Bus Floating Data

this project is a simple show case which shows how to create a streaming akka actor to ingest data from the http source to Kafka And consume those messages either via akka or spark to push them to a cassandra table.

The project is a multi project with cross compilation due to incompatabilities with kafka spark and akka when using the latest scala akka versions.

for a clean build from commandline:

sbt ";clean ;test; publishLocal"

or simply call

sbt create

To run the applications

ingest:

sbt ingest/run

select the 1st entry as the simple does all steps at once or better

sbt runIngest

to prepare the ingest container call:

sbt createIngestContainer

akka frontend: so service/run sbt runService

to prepare the service container:

sbt createServerContainer

spark digest: The spark digest starts a local spark master. As a spark job requires a fat jar first create that one:

sbt createDigestUberJar

to run the kafka->cassandra spark job: sbt submitKafkaCassandra

to run the cluster spark job: sbt submitClusterSpark

Pre-conditions to run those, have

a) a cassandra running b) a zookeeper running c) a kafka server running

for b) and c) you can use brew to install those. e.g. brew cask install zookeeper and brew install kafka

to run zookeper (not as a service)

zkServer start

to run kafka (not as a service)

kafka-server-start /usr/local/etc/kafka/server.properties

Operating Kafka

to clean old data from kafka turn down the retention time in kafka for the topic

kafka-configs --zookeeper localhost:2181 --entity-type topics --alter --add-config retention.ms=1000 --entity-name METRO-Vehicles

this will result in a clean up afte about a minute

after that you can configure back the time to about 2 days (more isn't worth when testing)

kafka-configs --zookeeper localhost:2181 --entity-type topics --alter --add-config retention.ms=172800000 --entity-name METRO-Vehicles

to check the current setting do the following:

kafka-configs --zookeeper localhost:2181 --entity-type topics --describe --entity-name METRO-Vehicles

Preparations for running on AWS

See also the terraform folder on how to run this on AWS. But prior to using it on AWS with a DC/OS cluster, make sure to create all required Docker images and deploy them to a place you have access to. To create those artefacts just call

sbt createAWS

this will create all uber-jars and runnable Docker images.

how to kill a lunatic running framework

curl -d'frameworkId=frameworkId' master.mesos:5050/master/teardown

which params to use for flink jobs

Run the Kafka to Cassandra app

Entry Class: de.nierbeck.floating.data.stream.flink.KafkaToCassandraFlinkApp

Programm Arguments:
METRO-Vehicles node.cassandra.l4lb.thisdcos.directory:9042 broker.kafka.l4lb.thisdcos.directory:9092

Run the cluster hotspot calculation

Entry Class: de.nierbeck.floating.data.stream.flink.CalcClusterFlinkApp

Programm Arguments: --connection node.cassandra.l4lb.thisdcos.directory:9042

Kubernetes

To browse the K8s cluster browse to:
kube-apiserver-0-instance.kubernetes.mesos:9000/api/v1/namespaces/kube-system/services/kubernetes-dashboard/proxy/

Marathon-LB

To browse the HA Proxy settings:
:9090/haproxy?stats

anierbeck / busfloatingdata 0.1.1