To deploy the stream, we can execute the following command on the Data Flow Shell:
stream deploy --name process-stock-change-events
You will see the message shown here when the request is sent for the creation of stream:
Deployment request has been sent for stream 'process-stock-change-events'
The following extract shows an extract from the Local Data Flow Server log:
o.s.c.d.spi.local.LocalAppDeployer : deploying app process-stock-change-events.event-store-sink instance 0
Logs will be in /var/folders/y_/x4jdvdkx7w94q5qsh745gzz00000gn/T/spring-cloud-dataflow-3084432375250471462/process-stock-change-events-1492100265496/process-stock-change-events.event-store-sink
o.s.c.d.spi.local.LocalAppDeployer : deploying app process-stock-change-events.stock-intelligence-processor instance 0
Logs will be in /var/folders/y_/x4jdvdkx7w94q5qsh745gzz00000gn/T/spring-cloud-dataflow-3084432375250471462/process-stock-change-events-1492100266448/process-stock-change-events.stock-intelligence-processor
o.s.c.d.spi.local.LocalAppDeployer : deploying app process-stock-change-events.significant-stock-change-source instance 0
Logs will be in /var/folders/y_/x4jdvdkx7w94q5qsh745gzz00000gn/T/spring-cloud-dataflow-3084432375250471462/process-stock-change-events-1492100267242/process-stock-change-events.significant-stock-change-source
A few important things to note are as follows:
- When we deploy a stream, Spring Cloud Data Flow will deploy all the applications in the stream and set up the connections between the applications through the message broker. The application code is independent of the message broker. Kafka has a different message broker setup compared to RabbitMQ. Spring Cloud Data Flow will take care of it. If you want to switch from RabbitMQ to Kafka, the application code does not need to change.
- The Local Data Flow Server log contains the path to logs of all the applications-- source, processor, and sink.