Reactive JDBC Experiment

This is a simple experiment to test Spring 5's Webflux Module's Functional Programming Model interaction with the Reactiverse reactive-pg-client.

Disclaimer: the reactive-pg-client does not implement the JDBC specification.

Prerequisites

• An account with Space Developer role access on a Cloud Foundry foundation, e.g., Pivotal Web Services
• CF CLI 6.37.0 or better if you want to push the application to a Cloud Foundry (CF) instance
• httpie 0.9.9 or better to simplify interaction with API endpoints
• Java JDK 1.8u172 or better to compile and run the code
• Gradle 4.8 or better to build and package source code
• Docker for Mac or Windows for spinning up a local instance of Postgres and Adminer (a database administration interface)

Clone

git clone https://github.com/pacphi/reactive-jdbc-demo.git

How to build

cd reactive-jdbc-demo
gradle build

How to run locally

1. Prepare database

   Open a Terminal session, then type

   docker-compose up -d

2. Login to Adminer interface

   Open a browser and visit http://localhost:9090

   Credentials are:

   • System => PostgreSQL
   • Server => db
   • Username => admin
   • Password => passw0rd
   • Database => people

   Click the Login button

3. Click on the SQL command link

   Link is in the upper left hand-corner of the interface

4. Cut-and-paste the contents of people.ddl into the text area, then click the Execute button

5. Start the application

   Start a new Terminal session and type

   gradle bootRun

6. Let's create some data using the API

   http POST localhost:8080/person firstName=Dweezil lastName=Zappa age=48
   
   HTTP/1.1 202 Accepted
   content-length: 0

7. Verify that we can find the person we added

   http localhost:8080/person
   
   HTTP/1.1 200 OK
   Content-Type: application/json
   transfer-encoding: chunked
   
   [
       {
           "age": 48,
           "firstName": "Dweezil",
           "id": "582279d1-9bd1-4e49-946c-ac720de0e04f",
           "lastName": "Zappa"
       }
   ]

8. Let's ask for a person by id

   http localhost:8080/person/582279d1-9bd1-4e49-946c-ac720de0e04f
   
   HTTP/1.1 200 OK
   Content-Length: 95
   Content-Type: application/json
   
   {
       "age": 48,
       "firstName": "Dweezil",
       "id": "582279d1-9bd1-4e49-946c-ac720de0e04f",
       "lastName": "Zappa"
   }

How to shutdown locally

1. Stop the application

   Visit the Terminal session where you started application and press Ctrl+c

2. Shutdown Postgres and Adminer interface

   Visit the Terminal session where you invoked docker-compose-up -d and type

   docker-compose down

   Note: the data volume is persistent! If you want to destroy all unused volumes and reclaim some additional space, type

   docker volume prune

How to run on Cloud Foundry

1. Authenticate to a foundation using the API endpoint.

   E.g., login to Pivotal Web Services

   cf login -a https:// api.run.pivotal.io

2. Push the app, but don't start it

   cf push reactive-jdbc-demo --random-route --no-start -p ./build/libs/reactive-jdbc-demo-0.0.1-SNAPSHOT.jar -m 1G -b https://github.com/cloudfoundry/java-buildpack.git

3. Let's fire fire up a Postgres instance

   We're going to use ElephantSQL

   cf cs elephantsql panda {service name}

   Note: this is going to cost you $19/month to keep alive Replace {service name} above with your desired service name

4. Next we'll bind the service to the application

   cf bs reactive-jdbc-demo {service name}

   Make sure {service name} above matches what you defined in Step 3

5. Let's verify that VCAP_SERVICES was properly injected

   cf env reactive-jdbc-demo
   
   Getting env variables for app reactive-jdbc-demo in org scooby-doo / space dev as [email protected]...
   OK
   
   System-Provided:
   {
   "VCAP_SERVICES": {
   "elephantsql": [
   {
       "binding_name": null,
       "credentials": {
       "max_conns": "20",
       "uri": "postgres://blxrphig:[email protected]:5432/banzlhig"
       },
       ...

   We're interested in vcap_services.elephantsql.uri The URI consists of {vendor}://{username}:{password}@{server}:5432/{database}

6. We'll set an environment variable

   cf set-env reactive-jdbc-demo PG_LOOKUP_KEY {service name}

   {service name} above should match value in steps 3 and 4

7. Now let's startup the application

   cf start reactive-jdbc-demo

8. Launch Adminer to administer the database

   The people table doesn't exist yet, so we need to create it

   docker-compose up -d

   Open a browser and visit http://localhost:9090

   Credentials are:

   • System => PostgreSQL
   • Server => {server}
   • Username => {username}
   • Password => {password}
   • Database => {database}

   Replace all bracketed values above with what you learned from Step 5

   Click the Login button

9. Click on the SQL command link

10. Cut-and-paste the contents of people.ddl into the text area, then click the Execute button

11. Follow steps 6-8 above in How to run locally to interact with API

    But replace occurrences of localhost:8080 with URL to application hosted on Cloud Foundry

Congratulations! You've just pushed and interacted with a 100% reactive and cloud native app.

How to spin down workloads on Cloud Foundry

1. Stop the application

   cf stop reactive-jdbc-demo

2. Unbind the database instance

   cf us reactive-jdbc-demo {service name}

   {service name} above should match value in How to run on Cloud Foundry steps 3 and 4

3. Delete the database instance

   cf ds {service name}

   {service name} above should match value in How to run on Cloud Foundry steps 3 and 4

4. Delete the application

   cf delete reactive-jdbc-demo

What to look forward to?

• Asynchronous Database Access (ADBA)
• ADBA over JDBC (AoJ)

Oracle continues to work on ADBA while having released AoJ under an Apache license to get community feedback.

Maybe we will see something concrete in JDK 11?

What else is there to play with?

• rxjava2-jdbc
• Vert.x JDBC Client
• Reactive Relational Database Connectivity Client (R2DBC)