- What is your original issue/pain point you want to solve?
The postinstall package script restarts nginx by restarting it directly:
echo "# Restarting webserver ${WEBSERVER_CMD}"
${INIT_CMD} restart ${WEBSERVER_CMD}
Nginx has a better reloading method, nginx -s reload, see Controlling NGINX Processes at Runtime | NGINX Documentation
- Which are one or two concrete situations where this problem hurts the most?
Other services running on the same service having its connections severed during a package upgrade.
- Why is it not solvable with the Zammad standard?
It’s an issue with how the script restarts nginx.
- What is your expectation/what do you want to achieve?
If possible, reload the nginx config via nginx -s reload or give some other control of this behaviour, possibly even allowing it to be disabled so one can manually reload nginx as needed.