This document provides a clear and informative overview of Multilauncher, a client-server application for remotely executing commands on your Raspberry Pi.
Multilauncher allows you to run scripts or commands on your Raspberry Pi by sending custom messages to a message broker (MQTT or RabbitMQ) or a network server (TCP or UDP). This enables you to manage your devices remotely without needing a direct connection.
By default, Multilauncher operates in MQTT mode, making it ideal for managing devices remotely. It utilizes an MQTT listener to receive specific messages on a designated topic. Upon receiving the message, it triggers the execution of a predefined command.
Imagine connecting Multilauncher on your Raspberry Pi to a free public MQTT broker like HiveMQ. You can then send a specific message from another device (like a smartphone with MQTT Dashboard for Android) to trigger an action on your home server (hidden behind a NAT firewall). This could include powering on the server, starting a container, or launching a virtual machine.
Multilauncher requires Python 3.9 and Bash to run and can be installed on Linux/Unix environments (including Windows under WSL). Its memory footprint is light, allowing it to function even on older Raspberry Pi models with limited RAM.
It's highly recommended to run Multilauncher within a Python virtual environment (VENV) for better isolation and dependency management. The instructions provide steps to install pyenv and switch to Python 3.9. It then guides you through creating and activating a virtual environment within the project directory.
- Install pyenv and switch to Python 3.9
- Move into setup/ folder and run build-environment.sh script to build environment
The config folder houses various configuration files:
- common_conf.py: Defines shared parameters like enabled listeners and message prefixes.
- mqtt_conf.py: Defines MQTT connection details like broker address, port, encryption, authentication, and topics (example provided for HiveMQ with encryption and authentication enabled).
- rabbitmq_conf.py: Defines RabbitMQ connection information like host and port.
- tcp_conf.py: Defines TCP server parameters like port and host.
- udp_conf.py: Defines UDP server configuration (port and host).
The document details how to configure mqtt_conf.py using HiveMQ settings, including creating an account and enabling TLS encryption and authentication.
The launcher folders contain JSON files for defining custom commands. The provided example showcases the structure and details of each attribute:
{
"name": "My-NAS",
"description": "Start/Stop NAS",
"start_message": "start nas",
"stop_message": "suspend nas",
"can_be_stopped": true,
"start_command": "etherwake -i wlan0 FF:FF:FF:FF:FF:FF",
"stop_command": "curl -k --user admin:password https://192.168.1.3/api/suspend",
"multiple_instances": false,
"enabled": true
}
- name: Unique identifier for the service.
- description: Brief explanation of the command's purpose.
- start_message: The message received from the listener (preceded by a prefix) that triggers the start_command.
- stop_message (optional): The message (preceded by a prefix) that triggers the stop_command.
- can_be_stopped: Boolean indicating whether the service can be stopped using the stop_command.
- start_command: The command to launch the service (e.g., etherwake to power on a NAS).
- stop_command: The command to stop the service (e.g., a curl command to call a suspend API).
- multiple_instances: Boolean indicating if multiple service instances can run concurrently.
- enabled: Boolean indicating if the service is available for use.
For Debian-based systems like Raspbian and Ubuntu, the script setup/install-as-service.sh allows you to install Multilauncher as a system service. This utilizes the systemctl command and requires root privileges. Remember to activate the virtual environment before running the script.
Use the setup/uninstall.sh script (run as root) to remove Multilauncher from your system.
Multilauncher is released under the Apache2 License.