scala-pocs

A collection of POC's written in Scala & running in Docker for Kafka, Hadoop and PostgreSQL

Setting up Kafka

Copy the docker/kafka-up.sh file into your host machine and run sh kafka-up.sh

NOTE: the following processes run interactively, so you will need to open multiple terminals.

Generate some data for a race with 5,000 participants and 2 checkpoints (start/finish) and stream it into Kafka on the 'reads' topic with the following command:

docker run --rm -i anaerobic/generate-race 5000 2 | docker run -i --rm --net host anaerobic/fsharp-kafka-producer reads http://kafka.lacolhost.com:9092

Now stream the race out of Kafka, into a microservice that bracketizes each participant and back into Kafka on the 'results' topic with the following command:

docker run --rm --net host anaerobic/fsharp-kafka-consumer reads http://kafka.lacolhost.com:9092 | docker run -i --rm anaerobic/streaming-fsharp | docker run -i --rm --net host anaerobic/fsharp-kafka-producer results http://kafka.lacolhost.com:9092

You can verify the data is on the 'results' topic using:

docker run -i --rm --net host anaerobic/fsharp-kafka-consumer results http://kafka.lacolhost.com:9092

Setting up PostgreSQL

Copy the docker/postgres-up.sh file into your host machine and run sh postgres-up.sh

Add a foo table to the postgres db (I used pgAdmin III) with the following schema:

CREATE TABLE public.foo
(
   id serial, 
   some_json json
) 
WITH (
  OIDS = FALSE
)
;

##Usage for kafka-to-postgres:

0. Build the kafka-to-postgres/target/kafka-to-postgres-1.0.jar with cd kafka-to-postgres and mvn package in a command prompt.
1. Copy the kafka-to-postgres/target/kafka-to-postgres-1.0.jar to the docker/kafka/share directory
2. Copy the docker/kafka directory to ~/kafka on your host machine
3. Run sh ~/kafka/docker-up.sh on your host machine
4. Run sh share/kafka-to-postgres.sh from inside the kafka container
5. Verify your data is in PostgreSQL
6. (╯°□°）╯︵ ┻━┻
7. ᕕ( ᐛ )ᕗ

##Usage for hdfs-to-postgres:

NOTE: I've got a forked Kafka to HDFS library working at https://github.com/anaerobic/kafka-hadoop-consumer but it is in Java and it doesn't use consumer groups, so the next best thing was to just copy the output of that into a file and copy it from local into hdfs instead...

0. Build the hdfs-to-postgres/target/hdfs-to-postgres-1.0.jar with cd hdfs-to-postgres and mvn package in a command prompt.
1. Copy the hdfs-to-postgres/target/hdfs-to-postgres-1.0.jar to the docker/hadoop/share directory
2. Copy the docker/hadoop directory to ~/hadoop on your host machine
3. Run sh ~/hadoop/docker-up.sh on your host machine
4. Run sh share/hdfs-to-postgres.sh from inside the Hadoop container
5. Verify your data is in PostgreSQL
6. (╯°□°）╯︵ ┻━┻
7. ᕕ( ᐛ )ᕗ