The project is composed of :
- README.md
- create_iot_mqtt_docker_image.sh
- delete_iot_mqtt_docker_image.sh
- run_emqtt_container_node_1.sh
- run_emqtt_container_node_1.sh
- run_publisher_container.sh
- run_subscriber_container.sh
- docker-files
- Dockerfile
- modify_emq_conf_file.sh
- publisher.sh
- subscriber.sh
- persisters_configuration_file.json
- jars
- iot-mqtt-1.0-SNAPSHOT-jar-with-dependencies.jar
This project enables to run 4 docker containers.
On two of these containers (named s1-emqtt-io and s2-emqtt-io) we will setup a mqtt cluster. We will also run a tcpdump capture in order to visualize the packets exchanged with clients.
On one of the two other containers (named publisher), we will run a publisher to send messages to the mqtt cluster. On the other containers (named subsciber) we will run a subscriber with the possibility to persist messages on console or on elasticsearch. With Kibana we will be able to visualize those messages.
Execute the script create_iot_mqtt_docker_image.sh to create the image iot-mqtt-docker-image:0.0.1.
If you want to delete this image, you can run the script delete_iot_mqtt_docker_image.sh.
To setup a mqtt cluster we decide to use the feature EMQTT because it is open-source and has more functionnalities than others (see https://github.com/mqtt/mqtt.github.io/wiki/server-support)
On two terminals run the scripts run_emqtt_container_node_1.sh and run_emqtt_container_node_2.sh. Two containers should be named s1-emqtt-io and s2-emqtt-io.
Then on each container
cd /home
sh modify_emq_conf_file.sh
Check that the file /etc/emqx/emqx.conf has been changed at the end.
emqx start
emqx_ctl cluster status
Mosquitto is an open-source project that enables to publish and subscriber to any mqtt cluster.
on s1-emqtt-io
mosquitto_sub -h s2-emqtt-io -t iot_data
on s2-emqtt-io
mosquitto_pub -h s1-emqtt-io -t iot_data -m "Hello IoT"
You should see the message "Hello IoT" display on s1-emqtt-io
On subscriber
su - elasticuser
export JAVA_HOME=/usr/bin/java/jdk1.8.0_191
./elasticsearch/bin/elasticsearch > /dev/null 2>&1 &
exit
curl -XPUT 'http://localhost:9200/index_test/messages/first' -H "Content-Type: application/json" -d '{"name" : "xebia & iot-ee"}'
on subscriber
/usr/share/kibana/bin/kibana > /dev/null 2>&1 &
On your local desktop, open a browser and go to
localhost:5601
Verify index "index_test" was created
On two terminals, execute run_publisher_container.sh and run_subscriber_container.sh. Two containers should be named publisher and subscriber.
On publisher
cd /home
sh publisher.sh tcp://s2-emqtt-io:1883 iot_data_test
On subscriber
cd /home
sh subscriber.sh tcp://s1-emqtt-io:1883 iot_data_test persisters_configuration_file.json
PS: Link followed https://github.com/emqtt/emqttd/issues/1274
curl -v --basic -u admin:public -H "Content-Type: application/json" -d '{"topic": "test","payload": "hello","qos": 1,"retain": false,"client_id": "C_1492145414740"}' -k http://localhost:8080/api/v2/mqtt/publish
Let's analyze the packets exchanged between clients and the MQTT cluster. You can execute on s1-emqtt-io, the following command:
tcpdump -w capture_network_exchange.pcap
With all methods we have seen, publish and subscribe messages from s1-emqtt-io, then retrieve the file capture_network_exchange.pcap on your local desktop by executing the command below on a new terminal:
docker cp <CONTAINER_ID_of_s1-emqtt-io>:/capture_network_exchange.pcap .
The file format pcap is readable with wireshark, thus download wireshark and open it. Finally, analyse the packet exchanged.