Chapter 2. Installing Change Data Capture connectors
Install Change Data Capture connectors through AMQ Streams by extending Kafka Connect with connector plugins. Following a deployment of AMQ Streams, you can deploy Change Data Capture as a connector configuration through Kafka Connect.
2.1. Prerequisites
A Change Data Capture installation requires the following:
- Red Hat Enterprise Linux version 7.x or 8.x with an x86_64 architecture.
-
Administrative privileges (
sudoaccess). AMQ Streams 1.3 on Red Hat Enterprise Linux is installed on the host machine.
- AMQ Streams must be running in one of the supported JVM versions.
-
Credentials for the
kafkauser that was created when AMQ Streams was installed. An AMQ Streams cluster is running.
- For instructions on running a basic, non-production AMQ Streams cluster containing a single ZooKeeper and a single Kafka node, see Running a single node AMQ Streams cluster.
If you have an earlier version of AMQ Streams, you need to upgrade to AMQ Streams 1.3. For upgrade instructions, see AMQ Streams and Kafka Upgrades.
Additional resources
- For more information about how to install AMQ Streams, see Installing AMQ Streams.
2.2. Kafka topic creation recommendations
Change Data Capture uses multiple Kafka topics for storing data. The topics have to be either created by an administrator, or by Kafka itself by enabling auto-creation for topics using the auto.create.topics.enable broker configuration.
The following list describes limitations and recommendations to consider when creating topics:
- Replication, a factor of at least 3 in production
- Single partition
- Infinite (or very long) retention if topic compaction is disabled
- Log compaction enabled, if you wish to only keep the last change event for a given record
Do not enable topic compaction for the database history topics used by the MySQL and SQL Server connectors.
If you relax the single partition rule, your application must be able to handle out-of-order events for different rows in the database (events for a single row are still fully ordered). If multiple partitions are used, Kafka will determine the partition by hashing the key by default. Other partition strategies require using Simple Message Transforms (SMTs) to set the key for each record.
For log compaction, configure the min.compaction.lag.ms and delete.retention.ms topic-level settings in Apache Kafka so that consumers have enough time to receive all events and delete markers. Specifically, these values should be larger than the maximum downtime you anticipate for the sink connectors, such as when you are updating them.
2.3. Deploying Change Data Capture with AMQ Streams on RHEL
This procedure describes how to set up connectors for Change Data Capture on Red Hat Enterprise Linux. Connectors are deployed to an AMQ Streams cluster using Kafka Connect, a framework for streaming data between Apache Kafka and external systems. Kafka Connect must be run in distributed mode rather than standalone mode.
This procedure assumes that AMQ Streams is installed and ZooKeeper and Kafka are running.
Procedure
- Visit Software Downloads on the Red Hat Customer Portal and download the Change Data Capture connector or connectors that you want to use. For example, download the Debezium 1.0.0 (Technical Preview) MySQL Connector to use Change Data Capture with a MySQL database.
In
/opt/kafka, create theconnector-pluginsdirectory if not already created for other Kafka Connect plugins:$ sudo mkdir /opt/kafka/connector-plugins
Extract the contents of the Change Data Capture connector archive to the
/opt/kafka/connector-pluginsdirectory:$ sudo unzip mysql-connector-java-8.0.16.jar -d /opt/kafka/connector-plugins
- Repeat the above step for each connector that you want to install.
Switch to the
kafkauser:$ su - kafka $ Password:
Check whether Kafka Connect is already running in distributed mode. If it is running, stop the associated process on all Kafka Connect worker nodes. For example:
$ jcmd | grep ConnectDistributed 18514 org.apache.kafka.connect.cli.ConnectDistributed /opt/kafka/config/connect-distributed.properties $ kill 18514
Edit the
connect-distributed.propertiesfile in/opt/kafka/config/and specify the location of the Change Data Capture connector:plugin.path=/opt/kafka/connector-plugins
Run Kafka Connect in distributed mode:
$ /opt/kafka/bin/connect-distributed.sh /opt/kafka/config/connect-distributed.properties
Kafka Connect runs. During startup, Change Data Capture connectors are loaded from the
connector-pluginsdirectory.- Repeat steps 6–8 for each Kafka Connect worker node.
Additional resources
Updating Kafka Connect
If you need to update your deployment, amend the Change Data Capture connector JAR files in the /opt/kafka/connector-plugins directory, and then restart Kafka Connect.
Next Steps
The Change Data Capture User Guide describes how to configure each connector and its source database for change data capture. Once configured, a connector will connect to the source database and produce events for each inserted, updated, and deleted row or document.
