![\"GitHub\"](\"https://fiware.github.io/tutorials.Historic-Context-NIFI/icon/GitHub-Mark-32px.png\"){kind=icon}
FIWARE 302: Persisting Context Data NIFIThis tutorial is an introduction to FIWARE Draco - a generic enabler which is used to persist context data into third-party databases creating a historical view of the context. The tutorial activates the IoT sensors connected in the previous tutorial and persists measurements\nfrom those sensors into a database for further analysis.
\nThe docker-compose files for this tutorial can be found on GitHub:
FIWARE 302: Persisting Context Data NIFI
\n\n\"Plots within plots, but all roads lead down the dragon’s gullet.\"
\n— George R.R. Martin (A Dance With Dragons)
\n
Previous tutorials have introduced a set of IoT Sensors (providing measurements of the\nstate of the real world), and two FIWARE Components - the Orion Context Broker and an IoT Agent. \nThis tutorial will introduct a new data persistance component - FIWARE Draco.
\nThe system so far has been built up to handle the current context, in other words it holds the data entities\ndefining the state of the real-world objects at a given moment in time.
\nFrom this definition you can see - context is only interested in the current state of the system\nIt is not the responsibility of any of the existing components to report on the historical state of the system,\nthe context is based on the last measurement each sensor has sent data to the context broker.
\nIn order to do this, we will need to extend the existing architecture to persist changes of state into a database whenever \nthe context is updated.
\nPersisting historical context data is useful for big data analysis - it can be used to discover trends, or data \ncan be sampled and aggregated to remove the influence of outlying data measurements. However within each Smart Solution,\nthe significance of each entity type will differ and entities and attributes may need to be sampled at different rates.
\nSince the business requirements for using context data differ from application to appliation, there is no one standard use \ncase for historical data persistence - each situation is unique - it is not the case that one size fits all.\nTherefore rather than overloading the context broker with the job of historical context data persistence, this role has been\nseparated out into a separate, highly configurable component - Draco.
\nAs you would expect, Draco, as part of an Open Source platform, is technology agnostic regarding the database \nto be used for data persistance. The database you choose to use will depend upon your own business needs.
\nHowever there is a cost to offering this flexibility - each part of the system must be separately configured and\nnotifications must be set up to only pass the minimal data required as necessary.
\nFor the purpose of this tutorial, a series of dummy IoT devices have been created, which will be attached to the context broker.\nThe state of each device can be seen on the UltraLight device monitor web-page found at: http://localhost:3000/device/monitor
![\"FIWARE](\"https://fiware.github.io/tutorials.Historic-Context-NIFI/img/device-monitor.png\")
This application builds on the components and dummy IoT devices created in \nprevious tutorials. It will make use of three FIWARE components - \nthe Orion Context Broker, the\nIoT Agent for UltraLight 2.0 and introduce the\nDraco Generic Enabler for persisting context data to a database.\nAdditional databases are now involved - both the Orion Context Broker and the IoT Agent rely on MongoDB technology to keep persistence of the information they hold, and we will be persisting our historical context data another database - either MySQL , PostgreSQL or Mongo-DB database.
\nTherefore the overall architecture will consist of the following elements:
\nSince all interactions between the elements are initiated by HTTP requests, the entities can be containerized and run from exposed ports.
\nThe specific architecture of each section of the tutorial is discussed below.
\nBefore you start you should ensure that you have obtained or built the necessary Docker images locally. Please run
\n./services create\n\n\nNote The
\ncontext-providerimage has not yet been pushed to Docker hub.\nFailing to build the Docker sources before proceeding will result in the following error:Pulling context-provider (fiware/cp-web-app:latest)...\nERROR: The image for the service you're trying to recreate has been removed.\n
Thereafter, all services can be initialised from the command line by running the services Bash script provided within the repository:
\n./services <command>\nWhere <command> will vary depending upon the databases we wish to activate.\nThis command will also import seed data from the previous tutorials and provision the dummy IoT sensors on startup.
\n\n","schema":"https://schema.getpostman.com/json/collection/v2.0.0/collection.json","toc":[{"content":"Data Persistence","slug":"data-persistence"},{"content":"Architecture","slug":"architecture"},{"content":"Start Up","slug":"start-up"}],"owner":"513743","collectionId":"33990d83-5855-48cb-b4ed-48fb14fb0a6b","publishedId":"S17qTpnL","public":true,"customColor":{"top-bar":"FFFFFF","right-sidebar":"303030","highlight":"233C68"},"publishDate":"2020-01-02T11:06:41.000Z"},"item":[{"name":"Mongo DB - Persisting Context","item":[{"name":"Draco - Check System Health","id":"30081785-d630-4b5e-86eb-14fec18334e9","protocolProfileBehavior":{"disableBodyPruning":true},"request":{"method":"GET","header":[],"url":"http://localhost:9090/nifi-api/system-diagnostics","description":":information_source: Note: If you want to clean up and start over again you can do so with the following command:
\n\n./services stop\n
Once Draco is running, you can check the status by making an HTTP request to the exposed draco port to /nifi-api/system-diagnostics. If the response is blank, this is usually because Draco is not running or is listening on another port.
Once a dynamic context system is up and running, we need to inform Draco of changes in context.
\nThis is done by making a POST request to the /v2/subscription endpoint of the Orion Context Broker.
fiware-service and fiware-servicepath headers are used to filter the subscription to only listen to measurements from the attached IoT SensorsidPattern in the request body ensures that Draco will be informed of all context data changes.url must match the configured NIFI_WEB_HTTP_PORT and the URL the sync is listening on (usually /v2/notify)throttling value defines the rate that changes are sampled.As you can see, the database used to persist context data has no impact on the details of the subscription. It is the same for each database.
\n","urlObject":{"protocol":"http","path":["v2","subscriptions",""],"host":["localhost:1026"],"query":[],"variable":[]}},"response":[],"_postman_id":"9758eb6b-ce4d-4373-b779-a32e4ecbf1b0"},{"name":"Orion - Check Subscription is working","id":"fc0179d3-2235-4135-9045-896f6daa0d77","protocolProfileBehavior":{"disableBodyPruning":true},"request":{"method":"GET","header":[{"key":"fiware-service","value":"openiot"},{"key":"fiware-servicepath","value":"/"}],"url":"http://localhost:1026/v2/subscriptions/","description":"If a subscription has been created, you can check to see if it is firing by making a GET \nrequest to the /v2/subscriptions endpoint.
Within the notification section of the response, you can see several additional attributes which describe the health of the subscription
If the criteria of the subscription have been met, timesSent should be greater than 0.\nA zero value would indicate that the subject of the subscription is incorrect or the subscription \nhas created with the wrong fiware-service-path or fiware-service header
The lastNotification should be a recent timestamp - if this is not the case, then the devices\nare not regularly sending data. Remember to unlock the Smart Door and switch on the Smart Lamp
The lastSuccess should match the lastNotification date - if this is not the case \nthen Draco is not receiving the subscription properly. Check that the host name\nand port are correct.
Finally, check that the status of the subscription is active - an expired subscription\nwill not fire.
Persisting historic context data using MongoDB technology is relatively simple to configure since we are already using a\nMongoDB instance to hold data related to the Orion Context Broker and the IoT Agent. The MongoDB instance is listening\non the standard 27017 port and the overall architecture can be seen below:
![](\"https://fiware.github.io/tutorials.Historic-Context-NIFI/img/mongo-draco-tutorial.png\")
mongo-db:\n image: mongo:3.6\n hostname: mongo-db\n container_name: db-mongo\n ports:\n - \"27017:27017\"\n networks:\n - default\n command: --bind_ip_all --smallfiles\ndraco:\n image: ging/fiware-draco:1.1.0\n container_name: draco\n depends_on:\n - mongo-db\n environment:\n - NIFI_WEB_HTTP_PORT=9090\n ports:\n - \"9090:9090\"\n - \"5050:5050\"\n healthcheck:\n test: curl --fail -s http://localhost:9090/nifi-api/system-diagnostics || exit 1\nThe draco container is listening on two ports:
5050 is where the service will be listening for notifications from the Orion\ncontext broker9090 is exposed purely for configuring the processorsTo start the system with a MongoDB database only, run the following command:
\n./services mongodb\nThen go to your browser and open Draco using this URL http://localhost:9090/nifi
Now go to the Components toolbar which is placed in the upper section of the NiFi GUI, find the template icon and drag\nand drop it inside the Draco user space. At this point, a popup should be displayed with a list of all the templates\navailable. Please select the template MONGO-TUTORIAL.
\n![](\"https://fiware.github.io/tutorials.Historic-Context-NIFI/img/mongo-tutorial-template.png\")
Select all the processors (press shift and click on every processor) and start them by clicking on the start button.\nNow, you can see that the status icon of each processor turned from red to green.
\nFor the purpose of this tutorial, we must be monitoring a system where the context is periodically being updated.\nThe dummy IoT Sensors can be used to do this. Open the device monitor page at http://localhost:3000/device/monitor\nand unlock a Smart Door and switch on a Smart Lamp. This can be done by selecting an appropriate the command \nfrom the drop down list and pressing the send button. The stream of measurements coming from the devices can then\nbe seen on the same page:
![](\"https://fiware.github.io/tutorials.Historic-Context/img/door-open.gif\")
To read mongo-db data from the command line, we will need access to the mongo tool run an interactive instance\nof the mongo image as shown to obtain a command line prompt:
docker run -it --network fiware_default --entrypoint /bin/bash mongo\nYou can then log into to the running mongo-db database by using the command line as shown:
mongo --host mongo-db\nTo show the list of available databases, run the statement as shown:
\nshow dbs\nadmin 0.000GB\niotagentul 0.000GB\nlocal 0.000GB\norion 0.000GB\norion-openiot 0.000GB\nsth_openiot 0.000GB\nThe result include two databases admin and local which are set up by default by MongoDB, along with four databases\ncreated by the FIWARE platform. The Orion Context Broker has created two separate database instance for each fiware-service
fiware-service and therefore are held within the orion database,\nwhereas the IoT device entities were created using the openiot fiware-service header and are held separately. The IoT Agent was intialized to hold the IoT sensor data in a separate MongoDB database called iotagentul.As a result of the subscription of Draco to Orion Context Broker, a new database has been created called sth_openiot. The default value for a Mongo DB database holding historic context consists of the sth_ prefix followed by the fiware-service header - therefore sth_openiot holds the historic context of the IoT devices.
use sth_openiot\nshow collections\nswitched to db sth_openiot\n\nsth_/_Door:001_Door\nsth_/_Door:001_Door.aggr\nsth_/_Lamp:001_Lamp\nsth_/_Lamp:001_Lamp.aggr\nsth_/_Motion:001_Motion\nsth_/_Motion:001_Motion.aggr\nLooking within the sth_openiot you will see that a series of tables have been created. The names of each table consist\nof the sth_ prefix followed by the fiware-servicepath header followed by the entity id. Two table are created for\neach entity - the .aggr table holds some aggregated data which will be accessed in a later tutorial. The raw data\ncan be seen in the tables without the .aggr suffix.
The historical data can be seen by looking at the data within each table, by default each row will contain the sampled value of a single attribute.
\ndb[\"sth_/_Door:001_Door\"].find().limit(10)\n{ \"_id\" : ObjectId(\"5b1fa48630c49e0012f7635d\"), \"recvTime\" : ISODate(\"2018-06-12T10:46:30.897Z\"), \"attrName\" : \"TimeInstant\", \"attrType\" : \"ISO8601\", \"attrValue\" : \"2018-06-12T10:46:30.836Z\" }\n{ \"_id\" : ObjectId(\"5b1fa48630c49e0012f7635e\"), \"recvTime\" : ISODate(\"2018-06-12T10:46:30.897Z\"), \"attrName\" : \"close_status\", \"attrType\" : \"commandStatus\", \"attrValue\" : \"UNKNOWN\" }\n{ \"_id\" : ObjectId(\"5b1fa48630c49e0012f7635f\"), \"recvTime\" : ISODate(\"2018-06-12T10:46:30.897Z\"), \"attrName\" : \"lock_status\", \"attrType\" : \"commandStatus\", \"attrValue\" : \"UNKNOWN\" }\n{ \"_id\" : ObjectId(\"5b1fa48630c49e0012f76360\"), \"recvTime\" : ISODate(\"2018-06-12T10:46:30.897Z\"), \"attrName\" : \"open_status\", \"attrType\" : \"commandStatus\", \"attrValue\" : \"UNKNOWN\" }\n{ \"_id\" : ObjectId(\"5b1fa48630c49e0012f76361\"), \"recvTime\" : ISODate(\"2018-06-12T10:46:30.836Z\"), \"attrName\" : \"refStore\", \"attrType\" : \"Relationship\", \"attrValue\" : \"Store:001\" }\n{ \"_id\" : ObjectId(\"5b1fa48630c49e0012f76362\"), \"recvTime\" : ISODate(\"2018-06-12T10:46:30.836Z\"), \"attrName\" : \"state\", \"attrType\" : \"Text\", \"attrValue\" : \"CLOSED\" }\n{ \"_id\" : ObjectId(\"5b1fa48630c49e0012f76363\"), \"recvTime\" : ISODate(\"2018-06-12T10:45:26.368Z\"), \"attrName\" : \"unlock_info\", \"attrType\" : \"commandResult\", \"attrValue\" : \" unlock OK\" }\n{ \"_id\" : ObjectId(\"5b1fa48630c49e0012f76364\"), \"recvTime\" : ISODate(\"2018-06-12T10:45:26.368Z\"), \"attrName\" : \"unlock_status\", \"attrType\" : \"commandStatus\", \"attrValue\" : \"OK\" }\n{ \"_id\" : ObjectId(\"5b1fa4c030c49e0012f76385\"), \"recvTime\" : ISODate(\"2018-06-12T10:47:28.081Z\"), \"attrName\" : \"TimeInstant\", \"attrType\" : \"ISO8601\", \"attrValue\" : \"2018-06-12T10:47:28.038Z\" }\n{ \"_id\" : ObjectId(\"5b1fa4c030c49e0012f76386\"), \"recvTime\" : ISODate(\"2018-06-12T10:47:28.081Z\"), \"attrName\" : \"close_status\", \"attrType\" : \"commandStatus\", \"attrValue\" : \"UNKNOWN\" }\nThe usual Mongo-DB query syntax can be used to filter appropriate fields and values. For example to read the rate at which the Motion Sensor with the id=Motion:001_Motion is accumulating, you would make a query as follows:
db[\"sth_/_Motion:001_Motion\"].find({attrName: \"count\"},{_id: 0, attrType: 0, attrName: 0 } ).limit(10)\n{ \"recvTime\" : ISODate(\"2018-06-12T10:46:18.756Z\"), \"attrValue\" : \"8\" }\n{ \"recvTime\" : ISODate(\"2018-06-12T10:46:36.881Z\"), \"attrValue\" : \"10\" }\n{ \"recvTime\" : ISODate(\"2018-06-12T10:46:42.947Z\"), \"attrValue\" : \"11\" }\n{ \"recvTime\" : ISODate(\"2018-06-12T10:46:54.893Z\"), \"attrValue\" : \"13\" }\n{ \"recvTime\" : ISODate(\"2018-06-12T10:47:00.929Z\"), \"attrValue\" : \"15\" }\n{ \"recvTime\" : ISODate(\"2018-06-12T10:47:06.954Z\"), \"attrValue\" : \"17\" }\n{ \"recvTime\" : ISODate(\"2018-06-12T10:47:15.983Z\"), \"attrValue\" : \"19\" }\n{ \"recvTime\" : ISODate(\"2018-06-12T10:47:49.090Z\"), \"attrValue\" : \"23\" }\n{ \"recvTime\" : ISODate(\"2018-06-12T10:47:58.112Z\"), \"attrValue\" : \"25\" }\n{ \"recvTime\" : ISODate(\"2018-06-12T10:48:28.218Z\"), \"attrValue\" : \"29\" }\nTo leave the MongoDB client and leave interactive mode, run the following:
\nexit\nexit\nOnce Draco is running, you can check the status by making an HTTP request to the exposed draco port to /nifi-api/system-diagnostics. If the response is blank, this is usually because Draco is not running or is listening on another port.
Once a dynamic context system is up and running, we need to inform Draco of changes in context.
\nThis is done by making a POST request to the /v2/subscription endpoint of the Orion Context Broker.
fiware-service and fiware-servicepath headers are used to filter the subscription to only listen to measurements from the attached IoT SensorsidPattern in the request body ensures that Draco will be informed of all context data changes.url must match the configured NIFI_WEB_HTTP_PORT and the URL the sync is listening on (usually /v2/notify)throttling value defines the rate that changes are sampled.As you can see, the database used to persist context data has no impact on the details of the subscription. It is the same for each database.
\n","urlObject":{"protocol":"http","path":["v2","subscriptions",""],"host":["localhost:1026"],"query":[],"variable":[]}},"response":[],"_postman_id":"2dc3dc3c-a4c5-4984-a6ac-eaf3f0983f6d"},{"name":"Orion - Check Subscription is working","id":"9b69e477-aa81-46b4-9d74-a15ecadebada","protocolProfileBehavior":{"disableBodyPruning":true},"request":{"method":"GET","header":[{"key":"fiware-service","value":"openiot"},{"key":"fiware-servicepath","value":"/"}],"url":"http://localhost:1026/v2/subscriptions/","description":"If a subscription has been created, you can check to see if it is firing by making a GET \nrequest to the /v2/subscriptions endpoint.
Within the notification section of the response, you can see several additional attributes which describe the health of the subscription
If the criteria of the subscription have been met, timesSent should be greater than 0.\nA zero value would indicate that the subject of the subscription is incorrect or the subscription \nhas created with the wrong fiware-service-path or fiware-service header
The lastNotification should be a recent timestamp - if this is not the case, then the devices\nare not regularly sending data. Remember to unlock the Smart Door and switch on the Smart Lamp
The lastSuccess should match the lastNotification date - if this is not the case \nthen Draco is not receiving the subscription properly. Check that the host name\nand port are correct.
Finally, check that the status of the subscription is active - an expired subscription\nwill not fire.
To persist historic context data into an alternative database such as PostgreSQL, we will need an additional\ncontainer which hosts the PostgreSQL server - the default Docker image for this data can be used. The PostgreSQL\ninstance is listening on the standard 5432 port and the overall architecture can be seen below:
![](\"https://fiware.github.io/tutorials.Historic-Context-NIFI/img/postgres-draco-tutorial.png\")
We now have a system with two databases, since the MongoDB container is still required to hold data related to the Orion\nContext Broker and the IoT Agent.
\npostgres-db:\n image: postgres:latest\n hostname: postgres-db\n container_name: db-postgres\n expose:\n - \"5432\"\n ports:\n - \"5432:5432\"\n networks:\n - default\n environment:\n - \"POSTGRES_PASSWORD=password\"\n - \"POSTGRES_USER=postgres\"\n - \"POSTGRES_DB=postgres\"\nThe postgres-db container is listening on a single port:
5432 is the default port for a PostgreSQL server. It has been exposed so you can also run the pgAdmin4 tool\nto display database data if you wishThe postgres-db container is driven by environment variables as shown:
\n\n:information_source: Note: Passing the Username and Password in plain text environment variables like this is a\nsecurity risk. Whereas this is acceptable practice in a tutorial, for a production environment, you can avoid this\nrisk by applying Docker Secrets
\n
draco:\n image: ging/fiware-draco:1.1.0\n container_name: draco\n depends_on:\n - postgres-db\n environment:\n - NIFI_WEB_HTTP_PORT=9090\n ports:\n - \"9090:9090\"\n - \"5050:5050\"\n healthcheck:\n test: curl --fail -s http://localhost:9090/nifi-api/system-diagnostics || exit 1\nThe draco container is listening on two ports:
5050 is where the service will be listening for notifications from the Orion\ncontext broker9090 is exposed purely for configuring the processors.To start the system with a PostgreSQL database run the following command:
\n./services postgres\nThen go to your browser and open Draco using this URL http://localhost:9090/nifi
Now go to the Components toolbar which is placed in the upper section of the NiFi GUI, find the template icon and drag\nand drop it inside the Draco user space. At this point, a popup should be displayed with a list of all the templates\navailable. Please select the template POSTGRESQL-TUTORIAL.
\n![](\"https://fiware.github.io/tutorials.Historic-Context-NIFI/img/postgres-tutorial-template.png\")
Before starting the processors, you need to set your PostgreSQL password and enable the DBCConnectionPool controller.\nFor doing that please follow the instructions:
\nDo right click on any part of the Draco GUI user space, and then click on configure.\n![](\"https://fiware.github.io/tutorials.Historic-Context-NIFI/img/step1.png\")
Go to the Controller Services Tab, at this point a list of controllers should be displayed, locate the\nDBCConnectionPool controller.
\nClick on the configuration button of the \"DBCPConnectionPool\"\n![](\"https://fiware.github.io/tutorials.Historic-Context-NIFI/img/step2.png\")
Go to the controller Properties tab and put \"password\" in the password field, then apply the changes.\n![](\"https://fiware.github.io/tutorials.Historic-Context-NIFI/img/controller-postgresql.png\")
Enable the processor by clicking on the thunder icon and then click on enable, then close the controller\nconfiguration page.
\n![](\"https://fiware.github.io/tutorials.Historic-Context-NIFI/img/step4.png\")
![](\"https://fiware.github.io/tutorials.Historic-Context-NIFI/img/step5.png\")
To start the system with a PostgreSQL database run the following command:
\n./services postgres\npostgres-db:\n image: postgres:latest\n hostname: postgres-db\n container_name: db-postgres\n expose:\n - \"5432\"\n ports:\n - \"5432:5432\"\n networks:\n - default\n environment:\n - \"POSTGRES_PASSWORD=password\"\n - \"POSTGRES_USER=postgres\"\n - \"POSTGRES_DB=postgres\"\n\nThe postgres-db container is listening on a single port:
5432 is the default port for a PostgreSQL server. It has been exposed so you can also run the pgAdmin4 tool to display database data if you wishThe postgres-db container is driven by environment variables as shown:
draco:\n image: fiware/draco-ngsi:latest\n hostname: draco\n container_name: fiware-draco\n networks:\n - default\n depends_on:\n - postgres-db\n expose:\n - \"5080\"\n ports:\n - \"5050:5050\"\n - \"5080:5080\"\n environment:\n - \"CYGNUS_POSTGRESQL_HOST=postgres-db\"\n - \"CYGNUS_POSTGRESQL_PORT=5432\"\n - \"CYGNUS_POSTGRESQL_USER=postgres\" \n - \"CYGNUS_POSTGRESQL_PASS=password\" \n - \"CYGNUS_LOG_LEVEL=DEBUG\"\n - \"CYGNUS_SERVICE_PORT=5050\"\n - \"CYGNUS_API_PORT=5080\"\n - \"CYGNUS_POSTGRESQL_ENABLE_CACHE=true\"\nThe draco container is listening on two ports:
5050 for notifications from the Orion context broker5080 is exposed purely for tutorial access - so that cUrl or Postman can make provisioning commands\nwithout being part of the same network.For the purpose of this tutorial, we must be monitoring a system where the context is periodically being updated.\nThe dummy IoT Sensors can be used to do this. Open the device monitor page at http://localhost:3000/device/monitor\nand unlock a Smart Door and switch on a Smart Lamp. This can be done by selecting an appropriate the command \nfrom the drop down list and pressing the send button. The stream of measurements coming from the devices can then\nbe seen on the same page:
To read PostgreSQL data from the command line, we will need access to the postgres client, to do this, run an\ninteractive instance of the postgresql-client image supplying the connection string as shown to obtain a command \nline prompt:
docker run -it --rm --network fiware_default jbergknoff/postgresql-client \\\n postgresql://postgres:password@postgres-db:5432/postgres\nTo show the list of available databases, run the statement as shown:
\n\\list\n Name | Owner | Encoding | Collate | Ctype | Access privileges \n-----------+----------+----------+------------+------------+-----------------------\n postgres | postgres | UTF8 | en_US.utf8 | en_US.utf8 | \n template0 | postgres | UTF8 | en_US.utf8 | en_US.utf8 | =c/postgres +\n | | | | | postgres=CTc/postgres\n template1 | postgres | UTF8 | en_US.utf8 | en_US.utf8 | =c/postgres +\n | | | | | postgres=CTc/postgres\n(3 rows)\nThe result include two template databases template0 and template1 as well as the postgres database setup when the docker container was started.
To show the list of available schemas, run the statement as shown:
\n\\dn\n List of schemas\n Name | Owner \n---------+----------\n openiot | postgres\n public | postgres\n(2 rows)\nAs a result of the subscription of Draco to Orion Context Broker, a new schema has been created called openiot. \nThe name of the schema matches the fiware-service header - therefore openiot holds the historic context of the\nIoT devices.
Once running a docker container within the network, it is possible to obtain information about the running\ndatabase.
\nSELECT table_schema,table_name\nFROM information_schema.tables\nWHERE table_schema ='openiot'\nORDER BY table_schema,table_name;\n table_schema | table_name \n--------------+-------------------\n openiot | door_001_door\n openiot | lamp_001_lamp\n openiot | motion_001_motion\n(3 rows)\nThe table_schema matches the fiware-service header supplied with the context data:
To read the data within a table, run a select statement as shown:
\nSELECT * FROM openiot.motion_001_motion limit 10;\n recvtimets | recvtime | fiwareservicepath | entityid | entitytype | attrname | attrtype | attrvalue | attrmd \n---------------+--------------------------+-------------------+------------+------------+-------------+--------------+--------------------------+------------------------------------------------------------------------------\n 1528803005491 | 2018-06-12T11:30:05.491Z | / | Motion:001 | Motion | TimeInstant | ISO8601 | 2018-06-12T11:30:05.423Z | []\n 1528803005491 | 2018-06-12T11:30:05.491Z | / | Motion:001 | Motion | count | Integer | 7 | [{\"name\":\"TimeInstant\",\"type\":\"ISO8601\",\"value\":\"2018-06-12T11:30:05.423Z\"}]\n 1528803005491 | 2018-06-12T11:30:05.491Z | / | Motion:001 | Motion | refStore | Relationship | Store:001 | [{\"name\":\"TimeInstant\",\"type\":\"ISO8601\",\"value\":\"2018-06-12T11:30:05.423Z\"}]\n 1528803035501 | 2018-06-12T11:30:35.501Z | / | Motion:001 | Motion | TimeInstant | ISO8601 | 2018-06-12T11:30:35.480Z | []\n 1528803035501 | 2018-06-12T11:30:35.501Z | / | Motion:001 | Motion | count | Integer | 10 | [{\"name\":\"TimeInstant\",\"type\":\"ISO8601\",\"value\":\"2018-06-12T11:30:35.480Z\"}]\n 1528803035501 | 2018-06-12T11:30:35.501Z | / | Motion:001 | Motion | refStore | Relationship | Store:001 | [{\"name\":\"TimeInstant\",\"type\":\"ISO8601\",\"value\":\"2018-06-12T11:30:35.480Z\"}]\n 1528803041563 | 2018-06-12T11:30:41.563Z | / | Motion:001 | Motion | TimeInstant | ISO8601 | 2018-06-12T11:30:41.520Z | []\n 1528803041563 | 2018-06-12T11:30:41.563Z | / | Motion:001 | Motion | count | Integer | 12 | [{\"name\":\"TimeInstant\",\"type\":\"ISO8601\",\"value\":\"2018-06-12T11:30:41.520Z\"}]\n 1528803041563 | 2018-06-12T11:30:41.563Z | / | Motion:001 | Motion | refStore | Relationship | Store:001 | [{\"name\":\"TimeInstant\",\"type\":\"ISO8601\",\"value\":\"2018-06-12T11:30:41.520Z\"}]\n 1528803047545 | 2018-06-12T11:30:47.545Z | / \nThe usual PostgreSQL query syntax can be used to filter appropriate fields and values. For example to read the rate at which the Motion Sensor with the id=Motion:001_Motion is accumulating, you would make a query as follows:
SELECT recvtime, attrvalue FROM openiot.motion_001_motion WHERE attrname ='count' limit 10;\n recvtime | attrvalue \n--------------------------+-----------\n 2018-06-12T11:30:05.491Z | 7\n 2018-06-12T11:30:35.501Z | 10\n 2018-06-12T11:30:41.563Z | 12\n 2018-06-12T11:30:47.545Z | 13\n 2018-06-12T11:31:02.617Z | 15\n 2018-06-12T11:31:32.718Z | 20\n 2018-06-12T11:31:38.733Z | 22\n 2018-06-12T11:31:50.780Z | 24\n 2018-06-12T11:31:56.825Z | 25\n 2018-06-12T11:31:59.790Z | 26\n(10 rows)\nTo leave the Postgres client and leave interactive mode, run the following:
\n\\q\nYou will then return to the commmand line.
\n","event":[{"listen":"prerequest","script":{"id":"9547319e-f5df-4a3d-9a69-86221f7aeb2f","type":"text/javascript","exec":[""]}},{"listen":"test","script":{"id":"eaf3c847-e1eb-48e2-aadd-35caf0463942","type":"text/javascript","exec":[""]}}],"_postman_id":"19c465ee-0a57-461b-8e27-8994a18df89a"},{"name":"MySQL - Persisting Context","item":[{"name":"Draco - Check System Health","id":"468b00a7-1a20-4e8f-b97e-b4e334fc649a","protocolProfileBehavior":{"disableBodyPruning":true},"request":{"method":"GET","header":[],"url":"http://localhost:9090/nifi-api/system-diagnostics","description":"Once Draco is running, you can check the status by making an HTTP request to the exposed draco port to /nifi-api/system-diagnostics. If the response is blank, this is usually because Draco is not running or is listening on another port.
Once a dynamic context system is up and running, we need to inform Draco of changes in context.
\nThis is done by making a POST request to the /v2/subscription endpoint of the Orion Context Broker.
fiware-service and fiware-servicepath headers are used to filter the subscription to only listen to measurements from the attached IoT SensorsidPattern in the request body ensures that Draco will be informed of all context data changes.url must match the configured NIFI_WEB_HTTP_PORT and the URL the sync is listening on (usually /v2/notify)throttling value defines the rate that changes are sampled.As you can see, the database used to persist context data has no impact on the details of the subscription. It is the same for each database.
\n","urlObject":{"protocol":"http","path":["v2","subscriptions",""],"host":["localhost:1026"],"query":[],"variable":[]}},"response":[],"_postman_id":"930c4b82-aec7-4aa4-b74c-b7d407b688bb"},{"name":"Orion - Check Subscription is working","id":"f44efe83-c13c-4608-ad5c-9d650245e02a","protocolProfileBehavior":{"disableBodyPruning":true},"request":{"method":"GET","header":[{"key":"fiware-service","value":"openiot"},{"key":"fiware-servicepath","value":"/"}],"url":"http://localhost:1026/v2/subscriptions/","description":"If a subscription has been created, you can check to see if it is firing by making a GET \nrequest to the /v2/subscriptions endpoint.
Within the notification section of the response, you can see several additional attributes which describe the health of the subscription
If the criteria of the subscription have been met, timesSent should be greater than 0.\nA zero value would indicate that the subject of the subscription is incorrect or the subscription \nhas created with the wrong fiware-service-path or fiware-service header
The lastNotification should be a recent timestamp - if this is not the case, then the devices\nare not regularly sending data. Remember to unlock the Smart Door and switch on the Smart Lamp
The lastSuccess should match the lastNotification date - if this is not the case \nthen Draco is not receiving the subscription properly. Check that the host name\nand port are correct.
Finally, check that the status of the subscription is active - an expired subscription\nwill not fire.
Similarly, to persisting historic context data into MySQL, we will again need an additional container which hosts\nthe MySQL server, once again the default Docker image for this data can be used. The MySQL instance is listening on the\nstandard 3306 port and the overall architecture can be seen below:
![](\"https://fiware.github.io/tutorials.Historic-Context-NIFI/img/mysql-draco-tutorial.png\")
Once again we have a system with two databases, since the MongoDB container is still required to hold data related to\nthe Orion Context Broker and the IoT Agent.
\nmysql-db:\n restart: always\n image: mysql:5.7\n hostname: mysql-db\n container_name: db-mysql\n expose:\n - \"3306\"\n ports:\n - \"3306:3306\"\n networks:\n - default\n environment:\n - \"MYSQL_ROOT_PASSWORD=123\"\n - \"MYSQL_ROOT_HOST=%\"\n\n\n:information_source: Note: Using the default
\nrootuser and displaying the password in an environment variables\nlike this is a security risk. Whereas this is acceptable practice in a tutorial, for a production environment, you can\navoid this risk by setting up another user and applying\nDocker Secrets
The mysql-db container is listening on a single port:
3306 is the default port for a MySQL server. It has been exposed so you can also run other database tools to\ndisplay data if you wishThe mysql-db container is driven by environment variables as shown:
draco:\n image: ging/fiware-draco:1.1.0\n container_name: draco\n depends_on:\n - mysql-db\n environment:\n - NIFI_WEB_HTTP_PORT=9090\n ports:\n - \"9090:9090\"\n - \"5050:5050\"\n healthcheck:\n test: curl --fail -s http://localhost:9090/nifi-api/system-diagnostics || exit 1\nThe draco container is listening on two ports:
5050 is where the service will be listening for notifications from the Orion\ncontext broker9090 is exposed purely for configuring the processorsTo start the system with a MySQL database run the following command:
\n./services mysql\nThen go to your browser and open Draco using this URL http://localhost:9090/nifi
Now go to the Components toolbar which is placed in the upper section of the NiFi GUI, find the template icon and drag\nand drop it inside the Draco user space. At this point, a popup should be displayed with a list of all the templates\navailable. Please select the template MYSQL-TUTORIAL.
\n![](\"https://fiware.github.io/tutorials.Historic-Context-NIFI/img/draco-template1.png\")
Before starting the processors, you need to set your MySQL password and enable the DBCConnectionPool controller. For\ndoing that please follow the instructions:
\nDo right click on any part of the Draco GUI user space, and then click on configure.\n
Go to the Controller Services Tab, at this point a list of controllers should be displayed, locate the\nDBCConnectionPool controller.
\nClick on the configuration button of the \"DBCPConnectionPool\"\n
Go to the controller Properties tab and put \"123\" in the password field, then apply the changes.\n![](\"https://fiware.github.io/tutorials.Historic-Context-NIFI/img/step3.png\")
Enable the processor by clicking on the thunder icon and then click on enable, then close the controller\nconfiguration page.
For the purpose of this tutorial, we must be monitoring a system where the context is periodically being updated.\nThe dummy IoT Sensors can be used to do this. Open the device monitor page at http://localhost:3000/device/monitor\nand unlock a Smart Door and switch on a Smart Lamp. This can be done by selecting an appropriate the command \nfrom the drop down list and pressing the send button. The stream of measurements coming from the devices can then\nbe seen on the same page:
To read MySQL data from the command line, we will need access to the mysql client, to do this, run an\ninteractive instance of the mysql image supplying the connection string as shown to obtain a command \nline prompt:
docker exec -it db-mysql mysql -h mysql-db -P 3306 -u root -p123\nTo show the list of available databases, run the statement as shown:
\nSHOW DATABASES;\n+--------------------+\n| Database |\n+--------------------+\n| information_schema |\n| mysql |\n| openiot |\n| performance_schema |\n| sys |\n+--------------------+\n5 rows in set (0.00 sec)\nTo show the list of available schemas, run the statement as shown:
\nSHOW SCHEMAS;\n+--------------------+\n| Database |\n+--------------------+\n| information_schema |\n| mysql |\n| openiot |\n| performance_schema |\n| sys |\n+--------------------+\n5 rows in set (0.00 sec)\nAs a result of the subscription of Draco to Orion Context Broker, a new schema has been created called openiot. \nThe name of the schema matches the fiware-service header - therefore openiot holds the historic context of the\nIoT devices.
Once running a docker container within the network, it is possible to obtain information about the running\ndatabase.
\nSHOW tables FROM openiot;\n table_schema | table_name \n--------------+-------------------\n openiot | door_001_door\n openiot | lamp_001_lamp\n openiot | motion_001_motion\n(3 rows)\nThe table_schema matches the fiware-service header supplied with the context data:
To read the data within a table, run a select statement as shown:
\nSELECT * FROM openiot.Motion_001_Motion limit 10;\n+---------------+-------------------------+-------------------+------------+------------+-------------+--------------+--------------------------+------------------------------------------------------------------------------+\n| recvTimeTs | recvTime | fiwareServicePath | entityId | entityType | attrName | attrType | attrValue | attrMd |\n+---------------+-------------------------+-------------------+------------+------------+-------------+--------------+--------------------------+------------------------------------------------------------------------------+\n| 1528804397955 | 2018-06-12T11:53:17.955 | / | Motion:001 | Motion | TimeInstant | ISO8601 | 2018-06-12T11:53:17.923Z | [] |\n| 1528804397955 | 2018-06-12T11:53:17.955 | / | Motion:001 | Motion | count | Integer | 3 | [{\"name\":\"TimeInstant\",\"type\":\"ISO8601\",\"value\":\"2018-06-12T11:53:17.923Z\"}] |\n| 1528804397955 | 2018-06-12T11:53:17.955 | / | Motion:001 | Motion | refStore | Relationship | Store:001 | [{\"name\":\"TimeInstant\",\"type\":\"ISO8601\",\"value\":\"2018-06-12T11:53:17.923Z\"}] |\n| 1528804403954 | 2018-06-12T11:53:23.954 | / | Motion:001 | Motion | TimeInstant | ISO8601 | 2018-06-12T11:53:23.928Z | [] |\n| 1528804403954 | 2018-06-12T11:53:23.954 | / | Motion:001 | Motion | count | Integer | 5 | [{\"name\":\"TimeInstant\",\"type\":\"ISO8601\",\"value\":\"2018-06-12T11:53:23.928Z\"}] |\n| 1528804403954 | 2018-06-12T11:53:23.954 | / | Motion:001 | Motion | refStore | Relationship | Store:001 | [{\"name\":\"TimeInstant\",\"type\":\"ISO8601\",\"value\":\"2018-06-12T11:53:23.928Z\"}] |\n| 1528804409970 | 2018-06-12T11:53:29.970 | / | Motion:001 | Motion | TimeInstant | ISO8601 | 2018-06-12T11:53:29.948Z | [] |\n| 1528804409970 | 2018-06-12T11:53:29.970 | / | Motion:001 | Motion | count | Integer | 7 | [{\"name\":\"TimeInstant\",\"type\":\"ISO8601\",\"value\":\"2018-06-12T11:53:29.948Z\"}] |\n| 1528804409970 | 2018-06-12T11:53:29.970 | / | Motion:001 | Motion | refStore | Relationship | Store:001 | [{\"name\":\"TimeInstant\",\"type\":\"ISO8601\",\"value\":\"2018-06-12T11:53:29.948Z\"}] |\n| 1528804446083 | 2018-06-12T11:54:06.83 | / | Motion:001 | Motion | TimeInstant | ISO8601 | 2018-06-12T11:54:06.062Z | [] |\n+---------------+-------------------------+-------------------+------------+------------+-------------+--------------+--------------------------+------------------------------------------------------------------------------+ \nThe usual MySQL query syntax can be used to filter appropriate fields and values. For example to read the rate at which the Motion Sensor with the id=Motion:001_Motion is accumulating, you would make a query as follows:
SELECT recvtime, attrvalue FROM openiot.Motion_001_Motion WHERE attrname ='count' LIMIT 10;\n+-------------------------+-----------+\n| recvtime | attrvalue |\n+-------------------------+-----------+\n| 2018-06-12T11:53:17.955 | 3 |\n| 2018-06-12T11:53:23.954 | 5 |\n| 2018-06-12T11:53:29.970 | 7 |\n| 2018-06-12T11:54:06.83 | 12 |\n| 2018-06-12T11:54:12.132 | 13 |\n| 2018-06-12T11:54:24.177 | 14 |\n| 2018-06-12T11:54:36.196 | 16 |\n| 2018-06-12T11:54:42.195 | 18 |\n| 2018-06-12T11:55:24.300 | 23 |\n| 2018-06-12T11:55:30.350 | 25 |\n+-------------------------+-----------+\n10 rows in set (0.00 sec)\nTo leave the MySQL client and leave interactive mode, run the following:
\n\\q\nYou will then return to the commmand line.
\n","event":[{"listen":"prerequest","script":{"id":"0a433918-9a68-4cbe-9bdc-031270322304","type":"text/javascript","exec":[""]}},{"listen":"test","script":{"id":"db11ec7d-1bde-4317-8318-db7b0ebf5c61","type":"text/javascript","exec":[""]}}],"_postman_id":"023cc10d-0fa3-46c1-b2bc-bad848e141fe"},{"name":"Multi-Agent Persisting to Multiple Databases","item":[{"name":"Draco - Obtain version information","id":"975983b4-c63c-4c0e-8d67-2e42dbd0acd8","protocolProfileBehavior":{"disableBodyPruning":true},"request":{"method":"GET","header":[],"url":"http://localhost:9090/v1/version","description":"Once Draco is running, You can check the status by making an HTTP request to the exposed CYGNUS_API_PORT port. \nIf the response is blank, this is usually because Draco is not running or is listening on another port.
\n\n","urlObject":{"protocol":"http","path":["v1","version"],"host":["localhost:9090"],"query":[],"variable":[]}},"response":[],"_postman_id":"975983b4-c63c-4c0e-8d67-2e42dbd0acd8"},{"name":"Orion - Subscribe to Context Changes","id":"949903c8-7764-485c-a825-af1cd48cc605","protocolProfileBehavior":{"disableBodyPruning":true},"request":{"method":"POST","header":[{"key":"Content-Type","value":"application/json"},{"key":"fiware-service","value":"openiot"},{"key":"fiware-servicepath","value":"/"}],"body":{"mode":"raw","raw":"{\n \"description\": \"Notify Draco of all context changes for MySQL on port 5050\",\n \"subject\": {\n \"entities\": [\n {\n \"idPattern\": \".*\"\n }\n ]\n },\n \"notification\": {\n \"http\": {\n \"url\": \"http://draco:5050/notify\"\n },\n \"attrsFormat\": \"legacy\"\n },\n \"throttling\": 5\n}"},"url":"http://localhost:1026/v2/subscriptions/","description":"Troubleshooting: What if the response is blank ?
\n\n
\n- To check that a docker container is running try
\n\ndocker ps\nYou should see several containers running. If
\ndracois not running, you can restart the containers as necessary.
Once a dynamic context system is up and running, we need to inform Draco of changes in context.
\nThis is done by making a POST request to the /v2/subscription endpoint of the Orion Context Broker.
fiware-service and fiware-servicepath headers are used to filter the subscription to only listen to measurements from the attached IoT SensorsidPattern in the request body ensures that Draco will be informed of all context data changes.attrsFormat=legacy is required since Draco currently only accepts notifications in the older NGSI v1 format.throttling value defines the rate that changes are sampled.When running in multi-agent mode, the notification url for each subscription must match the defaults for the given database.
The default port mapping can be seen below:
\nSince this subscription is using port 5050 the context data will eventually be persisted to the MySQL database.
If a subscription has been created, you can check to see if it is firing by making a GET \nrequest to the /v2/subscriptions endpoint.
Within the notification section of the response, you can see several additional attributes which describe the health of the subscription
If the criteria of the subscription have been met, timesSent should be greater than 0.\nA zero value would indicate that the subject of the subscription is incorrect or the subscription \nhas created with the wrong fiware-service-path or fiware-service header
The lastNotification should be a recent timestamp - if this is not the case, then the devices\nare not regularly sending data. Remember to unlock the Smart Door and switch on the Smart Lamp
The lastSuccess should match the lastNotification date - if this is not the case \nthen Draco is not receiving the subscription properly. Check that the host name\nand port are correct.
Finally, check that the status of the subscription is active - an expired subscription\nwill not fire.
It is also possible to configure Draco to populate multiple databases simultaneously. We can combine the architecture\nfrom the three previous examples and configure Draco to store data in multiple sinks.
\n![](\"https://fiware.github.io/tutorials.Historic-Context-NIFI/img/multiple-draco-tutorial.png\")
We now have a system with three databases, PostgreSQL and MySQL for data persistence and MongoDB for both data\npersistence and holding data related to the Orion Context Broker and the IoT Agent.
\ndraco:\n image: ging/fiware-draco:1.1.0\n container_name: draco\n depends_on:\n - mysql-db\n - mongo-db\n - postgres-db\n environment:\n - NIFI_WEB_HTTP_PORT=9090\n ports:\n - \"9090:9090\"\n - \"5050:5050\"\n healthcheck:\n test: curl --fail -s http://localhost:9090/nifi-api/system-diagnostics || exit 1\nThe draco container is listening on two ports:
5050 is where the service will be listening for notifications from the Orion\ncontext broker9090 is exposed purely for configuring the processorsTo start the system with multiple databases run the following command:
\n./services multiple\nThen go to your browser and open Draco using this URL http://localhost:9090/nifi
Now go to the Components toolbar which is placed in the upper section of the NiFi GUI, find the template icon and drag\nand drop it inside the Draco user space. At this point, a popup should be displayed with a list of all the templates\navailable. Please select the template MULTIPLE-SINKS-TUTORIAL.
\n![](\"https://fiware.github.io/tutorials.Historic-Context-NIFI/img/multiple-tutorial-template.png\")
Now repeat the process for setting the password in the controller \"DBCPConnectionPool\" each connection MySQL and\nPostgreSQL
\nSelect all the processors (press shift and click on every processor) and start them by clicking on the start button.\nNow, you can see that the status icon of each processor turned from red to green.
\nFor the purpose of this tutorial, we must be monitoring a system where the context is periodically being updated.\nThe dummy IoT Sensors can be used to do this. Open the device monitor page at http://localhost:3000/device/monitor\nand unlock a Smart Door and switch on a Smart Lamp. This can be done by selecting an appropriate the command \nfrom the drop down list and pressing the send button. The stream of measurements coming from the devices can then\nbe seen on the same page: