THIS REPO IS ARCHIVED Based on Security issues SEC ticket #SEC-2988

StreamX: Kafka Connect for S3

Forked from the awesome kafka-connect-hdfs

StreamX is a kafka-connect based connector to copy data from Kafka to Object Stores like Amazon s3, Google Cloud Storage and Azure Blob Store. It focusses on reliable and scalable data copying. It can write the data out in different formats (like parquet, so that it can readily be used by analytical tools) and also in different partitioning requirements.

Features :

StreamX inherits rich set of features from kafka-connect-hdfs.

In addition to these, we have made changes to the following to make it work efficiently with s3

Getting Started:

Pre-req : StreamX is based on Kafka Connect framework, which is part of Kafka project. Kafka Connect is added in Kafka 0.9, hence StreamX can only be used with Kafka version >= 0.9. To download Kafka binaries, check here.

Clone : git clone

Branch : For Kafka 0.9, use 2.x branch. For Kafka 0.10 and above, use master branch.

Build : mvn -DskipTests package

Once the build succeeds, StreamX packages all required jars under target/streamx-0.1.0-SNAPSHOT-development/share/java/streamx/* in StreamX repo. This directory needs to be in classpath.

Add Connector to Kafka Connect Classpath:

export CLASSPATH=$CLASSPATH:`pwd`/target/streamx-0.1.0-SNAPSHOT-development/share/java/streamx/*

Start Kafka Connect

In Kafka, change the following in config/ or config/ depending on what mode you want to use.

bootstrap.servers=set Kafka end-point (ex: localhost:9092)

Use ByteArrayConverter to copy data from Kafka as-is without any changes. (copy JSON/CSV)

Run Kafka Connect in Standalone mode

Set s3.url and hadoop-conf in StreamX config/ StreamX packages hadoop-conf directory at config/hadoop-conf for ease-of-use. Set s3 access and secret keys in config/hadoop-conf/hdfs-site.xml.

In Kafka, run

bin/connect-standalone etc/kafka/ \

You are done. Check s3 for ingested data!

Run Kafka Connect in distributed mode
bin/ config/

We have started the Kafka Connect framework and the S3 Connector is added to classpath. Kafka Connect framework starts a REST server (rest.port property in listening for Connect Job requests. The copy job can be submitted by hitting the REST end-point using curl or any REST clients.

For example, to submit a copy job from Kafka to S3

curl -i -X POST \
   -H "Accept:application/json" \
   -H "Content-Type:application/json" \
   -d \
}}' \

You have submitted the job, check S3 for output files. For the above copy job, it will create s3://streamx/demo/topics/adclicks/partition=x/

Note that, a single copy job could consume from multiple topics and writes to topic specific directory.

To delete a Connect job,

curl -i -X DELETE \
   -H "Accept:application/json" \
   -H "Content-Type:application/json" \

To list all Connect jobs,

curl -i -X GET \
   -H "Accept:application/json" \
   -H "Content-Type:application/json" \

Restarting Connect jobs : All Connect jobs are stored in a Kafka Queue. So, restarting the Kafka Connect will restart all the connectors submitted to it.

Docker Streamx supports Docker, but only in distributed mode To build your image,

docker build -t qubole/streamx .

When you run your container, you can override all the properties in file with env vars. env_vars will be of format CONNECT_BOOTSTRAP_SERVERS corresponding to bootstrap.servers. The convention is to prefix env vars with CONNECT. Example of how to run your container,

docker run -d -p 8083:8083 --env CONNECT_BOOTSTRAP_SERVERS=public_dns:9092 --env CONNECT_AWS_ACCESS_KEY=youracesskey --env CONNECT_AWS_SECRET_KEY=yoursecretkey qubole/streamx

You can also use Avro/Parquet format. Example:

docker run -d -p 8083:8083 --env CONNECT_BOOTSTRAP_SERVERS=public_dns:9092 --env CONNECT_AWS_ACCESS_KEY=youracesskey --env CONNECT_AWS_SECRET_KEY=yoursecretkey  --env CONNECT_KEY_CONVERTER=io.confluent.connect.avro.AvroConverter --env CONNECT_VALUE_CONVERTER=io.confluent.connect.avro.AvroConverter --env CONNECT_KEY_CONVERTER_SCHEMA_REGISTRY_URL= --env CONNECT_VALUE_CONVERTER_SCHEMA_REGISTRY_URL= qubole/streamx