Join us at Scylla Summit 2022 on Feb 9-10 Register for Free
Learn More

Kafka Sink Connector Quickstart

Topic: Kafka Connector

Learn: how to setup the ScyllaDB Sink Connector against a Dockerized ScyllaDB

Audience: Application Developer


This quickstart will show how to setup the ScyllaDB Sink Connector against a Dockerized ScyllaDB.

Preliminary setup

  1. Using Docker, follow the instructions to launch Scylla.

  2. Start the Docker container, replacing the --name and --host name parameters with your own information. For example:

    docker run --name some-scylla --hostname some-scylla -d scylladb/scylla
  3. Run docker ps to show the exposed ports. The output should be similar to this example:

    docker ps
    CONTAINER ID        IMAGE                     COMMAND                  CREATED             STATUS              PORTS                                              NAMES
    26cc6d47efe3        replace-with-image-name   "/docker-entrypoint.…"   4 hours ago         Up 23 seconds>1883/tcp,>9001/tcp   anonymous_my_1
  4. Continue with either Confluent or Manual Installation.

Install using Confluent Platform

If you are new to Confluent, download Confluent Platform.

  1. In the Self managed software box, click DOWNLOAD FREE

  2. Fill in your email address.

  3. Open the Select Deployment Type drop-down and select ZIP.

  4. Accept the Terms & Conditions and click DOWNLOAD FREE.

  5. You will receive an email with instructions. Download / move the file to the desired location.

  6. Continue with the setup following this document.

Install Kafka Connector manually

  1. Navigate to the Kafka Connect Scylladb Sink github page and clone the repository.

  2. Using a terminal, open the source code (src) folder.

  3. Run the command mvn clean install.

  4. Run the Integration Tests in an IDE. If tests fail run mvn clean install -DskipTests.


    To run Integration Tests, there is no need to run Confluent. Use docker-compose.yml file in the github repository and run the following command (it contains images to run Kafka and other services):

    docker-compose -f docker-compose.yml up

    After completion of the above steps, a folder named components will be created in the target folder of the source code folder. The Connector jar files are present in {source-code-folder}/target/components/packages/[jar-files] Create a folder by the name of ScyllaDB-Sink-Connector and copy the jar files into it. Navigate to your Confluent Platform installation directory and place this folder in {confluent-directory}/share/java.

Add Sink Connector plugin

The Scylla sink connector is used to publish records from a Kafka topic into Scylla. Adding a new connector plugin requires restarting Connect. Use the Confluent CLI to restart Connect.

  1. Run the following

    confluent local stop && confluent local start

    The output will be similar to:

    confluent local stop && confluent local start
    Starting zookeeper
    zookeeper is [UP]
    Starting Kafka
    Kafka is [UP]
    Starting schema-registry
    schema-registry is [UP]
    Starting kafka-rest
    kafka-rest is [UP]
    Starting connect
    connect is [UP]
  2. Check if the kafka-connect-scylladb connector plugin has been installed correctly and picked up by the plugin loader:

    curl -sS localhost:8083/connector-plugins | jq .[].class | grep ScyllaDbSinkConnector

    Your output should resemble:


Connector configuration

  1. Save the configuration settings in a file named kafka-connect-scylladb.json its contents should contain:

         "name" : "scylladb-sink-connector",
         "config" : {
           "connector.class" : "io.connect.scylladb.ScyllaDbSinkConnector",
           "tasks.max" : "1",
           "topics" : "topic1,topic2,topic3",
           "" : "scylladb-hosts",
           "scylladb.keyspace" : "test"
  2. Load the connector. Run the following command:

    curl -s -X POST -H 'Content-Type: application/json' --data @kafka-connect-scylladb.json http://localhost:8083/connectors
  3. Update the configuration of the existing connector.

    curl -s -X PUT -H 'Content-Type: application/json' --data @kafka-connect-scylladb.json http://localhost:8083/connectors/scylladb/config
  4. Once the Connector is up and running, use the command kafka-avro-console-producer to produce records(in AVRO format) into the Kafka topic.

    --broker-list localhost:9092
    --topic example
    --property parse.key=true
    --property key.schema='{"type":"record",name":"key_schema","fields":[{"name":"id","type":"int"}]}'
    --property "key.separator=$"
    --property value.schema='{"type":"record","name":"value_schema","fields":[{"name":"id","type":"int"},
  5. Test Scylla by running a select cql query:

    cqlsh>select * from demo.example;
     id | firstname | lastname
      1 |     first |     last

Scylla modes

There are two modes, Standalone and Distributed.

  • Standard - will use the properties based example.

  • Distributed - will use the JSON / REST examples.

Use this command to load the connector and connect to ScyllaDB instance without authentication:

curl -s -X POST -H 'Content-Type: application/json' --data @kafka-connect-scylladb.json http://localhost:8083/connectors

Select one of the following configuration methods based on how you have deployed |kconnect-long|. Distributed Mode will the JSON / REST examples. The standalone mode will use the properties based example.


Each json record should consist of a schema and payload.

Distributed Mode JSON example

  "name" : "scylladb-sink-connector",
  "config" : {
    "connector.class" : "io.connect.scylladb.ScyllaDbSinkConnector",
    "tasks.max" : "1",
    "topics" : "topic1,topic2,topic3",
    "" : "scylladb-hosts",
    "scylladb.keyspace" : "test",
    "key.converter" : "org.apache.kafka.connect.json.JsonConverter",
    "value.converter" : "org.apache.kafka.connect.json.JsonConverter"
    "key.converter.schemas.enable" : "true",
    "value.converter.schemas.enable" : "true",

    "transforms" : "createKey",
    "transforms.createKey.fields" : "[field-you-want-as-primary-key-in-scylla]",
    "transforms.createKey.type" : "org.apache.kafka.connect.transforms.ValueToKey"

Standalone Mode JSON example

To load the connector in Standalone mode use:

confluent local load scylladb-sink-conector -- -d

Use the following configuratopn settings:




For Example:

kafka-console-producer --broker-list localhost:9092 --topic sample-topic
>{"schema":{"type":"struct","fields":[{"type":"int32","optional":false,"field":"id"},{"type":"string","optional":false,"field":"name"},{"type":"string","optional":true,"field":"department"},"payload":{"id":10,"name":"John Doe10","department":"engineering"}}

Run the select cql query to view the data:

Select * from keyspace_name.topic-name;


To publish records in Avro Format use the following properties:



This example connects to a Scylla instance with security enabled and username / password authentication.

Select one of the following configuration methods based on how you have deployed |kconnect-long|. Distributed Mode will the JSON / REST examples. The standalone mode will use the properties based example.

Distributed Mode example

  "name" : "scylladbSinkConnector",
  "config" : {
    "connector.class" : "io.connect.scylladb.ScyllaDbSinkConnector",
    "tasks.max" : "1",
    "topics" : "topic1,topic2,topic3",
    "" : "cassandra",
    "scylladb.keyspace" : "test",
    "" : "true",
    "scylladb.username" : "example",
    "scylladb.password" : "password",
    **add other properties same as in the above example**

Standalone Mode example



To check logs for the Confluent Platform use:

confluent local log <service> -- [<argument>] --path <path-to-confluent>

To check logs for Scylla:

docker logs some-scylla | tail

Additional information