I've been banging my head on the wall trying to find the error in the code below. I have a Flask/PostgreSQL/SQLAlchemy program hosted on a GCP Compute Engine VM with Nginx and Gunicorn for a server solution.
My work is writing code, not deploying it. (It's better to have more skills, but it is what it is.) And I do not have anybody at this moment to delegate this task to or to get assistance from.
The app deploys locally without any issues. However, putting it on a GCP server is a bit of a problem. I have checked a few dozen Stackoverflow posts and Googled every phrase iteration, but I cannot figure out which part of the code is off.
Nginx server is online because when I check the domain in the browser, I get the 'Welcome to nginx!' page. But the service is failing after gunicorn is started. See the code below. Knowledgeable advice will be greatly appreciated.
my_project folder on the compute vm /home/user/
user@my_project-vm:~/my_project$ ls -la
total 72
drwxr-xr-x 8 user user 4096 Dec 27 21:00 .
drwxr-xr-x 8 user user 4096 Dec 27 21:13 ..
-rw-r--r-- 1 user user 2148 Dec 27 05:00 .env
-rw-r--r-- 1 user user 34 Dec 27 20:56 .flaskenv
drwxr-xr-x 8 user user 4096 Dec 27 20:56 .git
-rw-r--r-- 1 user user 1039 Dec 27 04:41 .gitignore
drwxr-xr-x 2 user user 4096 Dec 27 04:41 .vscode
-rw-r--r-- 1 user user 2322 Dec 27 04:41 GCP_email_notification.json
drwxr-xr-x 2 user www-data 4096 Dec 27 20:10 __pycache__
drwxr-xr-x 10 user user 4096 Dec 27 20:56 app
-rw-r--r-- 1 user user 7403 Dec 27 20:56 app.py
-rw-r--r-- 1 user user 604 Dec 27 04:41 extensions.py
srwxrwx--- 1 user www-data 0 Dec 27 21:00 my_project.sock
drwxr-xr-x 3 user user 4096 Dec 27 04:41 migrations
-rw-r--r-- 1 user user 1740 Dec 27 05:00 requirements.txt
-rw-r--r-- 1 user user 810 Dec 27 04:41 tmp-policy.json
drwxr-xr-x 6 user user 4096 Dec 27 04:48 venv
-rw-r--r-- 1 user user 45 Dec 27 20:10 wsgi.py
wsgi.py on /home/user/my_project/
from app import create_app
app = create_app()
app.py on /home/user/my_project/
from app import create_app
if __name__ == "__main__":
app = create_app()
app.run(host='0.0.0.0')
__init__.py on /home/user/my_project/app/
import ...
def create_app():
app = Flask(__name__,
static_url_path='/',
static_folder='../app/static',
template_folder='../app/templates')
...
app.config.from_object(config['default'])
config['default'].init_app(app)
@app.context_processor
...
@app.template_filter('datetime')
...
api.init_app(app)
db.init_app(app)
mail.init_app(app)
moment.init_app(app)
from app.models import ...
migrate = Migrate(app, db)
from .templates.about import about_bp
from .templates.errors import error_bp
from .templates.main import main_bp
...
app.register_blueprint(main_bp)
app.register_blueprint(error_bp)
app.register_blueprint(about_bp, url_prefix='/about')
...
return app
my_project on /etc/nginx/sites/available
server {
listen 80;
server_name my_domain.com www.my_domain.com;
location / {
proxy_pass http://unix:/home/user/my_project/my_project.sock
}
}
my_project.service on /etc/systemd/system
[Unit]
Description= You can change time text
After=network.target
[Service]
User=user
Group=www-data
WorkingDirectory=/home/user/my_project
Environment="PATH=/home/user/my_project/venv/bin"
ExecStart=/home/user/my_project/venv/bin/gunicorn --workers 3 --bind unix:my_project.sock -m 007 wsgi:app
[Install]
WantedBy=multi-user.target
This is what I get after sudo systemctl status my_project. The LAST line is where things go south...
my_project.service - my_project primary web server setup
Loaded: loaded (/etc/systemd/system/my_project.service; enabled; vendor preset: enabled)
Active: active (running) since Wed 2023-12-27 21:00:09 UTC; 1h 22min ago
Main PID: 131051 (gunicorn)
Tasks: 7 (limit: 2353)
Memory: 207.9M
CPU: 6.629s
CGroup: /system.slice/my_project.service
├─131051 /home/user/my_project/venv/bin/python3 /home/user/my_project/venv/bin/gunicorn --workers 3 --bind unix:my_project.sock -m 007 wsgi:app
├─131052 /home/user/my_project/venv/bin/python3 /home/user/my_project/venv/bin/gunicorn --workers 3 --bind unix:my_project.sock -m 007 wsgi:app
├─131053 /home/user/my_project/venv/bin/python3 /home/user/my_project/venv/bin/gunicorn --workers 3 --bind unix:my_project.sock -m 007 wsgi:app
└─131054 /home/user/my_project/venv/bin/python3 /home/user/my_project/venv/bin/gunicorn --workers 3 --bind unix:my_project.sock -m 007 wsgi:app
Dec 27 21:00:09 my_project systemd[1]: Started my_project primary web server setup.
Dec 27 21:00:09 my_project gunicorn[131051]: [2023-12-27 21:00:09 +0000] [131051] [INFO] Starting gunicorn 21.2.0
Dec 27 21:00:09 my_project gunicorn[131051]: [2023-12-27 21:00:09 +0000] [131051] [INFO] Listening at: unix:my_project.sock (131051)
Dec 27 21:00:09 my_project gunicorn[131051]: [2023-12-27 21:00:09 +0000] [131051] [INFO] Using worker: sync
Dec 27 21:00:09 my_project gunicorn[131052]: [2023-12-27 21:00:09 +0000] [131052] [INFO] Booting worker with pid: 131052
Dec 27 21:00:09 my_project gunicorn[131053]: [2023-12-27 21:00:09 +0000] [131053] [INFO] Booting worker with pid: 131053
Dec 27 21:00:09 my_project gunicorn[131054]: [2023-12-27 21:00:09 +0000] [131054] [INFO] Booting worker with pid: 131054
Dec 27 21:52:46 my_project systemd[1]: my_project.service: Current command vanished from the unit file, execution of the command list won't be resumed.
Thanks in advance. Enjoy the holiday season!
This particular problem was solved by changing the my_project.service file on /etc/systemd/system to include the --chdir flag in the ExecStart line as follows:
ExecStart=/home/user/my_project/venv/bin/gunicorn --workers 3 --bind unix:/home/user/my_project/my_project.sock -m 007 --chdir /home/user/my_project wsgi:app
The solution was adopted from How to run gunicorn from a folder that is not the django project folder