This project is no longer maintained. Please take a look at the following projects:
This project contains sources and tools to record flight data via MQTT and to store it in a PostgreSQL database.
- a mosquitto instance listens for sensor data, the MQTT protocol is used
- a Apache Camel instance listens for data on mosquitto and writes it into PostgreSQL. In addition, this instance provides a simple REST endpoint at path
/api
that can be used to read the data stored in PostgreSQL (experimental) - a PostgreSQL instance where data is written to by Apache Camel
These instructions will get you a copy of this mono repository and prepare it for development. Please be aware that the provided implementation and configuration is yet not suitable for productive deployment (work in progress).
The project supports Docker Compose and provides a docker-compose.yml
that can be used to start all required parts of the application.
To build the projects in the repository, the following prerequisites must be met:
- Docker
- Docker Compose
- Java JDK 11 or higher (e.g. OpenJDK) and Maven 3.x if you want to build and launch the integration pipeline based on Apache Camel locally (not strictly necessary but nice to have for development)
Get a copy of the repository, e.g. by cloning it from the following location:
git clone https://github.com/gappc/noi-flightdata.git
Change directory:
cd noi-flightdata/
Make a copy of the env.example
file named .env
. The .env
file contains the configuration and can be adjusted by your needs:
Note: the configuration in
env.example
allows anonymous MQTT access. If you want to enable MQTT authentication, use the fileenv-with-auth.example
instead and refer to the MQTT authentication for further information.
cp env.example .env
Start docker-compose
:
docker-compose up
Docker Compose should now start three containers:
- mosquitto (MQTT)
- postgres (Database)
- integrator (Apache Camel)
Note that the first launch may take some time, because the Docker images must be pulled and the
integrator
container has to download all necessary Maven dependencies.
To subscribe to all topics on the running mosquitto
instance, use the script mosquitto-docker-local-sub.sh
located inside the scripts
folder:
./scripts/mosquitto-docker-local-sub.sh
To publish a message to the flightdata
topic on the mosquitto
instance, use the script mosquitto-docker-local-pub.sh
located inside the scripts
folder. This script takes the message payload as its argument. Please be sure to provide valid JSON as payload, because PostgreSQL expects JSON as payload using its JSONB
datatype:
./scripts/mosquitto-docker-local-pub.sh '{"text": "Hello World!"}'
Data that is published to the MQTT broker should now be stored in the PostgreSQL instance. Use any database client you want to interact with PostgreSQL.
The integrator
container provides also a simple REST endpoint to read the data stored in PostgreSQL. You can access it at port 8080 using the path /api
, e.g. http://localhost:8080/api.
Note that the
/api
endpoint returns at most 100 results only.
If you use docker-compose
you can configure MQTT and PostgreSQL properties using environment variables. The environment variables may also be provided by a .env
file.
The project provides two environment example files that can be used as base configuration:
- copy the env.example file to
.env
to use MQTT without authentication - copy the env-with-auth.example file to
.env
to use MQTT with authentication
A description of the environment variables can be found in the example files mentioned above.
This README.md mostly describes a project usage where MQTT allows anonymous access. This is fine for development purpose, but should not be done in production environments. That's why the project can also be configured with MQTT authentication enabled.
If you take a look at env-with-auth.example you will find an environment configuration that enables MQTT authentication. Copy that file to .env
to use it as base for your own configuration with MQTT authentication enabled:
cp env-with-auth.example .env
The environment configuration defines mosquitto-with-auth.conf to be used as mosquitto config file using the MOSQUITTO_CONF
environment variable. This mosquitto config contains a property password_file
that references the password file passwordfile.example as it is mounted inside the docker container (take a look at the docker-compose.yml file to see how the mount points are defined).
The password file defines one example user user1
and its hashed password password1
Hint: https://mosquitto.org/documentation/authentication-methods/ provides more information about the password file specification and tooling.
Use the script
mosquitto-docker-local-update-passwordfile.sh
located inside thescripts
folder to hash the passwords of a plain-text password file in place. Take a look atmosquitto-docker-local-update-passwordfile.sh
description in the Scripts section for more information. Attention: the script does not check if the passwords are already hashed. If you apply the script several times to a password file, each time the passwords will be rehashed in place, rendering them useless.
Since MQTT is now password protected, the integrator
must also authenticate against the MQTT in order to be able to subscribe to topics. Set the INTEGRATOR_MQTT_USER
and INTEGRATOR_MQTT_PASS
environment variables (e.g. in your .env
file) to a user/password combination that is defined by your password file. If you use the provided env-with-auth.example
file you will see, that INTEGRATOR_MQTT_USER
is set to user1
and INTEGRATOR_MQTT_PASS
is set to password1
, which are the values defined in the password file and were mentioned above.
From here on, your MQTT instance has authentication enabled and your integrator is configured with the according credentials. Modify the password file as you want, but please keep in mind that a valid user / password combination must be defined for INTEGRATOR_MQTT_USER
and INTEGRATOR_MQTT_PASS
.
You can configure the MQTT further by adjusting the mosquitto config file that is mounted by docker-compose
. Set the MOSQUITTO_CONF
environment variable to define which mosquitto config file to use.
For development purpose, you can start any of the Docker Compose containers individually, e.g. docker-compose up mosquitto
.
Usually you want to work on the integrator
. The integrator reads from a MQTT broker (mosquitto) and writes to a PostgreSQL database. A common setup for development is therefor to launch mosquitto and PostgreSQL in their respective Docker containers and to start on integrator
locally. This way you can e.g. easily debug the integrator. To get the described setup, follow the steps below.
Start mosquitto
and postgres
containers using docker-compose:
docker-compose up mosquitto postgres
Run the integrator locally in dev mode (please note that you need to have Java 11 and Maven 3.x installed):
mvn -f ./integrator/pom.xml compile quarkus:dev
This should get you started with development.
Note that you don't have to start the
integrator
locally for development. You can start it as container together with the other containers withdocker-compose up
as described in the Installing section. The only drawback is, that it is not that easy to attach a debugger to the integrator inside the Docker container.
This section provides an overview of the available scripts that come with this project.
The script mosquitto-docker-local-pub.sh uses the official Docker image eclipse-mosquitto to start a local Docker container named mosquitto_pub
that can be used to publish messages to a local MQTT broker. The script expects one argument which is the payload. The payload is then published to the local MQTT broker on topic test/topic
.
Hint: Use mosquitto-docker-local-start.sh to start a local MQTT broker.
The script mosquitto-docker-local-sub.sh uses the official Docker image eclipse-mosquitto to start a local Docker container named mosquitto_pub
that can be used to subscribe to a local MQTT broker on all topics.
Hint: Use mosquitto-docker-local-start.sh to start a local MQTT broker.
The script mosquitto-docker-sub-to-official-test-server.sh uses the official Docker image eclipse-mosquitto to start a local Docker container named mosquitto-sub-test.mosquitto.org
that subscribes to all topics on the server test.mosquitto.org
. No ports are exposed. The output is written to stdout.
The reason for the script to exist is that it provides an easy start point to play around with MQTT. The MQTT broker at test.mosquitto.org
should be online and pusblish data 24 / 7.
The script mosquitto-docker-local-update-passwordfile.sh uses the official Docker image eclipse-mosquitto to start a local Docker container named mosquitto-update-password
. The script uses mosquitto_passwd
under the hood to hash the passwords of the provided password file in place. The script expects the fully qualified path to the password file as parameter.
Attention: the script does not check if the passwords are already hashed. If you apply the script several times to a password file, each time the passwords will be rehashed, rendering them useless.
Below you find an example contents of a plain-text password file whose passwords are hashed by the script:
user1:password1
Note: the password file may contain several users, each one on a new line.
When you apply the script to that file, its outcome may look as follows:
user1:$7$101$cNdSny7311rTLYeB$SSRWvaddDft6AnqSEtwfNoIQRVMOL2vlkPJWmlHMSYwbRQF4HZenS/CZO2H8K6tKl8tXDnJmPLaIVHLGfR+7YA==
You can see that the password was hashed. This file is now a valid mosquitto
password file.
This script triggers a mosquitto instance running in a Docker container named noi-flightdata_mosquitto_1
(as it is the case when the Docker containers are started with docker-compose
) to reload its configuration. This way, configuration changes become effective without restarting mosquitto (during a restart no messages can be received).
For support, please contact Christian Gapp.
If you'd like to contribute, please fork the repository and use a feature branch. Pull requests are warmly welcome.
This project uses SemVer for versioning. For the versions available, see the tags on this repository.
The code in this project is licensed under the MIT license. See the LICENSE file for more information.
- Christian Gapp - Initial work - gappc