Try a preview of AI-powered search below!

The Debezium connector for YugabyteDB captures row-level changes in the schemas of a YugabyteDB database.

The first time it connects to a YugabyteDB cluster or universe, the connector takes a consistent snapshot of the tables it is configured for. After that snapshot is complete, the connector continuously captures row-level changes that insert, update, and delete database content that are committed to a YugabyteDB database. The connector generates data change event records and streams them to Kafka topics. For each table, the default behavior is that the connector streams all generated events to a separate Kafka topic for that table. Applications and services consume data change event records from that topic.

Overview

The Debezium connector for YugabyteDB reads the changes produced by YugabyteDB. It uses the CDC service APIs implemented on the server side to get the changes.

The connector produces a change event for every row-level insert, update, and delete operation that was captured, and sends change event records for each table in separate Kafka topics. Client applications read the Kafka topics corresponding to database tables of interest, and can react to every row-level event they receive from those topics.

YugabyteDB normally purges write-ahead log (WAL) segments after some period of time. This means that the connector does not have the complete history of all changes that have been made to the database. Therefore, when the YugabyteDB connector first connects to a particular YugabyteDB database, it starts by taking a snapshot of each of the database schemas. After the connector completes the snapshot, it continues streaming changes from the exact point at which the snapshot was made. This way, the connector starts with a consistent view of all of the data, and does not omit any changes that were made while the snapshot was being taken.

The connector is tolerant of failures. As the connector reads changes and produces events, it records the WAL position for each event. If the connector stops for any reason (including communication failures, network problems, or crashes), upon restart the connector continues reading the WAL where it last left off using the WAL position called checkpoints managed on the Kafka side as well as on the YugabyteDB cluster.

How the connector works

To optimally configure and run a Debezium YugabyteDB connector, it is helpful to understand how the connector performs snapshots, streams change events, determines Kafka topic names, and uses metadata.

Security

Currently, for any user that has the access to the cluster, authentication is done via that user. SSL support-based verification is provided for all the required keys and certificates are passed to the connector.

Snapshots

Most YugabyteDB servers are configured to not retain the complete history of the database in the WAL segments. This means that the YugayteDB connector would be unable to see the entire history of the database by reading only the WAL. Consequently, the first time that the connector starts, it performs an initial consistent snapshot of the database. You can change this behavior by setting the snapshot.mode connector configuration property to a value other than initial.

After the connector completes its initial snapshot, the YugabyteDB connector continues streaming the changes. This ensures that the connector does not miss any updates. If the connector stops again for any reason, upon restart, the connector continues streaming changes from where it previously left off.

Options for the snapshot.mode connector configuration property are as follows:

Streaming changes

The YugabyteDB connector typically spends the vast majority of its time streaming changes from the YugabyteDB server to which it is connected.

The connector keeps polling for changes and whenever there is a change, the connector processes them, converts them to a specific format (Protobuf or JSON in the case of the Debezium plugin) and writes them on an output stream, which can then be consumed by clients.

The Debezium YugabyteDB connector acts as a YugabyteDB client. When the connector receives changes it transforms the events into Debezium create, update, or delete events that include the LSN of the event. The YugabyteDB connector forwards these change events in records to the Kafka Connect framework, which is running in the same process. The Kafka Connect process asynchronously writes the change event records in the same order in which they were generated to the appropriate Kafka topic.

Periodically, Kafka Connect records the most recent offset in another Kafka topic. The offset indicates source-specific position information that Debezium includes with each event.

When Kafka Connect gracefully shuts down, it stops the connectors, and flushes all event records to Kafka. Upon restart, the connector reads the last recorded offset from YugabyteDB server and then it sends a request to the YugabyteDB server to send the events starting just after that position.

Topic names

By default, the YugabyteDB connector writes change events for all INSERT, UPDATE, and DELETE operations that occur in a table to a single Apache Kafka topic that is specific to that table. The connector names change event topics as serverName.schemaName.tableName.

The components of a topic name are as follows:

• serverName - the logical name of the connector, as specified by the database.server.name configuration property.
• schemaName - the name of the database schema in which the change event occurred.
• tableName - the name of the database table in which the change event occurred.

For example, suppose that dbserver is the logical server name in the configuration for a connector that is capturing changes in a YugabyteDB installation that has a yugabyte database and an inventory schema that contains four tables: products, products_on_hand, customers, and orders. The connector would stream records to these four Kafka topics:

• dbserver.inventory.products
• dbserver.inventory.products_on_hand
• dbserver.inventory.customers
• dbserver.inventory.orders

Now suppose that the tables are not part of a specific schema but were created in the default public YugabyteDB schema. The names of the Kafka topics would be:

• dbserver.public.products
• dbserver.public.products_on_hand
• dbserver.public.customers
• dbserver.public.orders

The connector applies similar naming conventions to label its transaction metadata topics.

If the default topic names don't meet your requirements, you can configure custom topic names. To configure custom topic names, you specify regular expressions in the logical topic routing SMT. For more information about using the logical topic routing SMT to customize topic naming, see the Debezium documentation on Topic routing.

Meta information

In addition to the data change event, each record produced by the YugabyteDB connector contains some metadata. Metadata includes information about which tablet caused the change event to occur, the commit time, table, database, offset of the event, for example:

"source": {
  "version": "1.9.5.y.21",
  "connector": "yugabytedb",
  "name": "dbserver1",
  "snapshot": "false",
  "db": "yugabyte",
  "sequence": "[null,\"1:4::0:0\"]",
  "schema": "public",
  "table": "customers",
  "txId": "",
  "lsn": "1:4::0:0"
}

• version is the version number of the connector which is being used.
• name always defaults to the database.server.name connector configuration property.
• db is the database name on which the connector is configured.
• sequence and lsn indicate the offset to which the change event belongs.
• schema is the schema name to which the table belongs.
• table is the name of the table to which the change event belongs.
• txId contains the transaction ID if the change event is a part of any transaction; otherwise it is empty.

Transaction metadata

Debezium can generate events that represent transaction boundaries and that enrich data change event messages.

For every transaction BEGIN and END, Debezium generates an event containing the following fields:

• status - BEGIN or END
• id - string representation of unique transaction identifier
• event_count (for END events) - total number of events emitted by the transaction
• data_collections (for END events) - an array of pairs of data_collection and event_count that provides the number of events emitted by changes originating from given data collection

For example:

{
  "status": "BEGIN",
  "id": "571",
  "event_count": null,
  "data_collections": null
}

{
  "status": "END",
  "id": "571",
  "event_count": 2,
  "data_collections": [
    {
      "data_collection": "s1.a",
      "event_count": 1
    },
    {
      "data_collection": "s2.a",
      "event_count": 1
    }
  ]
}

Unless overridden via the transaction.topic option, transaction events are written to the topic and named database.server.name.transaction.

Change data event enrichment

When transaction metadata is enabled the data message Envelope is enriched with a new transaction field. This field provides information about every event in the form of a composite of fields:

• id - string representation of unique transaction identifier
• total_order - absolute position of the event among all events generated by the transaction
• data_collection_order - the per-data collection position of the event among all events emitted by the transaction

For example:

{
  "before": null,
  "after": {
    "pk": "2",
    "aa": "1"
  },
  "source": {
   ...
  },
  "op": "c",
  "ts_ms": "1580390884335",
  "transaction": {
    "id": "571",
    "total_order": "1",
    "data_collection_order": "1"
  }
}

Data change events

The Debezium YugabyteDB connector generates a data change event for each row-level INSERT, UPDATE, and DELETE operation. Each event contains a key and a value. The structure of the key and the value depends on the table that was changed.

Debezium and Kafka Connect are designed around continuous streams of event messages. However, the structure of these events may change over time, which can be difficult for consumers to handle. To address this, each event contains the schema for its content. This makes each event self-contained.

The following skeleton JSON shows the basic four parts of a change event. However, how you configure the Kafka Connect converter that you choose to use in your application determines the representation of these four parts in change events. A schema field is in a change event only when you configure the converter to produce it. Likewise, the event key and event payload are in a change event only if you configure a converter to produce it.

If you use the JSON converter and you configure it to produce all four basic change event parts, change events have the following structure:

{
 "schema": { --> 1
   ...
  },
 "payload": { --> 2
   ...
 },
 "schema": { --> 3
   ...
  },
 "payload": { --> 4
   ...
 }
}

Change event keys

For a given table, the change event's key has a structure that contains a field for each column in the primary key of the table at the time the event was created.

Consider a customers table defined in the public database schema and the example of a change event key for that table:

CREATE TABLE customers (
  id SERIAL,
  name VARCHAR(255),
  email TEXT,
  PRIMARY KEY(id)
);

Example change event key

If the database.server.name connector configuration property has the value dbserver1, every change event for the customers table while it has this definition has the same key structure, which in JSON looks like the following:

{
  "schema": { --> 1
    "type": "struct",
    "name": "dbserver1.public.customers.Key", --> 2
    "optional": false, --> 3
    "fields": [ --> 4
          {
              "name": "id",
              "index": "0",
              "schema": {
                  "type": "INT32",
                  "optional": "false"
              }
          }
      ]
  },
  "payload": { --> 5
      "id": {
        "value":"1"
      }
  },
}

This change event key has the following structure:

Change event values

The value in a change event is a bit more complicated than the key. Like the key, the value has a schema section and a payload section. The schema section contains the schema that describes the Envelope structure of the payload section, including its nested fields. Change events for operations that create, update or delete data all have a value payload with an envelope structure.

Create events

For a given table, the change event has a structure that contains a field for each column of the table at the time the event was created.

Now suppose a row is inserted to the table:

INSERT INTO customers (name, email) VALUES ('Vaibhav Kushwaha', 'vaibhav@example.com');

The following example shows the value portion of a change event that the connector generates for an operation that creates data in the customers table:

{
  "schema": { --> 1
    "type": "struct",
    "fields": [
      {
        "type": "struct",
        "fields": [ --> 2
          {
            "type": "int32",
            "optional": false,
            "field": "id"
          },
          {
            "type": "string",
            "optional": true,
            "field": "name"
          },
          {
            "type": "string",
            "optional": true,
            "field": "email"
          }
        ],
        "optional": true,
        "name": "dbserver1.public.customers.Value",
        "field": "before"
      },
      {
        "type": "struct",
        "fields": [
          {
            "type": "int32",
            "optional": false,
            "field": "id"
          },
          {
            "type": "string",
            "optional": true,
            "field": "name"
          },
          {
            "type": "string",
            "optional": true,
            "field": "email"
          }
        ],
        "optional": true,
        "name": "dbserver1.public.customers.Value",
        "field": "after"
      },
      {
        "type": "struct",
        "fields": [
          {
            "type": "string",
            "optional": false,
            "field": "version"
          },
          {
            "type": "string",
            "optional": false,
            "field": "connector"
          },
          {
            "type": "string",
            "optional": false,
            "field": "name"
          },
          {
            "type": "int64",
            "optional": false,
            "field": "ts_ms"
          },
          {
            "type": "string",
            "optional": true,
            "name": "io.debezium.data.Enum",
            "version": 1,
            "parameters": {
              "allowed": "true,last,false"
            },
            "default": "false",
            "field": "snapshot"
          },
          {
            "type": "string",
            "optional": false,
            "field": "db"
          },
          {
            "type": "string",
            "optional": true,
            "field": "sequence"
          },
          {
            "type": "string",
            "optional": false,
            "field": "schema"
          },
          {
            "type": "string",
            "optional": false,
            "field": "table"
          },
          {
            "type": "string",
            "optional": true,
            "field": "txId"
          },
          {
            "type": "string",
            "optional": true,
            "field": "lsn"
          },
          {
            "type": "int64",
            "optional": true,
            "field": "xmin"
          }
        ],
        "optional": false,
        "name": "io.debezium.connector.postgresql.Source",
        "field": "source"
      },
      {
        "type": "string",
        "optional": false,
        "field": "op"
      },
      {
        "type": "int64",
        "optional": true,
        "field": "ts_ms"
      },
      {
        "type": "struct",
        "fields": [
          {
            "type": "string",
            "optional": false,
            "field": "id"
          },
          {
            "type": "int64",
            "optional": false,
            "field": "total_order"
          },
          {
            "type": "int64",
            "optional": false,
            "field": "data_collection_order"
          }
        ],
        "optional": true,
        "field": "transaction"
      }
    ],
    "optional": false,
    "name": "dbserver1.public.customers.Envelope"
  },
  "payload": { --> 3
    "before": null, --> 4
    "after": { --> 5
      "id": {
        "value":1
      },
      "name": {
        "value":"Vaibhav Kushwaha"
      },
      "email": {
        "value":"vaibhav@example.com"
      }
    },
    "source": { --> 6
      "version": "1.9.5.y.11",
      "connector": "yugabytedb",
      "name": "dbserver1",
      "ts_ms": -8898156066356,
      "snapshot": "false",
      "db": "yugabyte",
      "sequence": "[null,\"1:4::0:0\"]",
      "schema": "public",
      "table": "customers",
      "txId": "",
      "lsn": "1:4::0:0",
      "xmin": null
    },
    "op": "c", --> 7
    "ts_ms": 1646145062480, --> 8
    "transaction": null
  }
}

The fields in the create event are as follows:

Update events

The value of a change event for an update in the sample customers table has the same schema as a create event for that table. Likewise, the event value's payload has the same structure. However, the event value payload contains different values in an update event.

Note that updating the columns for a row's primary/unique key changes the value of the row's key. When a key changes, Debezium outputs three events: a DELETE event and a tombstone event with the old key for the row, followed by an event with the new key for the row. See Primary key updates on this page for details.

The following example shows a change event value in an event that the connector generates for an update in the customers table:

UPDATE customers SET email = 'service@example.com' WHERE id = 1;

The update event is as follows:

{
  "schema": {...},
  "payload": {
    "before": null, --> 1
    "after": { --> 2
      "id": {
        "value": 1
      },
      "name": {
        "value": "Vaibhav Kushwaha"
      },
      "email": {
        "value": "service@example.com"
      }
    },
    "source": { --> 3
      "version": "1.9.5.y.11",
      "connector": "yugabytedb",
      "name": "dbserver1",
      "ts_ms": -8881476960074,
      "snapshot": "false",
      "db": "yugabyte",
      "sequence": "[null,\"1:5::0:0\"]",
      "schema": "public",
      "table": "customers",
      "txId": "",
      "lsn": "1:5::0:0",
      "xmin": null
    },
    "op": "u", --> 4
    "ts_ms": 1646149134341,
    "transaction": null
  }
}

The fields in the update event are:

Primary key updates

An UPDATE operation that changes a row's primary key field(s) is known as a primary key change. For a primary key change, in place of sending an UPDATE event record, the connector sends a DELETE event record for the old key, and a CREATE event record for the new (updated) key. These events have the usual structure and content, and in addition, each one has a message header related to the primary key change:

• The DELETE event record has __debezium.newkey as a message header. The value of this header is the new primary key for the updated row.

• The CREATE event record has __debezium.oldkey as a message header. The value of this header is the previous (old) primary key for the updated row.

Delete events

The value in a delete change event has the same schema portion as create and update events for the same table. The payload portion in a delete event for the sample customers table looks like the following:

DELETE FROM customers WHERE id = 1;

{
  "schema": {...},
  "payload": {
    "before": { --> 1
      "id": {
        "value": 1
      },
      "name": null,
      "email": null
    },
    "after": null, --> 2
    "source": {
      "version": "1.9.5.y.11",
      "connector": "yugabytedb",
      "name": "dbserver1",
      "ts_ms": -8876894517738,
      "snapshot": "false",
      "db": "yugabyte",
      "sequence": "[null,\"1:6::0:0\"]",
      "schema": "public",
      "table": "customers",
      "txId": "",
      "lsn": "1:6::0:0",
      "xmin": null
    },
    "op": "d", --> 3
    "ts_ms": 1646150253203,
    "transaction": null
  }
}

The fields in this event are:

A delete change event record provides a consumer with the information it needs to process the removal of this row.

Tombstone events

When a row is deleted, the delete event value still works with log compaction, because Kafka can remove all earlier messages that have that same key. However, for Kafka to remove all messages that have that same key, the message value must be null. To make this possible, the YugabyteDB connector follows a delete event with a special tombstone event that has the same key but a null value.

Datatype mappings

The YugabyteDB connector represents changes to rows with events that are structured like the table in which the row exists. The event contains a field for each column value. How that value is represented in the event depends on the YugabyteDB data type of the column. The following sections describe how the connector maps YugabyteDB data types to a literal type and a semantic type in event fields.

• The literal type describes how the value is literally represented using Kafka Connect schema types: INT8, INT16, INT32, INT64, FLOAT32, FLOAT64, BOOLEAN, STRING, BYTES, ARRAY, MAP, and STRUCT.
• The semantic type describes how the Kafka Connect schema captures the meaning of the field using the name of the Kafka Connect schema for the field.

Default values

If there is a default value for any column in a the YugabyteDB database schema, the YugabyteDB Debezium connector propagates the same value to the Kafka schema.

Basic types

The following table describes mappings for YugabyteDB basic data types.

Temporal types

Other than YugabyteDB's TIMESTAMPTZ and TIMETZ data types, which contain time zone information, how temporal types are mapped depends on the value of the time.precision.mode connector configuration property. The following sections describe these mappings:

• time.precision.mode=adaptive
• time.precision.mode=adaptive_time_microseconds
• time.precision.mode=connect

Adaptive mode

When the time.precision.mode property is set to adaptive (the default), the connector determines the literal type and semantic type based on the column's data type definition. This ensures that events exactly represent the values in the database.

The following table describes mappings when time.precision.mode is adaptive.

Adaptive (microseconds) mode

When the time.precision.mode configuration property is set to adaptive_time_microseconds, the connector determines the literal type and semantic type for temporal types based on the column's data type definition. This ensures that events exactly represent the values in the database, except all TIME fields are captured as microseconds.

The following table describes mappings when time.precision.mode is adaptive_time_microseconds.

Connect mode

When the time.precision.mode configuration property is set to connect, the connector uses Kafka Connect logical types. This may be beneficial when consumers can handle only the built-in Kafka Connect logical types and are unable to handle variable-precision time values. However, because YugabyteDB supports microsecond precision, the events generated by a connector with the connect time precision mode results in a loss of precision when the database column has a fractional second precision value that is greater than 3.

The following table describes mappings when time.precision.mode is connect.

TIMESTAMP type

The TIMESTAMP type represents a timestamp without time zone information. Such columns are converted into an equivalent Kafka Connect value based on UTC. For example, the TIMESTAMP value 2022-03-03 16:51:30 is represented by an io.debezium.time.Timestamp with the value 1646326290000 when time.precision.mode is set to any value other than connect.

The timezone of the JVM running Kafka Connect and Debezium does not affect this conversion.

YugabyteDB supports using +/-infinity values in TIMESTAMP columns. These special values are converted to timestamps with value 9223372036825200000 in case of positive infinity or -9223372036832400000 in case of negative infinity.

Decimal types

The setting of the YugabyteDB connector configuration property decimal.handling.mode determines how the connector maps decimal types.

When the decimal.handling.mode property is set to double, the connector represents all DECIMAL, NUMERIC, and MONEY values as Java double values and encodes them as shown in the following table.

The following table describes mappings when decimal.handling.mode is double.

The other possible value for decimal.handling.mode is string. In this case, the connector represents DECIMAL, NUMERIC, and MONEY values as their formatted string representation, and encodes them as shown in the following table.

The following table describes mappings when decimal.handling.mode is string.

Network address types

YugabyteDB has data types that can store IPv4, IPv6, and MAC addresses. You should use these types instead of plain text types to store network addresses, as network address types offer input error checking and specialized operators and functions.

The following table describes mappings for network address types.

Example data type behaviors

Unsupported data types

Support for the following YugabyteDB data types will be enabled in future releases:

• BOX
• CIRCLE
• LINE
• LSEG
• PATH
• PG_LSN
• POINT
• POLYGON
• TSQUERY
• TSVECTOR
• TXID_SNAPSHOT

Setting up YugabyteDB

Before using the YugabyteDB connector to monitor the changes on a YugabyteDB server, you need to ensure the following:

• You have a stream ID created on the database you want to monitor the changes for. The stream can be created using the yb-admin create_change_data_stream command.
• The table which is supposed to be monitored should have a primary key. Only tables which have a primary key can be streamed. See limitations.

WAL disk space consumption

In certain cases, it is possible for YugabyteDB disk space consumed by WAL files to spike or increase out of proportion. There are some possible reasons for this.

For example, the connector is lagging behind in streaming the changes. In this case, the latest checkpoint the connector has received is way behind the last record available in WAL. Because the latest changes are not consumed yet, CDC will not allow the WAL files to be cleaned up, thus causing higher disk consumption. This is the expected behaviour and no action is needed. However, the efficiency of the connector can be increased by increasing the number of tasks so that more processing can happen in parallel.

Deployment

To deploy a Debezium YugabyteDB connector, you install the Debezium YugabyteDB connector archive, configure the connector, and start the connector by adding its configuration to Kafka Connect. For complete steps, follow the guide to running the Debezium connector for YugabyteDB.

Connector configuration example

Following is an example of the configuration for a YugabyteDB connector that connects to a YugabyteDB server on port 5433 at 127.0.0.1, whose logical name is dbserver1. Typically, you configure the Debezium YugabyteDB connector in a JSON file by setting the configuration properties available for the connector.

You can choose to produce events for a subset of the schemas and tables in a database. Optionally, you can ignore, mask, or truncate columns that contain sensitive data, are larger than a specified size, or that you do not need.

{
  "name": "ybconnector", --> 1
  "config": {
    "connector.class": "io.debezium.connector.yugabytedb.YugabyteDBConnector", --> 2
    "database.hostname": "127.0.0.1", --> 3
    "database.port": "5433", --> 4
    "database.master.addresses": "127.0.0.1:7100", --> 5
    "database.streamid": "d540f5e4890c4d3b812933cbfd703ed3", --> 6
    "database.user": "yugabyte", --> 7
    "database.password": "yugabyte", --> 8
    "database.dbname": "yugabyte", --> 9
    "database.server.name": "dbserver1", --> 10
    "table.include.list": "public.test" --> 11
  }
}

1. The name of the connector when registered with a Kafka Connect service.
2. The name of this YugabyteDB connector class.
3. The address of this YugabyteDB server.
4. The port number of the YugabyteDB YSQL process.
5. List of comma separated values of master nodes of the YugabyteDB server. Usually in the form host:port.
6. The DB stream ID created using yb-admin.
7. The name of the YugabyteDB user having the privileges to connect to the database.
8. The password for the above specified YugabyteDB user.
9. The name of the YugabyteDB database to connect to.
10. The logical name of the YugabyteDB server/cluster, which forms a namespace and is used in all the names of the Kafka topics to which the connector writes and the Kafka Connect schema names.
11. A list of all tables hosted by this server that this connector will monitor. This is optional, and there are other properties for listing the schemas and tables to include or exclude from monitoring.

You can send this configuration with a POST command to a running Kafka Connect service. The service records the configuration and starts one connector task that performs the following actions:

• Connects to the YugabyteDB database.
• Reads the transaction log.
• Streams change event records to Kafka topics.

Custom record extractor

YugabyteDB uses a custom record extractor (YBExtractNewRecordState) so that the sinks understand the format in which data is sent. For example, if you are using a JDBC sink connector, you need to add two more properties to the sink configuration:

See Transformers.

Adding connector configuration

To run a YugabyteDB connector, create a connector configuration and add the configuration to your Kafka Connect cluster.

Prerequisites

• YugabyteDB is configured for change data capture.
• The YugabyteDB connector is installed.

Procedure

1. Create a configuration for the YugabyteDB connector.
2. Use the Kafka Connect REST API to add that connector configuration to your Kafka Connect cluster.

Results

After the connector starts, it will perform a snapshot of the tables depending on the configuration and if the connector is set to take snapshots. The connector then starts generating data change events for row-level operations and streaming change event records to Kafka topics.

Connector configuration properties

The Debezium YugabyteDB connector has many configuration properties that you can use to achieve the right connector behavior for your application. Many properties have default values.

The following properties are required unless a default value is available:

Advanced connector configuration properties:

Transformers

The following three transformers are available: YBExtractNewRecordState, ExtractTopic, and PGCompatible.

YBExtractNewRecordState SMT

Transformer type: io.debezium.connector.yugabytedb.transforms.YBExtractNewRecordState

Unlike the Debezium Connector for PostgreSQL, the YugabyteDB connector only sends the after image of the "set of columns" that are modified. PostgreSQL sends the complete after image of the row which has changed. So by default if the column was not changed, it is not a part of the payload that is sent and the default value is set to null.

To differentiate between the case where a column is set to null and the case in which it's not modified, the YugabyteDB connector changes the value type to a struct. In this structure, an unchanged column is {'value': null}, whereas the column changed to a null value is {'value': null, 'set': true}.

A schema registry requires that, once a schema is registered, records must contain only payloads with that schema version. If you're using a schema registry, the YugabyteDB Debezium connector's approach can be problematic, as the schema may change with every message. For example, if we keep changing the record to only include the value of modified columns, the schema of each record will be different (the total number unique schemas will be a result of making all possible combinations of columns) and thus would require sending a schema with every record.

To avoid this problem when you're using a schema registry, use the YBExtractNewRecordState SMT (Single Message Transformer for Kafka), which interprets these values and sends the record in the correct format (by removing the unmodified columns from the JSON message). Records transformed by YBExtractNewRecordState are compatible with all sink implementations. This approach ensures that the schema doesn't change with each new record and it can work with a schema registry.

ExtractTopic

Transformer type: io.aiven.kafka.connect.transforms.ExtractTopic

This transformer extracts a string value from the record and uses it as the topic name.

The transformation can use either the whole key or value (in this case, it must have INT8, INT16, INT32, INT64, FLOAT32, FLOAT32, BOOLEAN, or STRING type) or a field in them (in this case, it must have STRUCT type and the field's value must be INT8, INT16, INT32, INT64, FLOAT32, FLOAT32, BOOLEAN, or STRING).

ExtractTopic exists in two variants:

• io.aiven.kafka.connect.transforms.ExtractTopic$Key - works on keys
• io.aiven.kafka.connect.transforms.ExtractTopic$Value - works on values

The transformation defines the following configurations:

• field.name - The name of the field which should be used as the topic name. If null or empty, the entire key or value is used (and assumed to be a string). By default is null.
• skip.missing.or.null - In case the source of the new topic name is null or missing, should a record be silently passed without transformation. By default, is false.

Here is an example of this transformation configuration:

...
"transforms":"ExtractTopicFromValueField",
"transforms.ExtractTopicFromValueField.type":"io.aiven.kafka.connect.transforms.ExtractTopic$Value",
"transforms.ExtractTopicFromValueField.field.name":"inner_field_name",
...

PGCompatible SMT

Transformer type: io.debezium.connector.yugabytedb.transforms.PGCompatible

By default, the YugabyteDB CDC service publishes events with a schema that only includes columns that have been modified. The source connector then sends the value as null for columns that are missing in the payload. Each column payload includes a set field that is used to signal if a column has been set to null because it wasn't present in the payload from YugabyteDB.

However, some sink connectors may not understand the preceding format. PGCompatible transforms the payload to a format that is compatible with the format of the standard change data events. Specifically, it transforms column schema and value to remove the set field and collapse the payload such that it only contains the data type schema and value.

PGCompatible differs from YBExtractNewRecordState by recursively modifying all the fields in a payload.

Transaction ordering

In a CDC Stream, events from different transactions in different tablets across tables may appear at different times. This works well in use cases such as archiving, or with applications where only eventual consistency is required. There is another class of applications, where the end destination is another OLTP / operational database. These databases can have constraints (such as foreign keys) and strict transactional consistency requirements. In these cases, the stream of events cannot be applied as is, as events of the same transaction may appear out of order because transactions in YugabyteDB can span two tablets.

The YugabyteDB source connector supports transaction ordering, which guarantees consistent streaming of records in the sorted order based on time. With transaction ordering enabled, the connector can be used to stream change events while honoring the constraints.

To use transaction ordering, you need to set the configuration property transaction.ordering to true. Additionally, a transformer ByLogicalTableRouter is required to send all the events to a common topic to ensure that the published change events are published in the same sorted order as they are meant to be.

The following table describes properties for configuring transaction ordering.

For usage example, refer to YugabyteDB CDC Consistent Streaming Pipeline in the example repository.

Transaction boundaries

The connector publishes metadata that can be used to distinguish transaction boundaries for a downstream application to implement atomicity. Once the configuration property provide.transaction.metadata is enabled, the connector will also publish events indicating the beginning and end of the transaction. For more information, see Transaction metadata.

Prerequisites

• Create the Stream ID should in the EXPLICIT checkpointing mode. For more information, see yb-admin create_change_data_stream.
• You should always run the connector with a single task, that is, tasks.max should always be set to 1.

Known limitations

• Transactional ordering is currently not supported with schema evolution. See issue 18476.

Monitoring

For information on monitoring CDC, refer to Monitor.

Behavior when things go wrong

Debezium is a distributed system that captures all changes in multiple upstream databases; it never misses or loses an event. When the system is operating normally or being managed carefully then Debezium provides exactly once delivery of every change event record.

If a fault does happen then the system does not lose any events. However, while it is recovering from the fault, it might repeat some change events. In these abnormal situations, Debezium, like Kafka, provides at least once delivery of change events.

The rest of this section describes how Debezium handles various kinds of faults and problems.

Configuration and startup errors

In the following situations, the connector fails when trying to start, reports an error/exception in the log, and stops running:

• The connector's configuration is invalid.
• The connector cannot successfully connect to YugabyteDB by using the specified connection parameters.
• The connector is restarting from a previously-recorded checkpoint and YugabyteDB no longer has that history available.

In these cases, the error message has details about the problem and possibly a suggested workaround. After you correct the configuration or address the YugabyteDB problem, restart the connector.

YB-TServer becomes unavailable

In case one of the tablet servers crashes, the replicas on other YB-TServer nodes will become the leader for the tablets that were hosted on the crashed server. The YugabyteDB connector will figure out the new tablet leaders and start streaming from the checkpoint the Debezium maintains.

YugabyteDB server failures

In case of YugabyteDB server failures, the Debezium YugabyteDB connector will try for a configurable (using a flag) amount of time for the availability of the YB-TServer and will stop if the cluster cannot start. When the cluster is restarted, the connector can be run again and it will start processing the changes with the committed checkpoint.

Connector unable to find table association with stream ID

In this case, the Connector throws an exception with an error message that the table is not a part of the stream ID.

This can happen in the following 2 scenarios:

• The stream ID you have created might belong to any other database than the one being polled.
• The table you are asking to poll for has no primary keys on it. In this case, the table will not be a part of the stream ID. To continue, add a primary key on the table and create a new stream ID on the database.

YugabyteDB server becomes unavailable

When the connector is running, the YugabyteDB server that it is connected to could become unavailable for any number of reasons. If this happens, the connector fails with an error and stops. When the server is available again, restart the connector.

The YugabyteDB connector externally stores the last processed offset in the form of a checkpoint. After a connector restarts and connects to a server instance, the connector communicates with the server to continue streaming from that particular offset. This offset is available as long as the Debezium replication slot remains intact. Never drop a replication slot on the primary server or you will lose data.

Kafka Connect process stops gracefully

Suppose that Kafka Connect is being run in distributed mode and a Kafka Connect process is stopped gracefully. Prior to shutting down that process, Kafka Connect migrates the process's connector tasks to another Kafka Connect process in that group. The new connector tasks start processing exactly where the prior tasks stopped. There is a short delay in processing while the connector tasks are stopped gracefully and restarted on the new processes.

Kafka Connect process crashes

If the Kafka Connector process stops unexpectedly, any connector tasks it was running terminate without recording their most recently processed offsets. When Kafka Connect is being run in distributed mode, Kafka Connect restarts those connector tasks on other processes. However, YugabyteDB connectors resume from the last offset that was recorded by the earlier processes. This means that the new replacement tasks might generate some of the same change events that were processed just prior to the crash. The number of duplicate events depends on the offset flush period and the volume of data changes just before the crash.

Because there is a chance that some events might be duplicated during a recovery from failure, consumers should always anticipate some duplicate events. Debezium changes are idempotent, so a sequence of events always results in the same state.

In each change event record, Debezium connectors insert source-specific information about the origin of the event, including the YugabyteDB server's time of the event, the ID of the server transaction, and the position in the write-ahead log where the transaction changes were written. Consumers can keep track of this information, especially the LSN, to determine whether an event is a duplicate.

Kafka becomes unavailable

As the connector generates change events, the Kafka Connect framework records those events in Kafka by using the Kafka producer API. Periodically, at a frequency that you specify in the Kafka Connect configuration, Kafka Connect records the latest offset that appears in those change events. If the Kafka brokers become unavailable, the Kafka Connect process that is running the connectors repeatedly tries to reconnect to the Kafka brokers. In other words, the connector tasks pause until a connection can be re-established, at which point the connectors resume exactly where they left off.

Connector is stopped for a duration

If the connector is gracefully stopped, the database can continue to be used. Any changes are recorded in the YugabyteDB WAL. When the connector restarts, it resumes streaming changes where it left off. That is, it generates change event records for all database changes that were made while the connector was stopped.

A properly configured Kafka cluster is able to handle massive throughput. Kafka Connect is written according to Kafka best practices, and given enough resources a Kafka Connect connector can also handle very large numbers of database change events. Because of this, after being stopped for a while, when a Debezium connector restarts, it is very likely to catch up with the database changes that were made while it was stopped. How quickly this happens depends on the capabilities and performance of Kafka and the volume of changes being made to the data in YugabyteDB.