This quickstart will show you how to setup the Scylla CDC Source Connector to replicate changes made in a Scylla table using Scylla CDC.
First, let's setup a Scylla cluster and create a CDC-enabled table.
For the purpose of this quickstart, we will configure a Scylla instance using Docker. You can skip this section if you have already installed Scylla. To learn more about installing Scylla in production environments, please refer to the Install Scylla page.
- Using Docker, follow the instructions to launch Scylla.
- Start the Docker container, replacing the
--name
and--host name
parameters with your own information. For example:docker run --name scylla-cdc-quickstart --hostname scylla-cdc-quickstart -d scylladb/scylla
- 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 4fca02217055 scylladb/scylla "/docker-entrypoint.…" 8 seconds ago Up 7 seconds 22/tcp, 7000-7001/tcp, 9042/tcp, 9160/tcp, 9180/tcp, 10000/tcp scylla-cdc-quickstart
Let's connect to your Scylla cluster and create a new CDC-enabled table. We will create an example table by issuing the following CQL query and insert some example data:
CREATE KEYSPACE quickstart_keyspace WITH REPLICATION = {'class': 'SimpleStrategy', 'replication_factor': 1};
CREATE TABLE quickstart_keyspace.orders(
customer_id int,
order_id int,
product text,
PRIMARY KEY(customer_id, order_id)) WITH cdc = {'enabled': true};
INSERT INTO quickstart_keyspace.orders(customer_id, order_id, product) VALUES (1, 1, 'pizza');
INSERT INTO quickstart_keyspace.orders(customer_id, order_id, product) VALUES (1, 2, 'cookies');
INSERT INTO quickstart_keyspace.orders(customer_id, order_id, product) VALUES (1, 3, 'tea');
If you already have a table you wish to use, but it does not have CDC enabled, you can turn it on by using the following CQL query:
ALTER TABLE keyspace.table_name with cdc = {'enabled': true};
To learn more about Scylla CDC, visit Change Data Capture (CDC) page.
Scylla CDC Source Connector works well with both open-source Kafka and Confluent Platform. Kafka version 2.6.0 or greater is required. In this quickstart we will show how to install the Confluent Platform and deploy the connector (applicable to both open-source Kafka and Confluent Platform).
If you are new to Confluent, download Confluent Platform.
- In the Download Confluent Platform section fill in your email address
- Open the Select Deployment Type drop-down and select ZIP
- Accept the Terms & Conditions and click DOWNLOAD FREE
- You will receive an email with instructions. Download / move the file to the desired location
- Continue with the setup following this document
- Download or build Scylla CDC Source Connector using the project build instructions
- Deploy the connector:
- If you use Confluent Platform, move connector JAR files to the
share/java
folder - If you use open-source Kafka, make sure that
plugin.path
of Kafka Connect configuration contains the folder with connector JAR files
- If you use Confluent Platform, move connector JAR files to the
After you have successfully configured Scylla and Kafka, the next step is to configure the connector and start it up.
If you use Confluent Platform, the easiest way to configure and start up the Scylla CDC Source Connector is to use Confluent Control Center web interface.
-
Open the Confluent Control Center. By default, it is started at port
9021
: -
Click on the cluster you want to start the connector in and open the "Connect" tab:
-
Configure the connector. You need to fill in these required configuration parameters:
- Name: the name of this configuration
- Key converter class, value converter class: converters that determine the format of produced messages. You can read more about them at Kafka Connect Deep Dive – Converters and Serialization Explained
- Hosts: contact points of Scylla
- Namespace: a unique name that identifies the Scylla cluster and that is used as a prefix for all schemas, topics.
- Table names: the names of CDC-enabled tables you want to replicate
For the quickstart example here are the values we will use:
- Name:
QuickstartConnector
- Key converter class, value converter class:
org.apache.kafka.connect.json.JsonConverter
- Hosts:
172.17.0.2:9042
(Scylla started in Docker) - Namespace:
QuickstartConnectorNamespace
- Table names:
quickstart_keyspace.orders
-
Click "Continue" and "Launch"
-
After a short while, a new
QuickstartConnectorNamespace.quickstart_keyspace.orders
topic will be automatically created and inserted rows will be replicated. You can browse them by going to the "Topics" tab, selectingQuickstartConnectorNamespace.quickstart_keyspace.orders
topic, going to "Message" tab and inputting0
to "Jump to offset" field:
-
Start Kafka Connect standalone using this guide. You will have to create a
connector.properties
file with the following contents:name = QuickstartConnector connector.class = com.scylladb.cdc.debezium.connector.ScyllaConnector key.converter = org.apache.kafka.connect.json.JsonConverter value.converter = org.apache.kafka.connect.json.JsonConverter scylla.cluster.ip.addresses = 172.17.0.2:9042 scylla.name = QuickstartConnectorNamespace scylla.table.names = quickstart_keyspace.orders
-
After starting the connector, you can see the generated messages by using
kafka-console-consumer
tool:bin/kafka-console-consumer --bootstrap-server localhost:9092 --topic QuickstartConnectorNamespace.quickstart_keyspace.orders --from-beginning