansible
executable location = /home/devops/python-venv/ansible2.9/bin/ansible
python version = 3.6.8 (default, Jan 09 2021, 10:57:11) [GCC 8.3.1 20191121 (Red Hat 8.3.1-5)]
Install Ansible roles or collections
(ansible2.9)$ ansible-galaxy collection install \
kubernetes.core:==1.2.1 -p collections
Deactivate a Python virtual environment
(ansible2.9)$ deactivate
Create another Python virtual environment for Ansible 3.0
$ python3 -m venv ansible3.0
$ ls -1
ansible2.9
ansible3.0
$ source ansible3.0/bin/activate
(ansible3.0)$ which python
~/python-venv/ansible3.0/bin/python
(ansible3.0)$ python3 -m pip install --upgrade pip
(ansible3.0)$ python3 -m pip install ansible==3.0
Sources
This article follows the attribution requirements of Stack Overflow and is licensed under CC BY-SA 3.0.
Source: Stack Overflow
Solution | Source |
---|---|
Solution 1 | Tanuki |
Solution 2 | Muhammad Fayzan |