Skip to content
Switch branches/tags
Go to file

Latest commit


Git stats


Failed to load latest commit information.
Latest commit message
Commit time

An Implementation of Kafka + Storm


Within the Lambda Architecture repository are the following directories:

  • twitter-example - a Storm topology that pulls from a Kafka 0.7 server using KafkaSpout. The KafkaSpout pulls tweets from the Kafka server that originally come from the Twitter dev stream via the twitter-kafka-producer.
  • twitter-kafka-producer – a Kafka producer a.k.a. publisher that links the Twitter dev testing stream to Kafka


Even though Kafka v0.7 is not the most recent version of Kafka, it was the only one that I was able to get to work at the time with Storm with relatively little effort. Others also have trouble with getting KafkaSpout from the current version of storm-kafka (v 0.9x) to work with Kafka v0.8 (it only works with Kafka v0.7) -- see this email group thread!topic/storm-user/OXQ0a9ppcYM

Those with extra time should consider updating the latest version KafkaSpout to work with the latest version of Kafka, which is said to be much better than Kafka v0.7.

In order to target Storm version 0.8.2, use storm-kafka version 0.8.0-wip4 -- see this email group thread!topic/storm-user/DEqXfOSv_MA


Kafka v0.7

Download a v 0.7.x Kafka release and follow the Kafka v0.7 documentation's Quickstart instructions for downloading,installing, and running.

Documentation for Kafka v0.7 can be found here.

In order to install a multi-broker Kafka cluster, you can adapt the instructions from Quickstart of the latest Kafka documentation:

Configure the built-in Kafka Zookeeper instance

  • ensure in config/
    • clientPort=2181
Configure the Kafka brokers
  • cp config/ config/
    • edit config/
      • brokerid=0
      • port=9092
      • log.dir=/tmp/kfk0.7-logs-0
  • cp config/ config/
    • edit config/
      • brokerid=1
      • port=9093
      • log.dir=/tmp/kfk0.7-logs-1
  • cp config/ config/
    • edit config/
      • brokerid=2
      • port=9094
      • log.dir=/tmp/kfk0.7-logs-2

Configure the Twitter Dev Stream

  • Configure the access credentials
  • Configure the terms used to filter Twitter stream
    • In ./twitter-kafka-producer/src/twitter_kafka_producer/core.clj:
      • Edit the track-terms vector to contain the strings used to filter the Tweet stream


  • Run the Kafka Zookeeper instance
    • bin/ config/
  • Start the Kafka brokers, each with its own JMX port number
    • JMX_PORT=2002 bin/ config/
    • JMX_PORT=2003 bin/ config/
    • JMX_PORT=2004 bin/ config/
  • In Kafka v0.7, topics are created automatically when you write to them. In v0.8, you must create the topic manually. Nothing to do here for Kafka topic creation in this particular example implementation.
    • Debugging of topics can be done through running the console producer and consumer as:
      • console producer - bin/ --zookeeper localhost:2181 --topic twitter-to-storm
      • console consumer -bin/ --zookeeper localhost:2181 --topic twitter-to-storm --from-beginning
  • Run the Kafka producer
    • cd ./twitter-kafka-producer
    • lein do clean, run (this is lein clean and lein run combined as one call to lein - note the spacing around do and commas)
  • Run the Storm topology (that is, a Storm instance with this topology)
    • cd ./twitter-example
    • lein do clean, run


For me, so far, the hardest part by far of getting Kafka, Storm, and KafkaSpout to work together has been a matter of getting the correct versions to match up.

The code currently uses the following versions for the components:

  • Kafka v 0.7.2
  • Kafka client API via clj-kafka v 0.0.7-0.7 for a Kafka producer
  • Storm v 0.8.2
  • storm-kafka v 0.8.0-wip4 for a KafkaSpout (that version will work with Storm v 0.8.2, and it works with Kafka only for Kafka v 0.7.2)
  • twitter4j-core and twitter4j-stream v 3.0.6 (the versions for both should be the same)

The version numbers are as used within Leiningen (therefore, you should be able to use them via Maven (either in Maven Central or Clojars))


The license for all of the code in the repository (including twitter-example and twitter-kafka-producer) is EPL, the same as Clojure.


No releases published


No packages published