Debezium Connector for Oracle

Overview

Debezium ingests change events from Oracle using the XStream API or directly by LogMiner. In order to use the XStream API, you need to have a license for the GoldenGate product (though it is not required that GoldenGate itself is installed).

Setting up Oracle

The following steps need to be performed in order to prepare the database so the Debezium connector can be used. This assumes the multi-tenancy configuration (with a container database and at least one pluggable database); if you’re not using this model, adjust the steps accordingly.

You can find a template for setting up Oracle in a virtual machine (via Vagrant) in the oracle-vagrant-box/ repository.

Preparing the Database

XStream

Enable GoldenGate replication and archive log mode:

  1. ORACLE_SID=ORCLCDB dbz_oracle sqlplus /nolog
  2. CONNECT sys/top_secret AS SYSDBA
  3. alter system set db_recovery_file_dest_size = 5G;
  4. alter system set db_recovery_file_dest = '/opt/oracle/oradata/recovery_area' scope=spfile;
  5. alter system set enable_goldengate_replication=true;
  6. shutdown immediate
  7. startup mount
  8. alter database archivelog;
  9. alter database open;
  10. -- Should show "Database log mode: Archive Mode"
  11. archive log list
  12. exit;

Furthermore, in order to capture the before state of changed rows, supplemental logging must be enabled for the captured tables or the database in general. E.g. like so for a specific table:

  1. ALTER TABLE inventory.customers ADD SUPPLEMENTAL LOG DATA (ALL) COLUMNS;

LogMiner

The LogMiner implementation requires the database to be configured with a minimum number of redo log groups and files per group. It’s recommended to have a minimum of 5 redo log groups each with 2 redo files per group. Please refer to the Oracle documentation on how to manipulate redo log groups and files for your database version.

Enable archive log mode:

  1. ORACLE_SID=ORACLCDB dbz_oracle sqlplus /nolog
  2. CONNECT sys/top_secret AS SYSDBA
  3. alter system set db_recovery_file_dest_size = 10G;
  4. alter system set db_recovery_file_dest = '/opt/oracle/oradta/recovery_area' scope=spfile;
  5. shutdown immediate
  6. startup mount
  7. alter database archivelog;
  8. alter database open;
  9. -- Should now "Database log mode: Archive Mode"
  10. archive log list
  11. exit;

Furthermore, in order to capture the before state of changed rows, supplemental logging must be enabled for the captured tables or the database in general. E.g. like so for a specific table:

  1. ALTER TABLE inventory.customers ADD SUPPLEMENTAL LOG DATA (ALL) COLUMNS;

Creating Users for the connector

XStream

Create an XStream admin user in the container database (used per Oracle’s recommendation for administering XStream):

  1. sqlplus sys/top_secret@//localhost:1521/ORCLCDB as sysdba
  2. CREATE TABLESPACE xstream_adm_tbs DATAFILE '/opt/oracle/oradata/ORCLCDB/xstream_adm_tbs.dbf'
  3. SIZE 25M REUSE AUTOEXTEND ON MAXSIZE UNLIMITED;
  4. exit;
  1. sqlplus sys/top_secret@//localhost:1521/ORCLPDB1 as sysdba
  2. CREATE TABLESPACE xstream_adm_tbs DATAFILE '/opt/oracle/oradata/ORCLCDB/ORCLPDB1/xstream_adm_tbs.dbf'
  3. SIZE 25M REUSE AUTOEXTEND ON MAXSIZE UNLIMITED;
  4. exit;
  1. sqlplus sys/top_secret@//localhost:1521/ORCLCDB as sysdba
  2. CREATE USER c##xstrmadmin IDENTIFIED BY xsa
  3. DEFAULT TABLESPACE xstream_adm_tbs
  4. QUOTA UNLIMITED ON xstream_adm_tbs
  5. CONTAINER=ALL;
  6. GRANT CREATE SESSION, SET CONTAINER TO c##xstrmadmin CONTAINER=ALL;
  7. BEGIN
  8. DBMS_XSTREAM_AUTH.GRANT_ADMIN_PRIVILEGE(
  9. grantee => 'c##xstrmadmin',
  10. privilege_type => 'CAPTURE',
  11. grant_select_privileges => TRUE,
  12. container => 'ALL'
  13. );
  14. END;
  15. /
  16. exit;

Create XStream user (used by the Debezium connector to connect to the XStream outbound server):

  1. sqlplus sys/top_secret@//localhost:1521/ORCLCDB as sysdba
  2. CREATE TABLESPACE xstream_tbs DATAFILE '/opt/oracle/oradata/ORCLCDB/xstream_tbs.dbf'
  3. SIZE 25M REUSE AUTOEXTEND ON MAXSIZE UNLIMITED;
  4. exit;
  1. sqlplus sys/top_secret@//localhost:1521/ORCLPDB1 as sysdba
  2. CREATE TABLESPACE xstream_tbs DATAFILE '/opt/oracle/oradata/ORCLCDB/ORCLPDB1/xstream_tbs.dbf'
  3. SIZE 25M REUSE AUTOEXTEND ON MAXSIZE UNLIMITED;
  4. exit;
  1. sqlplus sys/top_secret@//localhost:1521/ORCLCDB as sysdba
  2. CREATE USER c##xstrm IDENTIFIED BY xs
  3. DEFAULT TABLESPACE xstream_tbs
  4. QUOTA UNLIMITED ON xstream_tbs
  5. CONTAINER=ALL;
  6. GRANT CREATE SESSION TO c##xstrm CONTAINER=ALL;
  7. GRANT SET CONTAINER TO c##xstrm CONTAINER=ALL;
  8. GRANT SELECT ON V_$DATABASE to c##xstrm CONTAINER=ALL;
  9. GRANT FLASHBACK ANY TABLE TO c##xstrm CONTAINER=ALL;
  10. exit;

LogMiner

Create LogMiner user (used by the Debezium connector to connect to Oracle):

  1. sqlplus sys/top_secret@//localhost:1521/ORCLCDB as sysdba
  2. CREATE TABLESPACE logminer_tbs DATAFILE '/opt/oracle/oradata/ORCLCDB/logminer_tbs.dbf'
  3. SIZE 25M REUSE AUTOEXTEND ON MAXSIZE UNLIMITED;
  4. exit;
  1. sqlplus sys/top_secret@//localhost:1521/ORCLPDB1 as sysdba
  2. CREATE TABLESPACE logminer_tbs DATAFILE '/opt/oracle/oradata/ORCLCDB/ORCLPDB1/logminer_tbs.dbf'
  3. SIZE 25M REUSE AUTOEXTEND ON MAXSIZE UNLIMITED;
  4. exit;
  1. sqlplus sys/top_secret@//localhost:1521/ORCLCDB as sysdba
  2. CREATE USER c##logminer IDENTIFIED BY lm
  3. DEFAULT TABLESPACE logminer_tbs
  4. QUOTA UNLIMITED ON logminer_tbs
  5. CONTAINER=ALL;
  6. GRANT CREATE SESSION TO c##logminer CONTAINER=ALL;
  7. GRANT SET CONTAINER TO c##logminer CONTAINER=ALL;
  8. GRANT SELECT ON V_$DATABASE to c##logminer CONTAINER=ALL;
  9. GRANT FLASHBACK ANY TABLE TO c##logminer CONTAINER=ALL;
  10. GRANT LOGMINING TO c##logminer CONTAINER=ALL;
  11. GRANT LOCK ANY TABLE TO c##logminer CONTAINER=ALL;
  12. exit;

Create an XStream Outbound Server

If you’re using the LogMiner implementation, this step is not necessary.

Create an XStream Outbound server (given the right privileges, this may be done automatically by the connector going forward, see DBZ-721):

  1. sqlplus c##xstrmadmin/xsa@//localhost:1521/ORCLCDB
  2. DECLARE
  3. tables DBMS_UTILITY.UNCL_ARRAY;
  4. schemas DBMS_UTILITY.UNCL_ARRAY;
  5. BEGIN
  6. tables(1) := NULL;
  7. schemas(1) := 'debezium';
  8. DBMS_XSTREAM_ADM.CREATE_OUTBOUND(
  9. server_name => 'dbzxout',
  10. table_names => tables,
  11. schema_names => schemas);
  12. END;
  13. /
  14. exit;

Alter the XStream Outbound server to allow the xstrm user to connect to it:

  1. sqlplus sys/top_secret@//localhost:1521/ORCLCDB as sysdba
  2. BEGIN
  3. DBMS_XSTREAM_ADM.ALTER_OUTBOUND(
  4. server_name => 'dbzxout',
  5. connect_user => 'c##xstrm');
  6. END;
  7. /
  8. exit;

Note that a given outbound server must not be used by multiple connector instances at the same time. If you wish to set up multiple instances of the Debezium Oracle connector, a specific XStreamOutbound server is needed for each of them.

Supported Configurations

So far, the connector has been tested with the pluggable database set-up (CDB/PDB model). It should monitor a single PDB in this model. It should also work with traditional (non-CDB) set-ups, though this could not be tested so far.

How the Oracle Connector Works

Database Schema History

tbd.

Snapshots

Most Oracle servers are configured to not retain the complete history of the database in the redo logs, so the Debezium Oracle connector would be unable to see the entire history of the database by simply reading the logs. So, by default (snapshotting mode initial) the connector will upon first startup perform an initial consistent snapshot of the database (meaning the structure and data within any tables to be captured as per the connector’s filter configuration).

Each snapshot consists of the following steps:

  1. Determine the tables to be captured

  2. Obtain an IN EXCLUSIVE MODE lock on each of the monitored tables to ensure that no structural changes can occur to any of the tables.

  3. Read the current SCN (“system change number”) position in the server’s redo log.

  4. Capture the structure of all relevant tables.

  5. Release the locks obtained in step 2, i.e. the locks are held only for a short period of time.

  6. Scan all of the relevant database tables and schemas as valid at the SCN position read in step 3 (SELECT * FROM …​ AS OF SCN 123), and generate a READ event for each row and write that event to the appropriate table-specific Kafka topic.

  7. Record the successful completion of the snapshot in the connector offsets.

If the connector fails, is rebalanced, or stops after step 1 begins but before step 7 completes, upon restart the connector will begin a new snapshot. Once the Oracle connector does complete its initial snapshot, it continues streaming from the position read during step 3, ensuring that it does not miss any updates that occurred while the snapshot was taken. If the connector stops again for any reason, upon restart it will simply continue streaming changes from where it previously left off.

A second snapshotting mode is schema_only. In this case, step 6 from the snapshotting routine described above is not applied. In other words, the connector still captures the structure of the relevant tables, but it does not create any READ events representing the complete dataset at the point of connector start-up. This can be useful if you are interested in data changes only from now onwards but not the complete current state of all records.

Reading the Redo Log

Upon first start-up, the connector takes a snapshot of the structure of the captured tables (DDL) and persists this information in its internal database history topic. It then proceeds to listen for change events right from the SCN at which the schema structure was captured. Processed SCNs are passed as offsets to Kafka Connect and regularly acknowledged with the database server (allowing it to discard older log files). After restart, the connector will resume from the offset (SCN) where it left off before.

Topics Names

Schema Change Topic

The format of the schema change topic messages is in incubating state and it can change without further notice.

Oracle connector stores the historic schema structure of database tables in a database history topic. This topic should be considered as an internal state of the connector and should not be used by the user. If the application needs to track changes in the source database there is the public-facing schema change topic. The topic name is the same as the logical server name configured in connector configuration.

Debezium emits a new message to this topic whenever a new table is streamed from or when the structure of the table is altered (schema evolution procedure must be followed). The message contains a logical representation of the table schema.

The example of the message is:

  1. {
  2. "schema": {
  3. ...
  4. },
  5. "payload": {
  6. "source": {
  7. "version": "1.3.1.Final",
  8. "connector": "oracle",
  9. "name": "server1",
  10. "ts_ms": 1588252618953,
  11. "snapshot": "true",
  12. "db": "ORCLPDB1",
  13. "schema": "DEBEZIUM",
  14. "table": "CUSTOMERS",
  15. "txId" : null,
  16. "scn" : "1513734",
  17. "commit_scn": "1513734",
  18. "lcr_position" : null
  19. },
  20. "databaseName": "ORCLPDB1",
  21. "schemaName": "DEBEZIUM",
  22. "ddl": "CREATE TABLE \"DEBEZIUM\".\"CUSTOMERS\" \n ( \"ID\" NUMBER(9,0) NOT NULL ENABLE, \n \"FIRST_NAME\" VARCHAR2(255), \n \"LAST_NAME" VARCHAR2(255), \n \"EMAIL\" VARCHAR2(255), \n PRIMARY KEY (\"ID\") ENABLE, \n SUPPLEMENTAL LOG DATA (ALL) COLUMNS\n ) SEGMENT CREATION IMMEDIATE \n PCTFREE 10 PCTUSED 40 INITRANS 1 MAXTRANS 255 \n NOCOMPRESS LOGGING\n STORAGE(INITIAL 65536 NEXT 1048576 MINEXTENTS 1 MAXEXTENTS 2147483645\n PCTINCREASE 0 FREELISTS 1 FREELIST GROUPS 1\n BUFFER_POOL DEFAULT FLASH_CACHE DEFAULT CELL_FLASH_CACHE DEFAULT)\n TABLESPACE \"USERS\" ",
  23. "tableChanges": [
  24. {
  25. "type": "CREATE",
  26. "id": "\"ORCLPDB1\".\"DEBEZIUM\".\"CUSTOMERS\"",
  27. "table": {
  28. "defaultCharsetName": null,
  29. "primaryKeyColumnNames": [
  30. "ID"
  31. ],
  32. "columns": [
  33. {
  34. "name": "ID",
  35. "jdbcType": 2,
  36. "nativeType": null,
  37. "typeName": "NUMBER",
  38. "typeExpression": "NUMBER",
  39. "charsetName": null,
  40. "length": 9,
  41. "scale": 0,
  42. "position": 1,
  43. "optional": false,
  44. "autoIncremented": false,
  45. "generated": false
  46. },
  47. {
  48. "name": "FIRST_NAME",
  49. "jdbcType": 12,
  50. "nativeType": null,
  51. "typeName": "VARCHAR2",
  52. "typeExpression": "VARCHAR2",
  53. "charsetName": null,
  54. "length": 255,
  55. "scale": null,
  56. "position": 2,
  57. "optional": false,
  58. "autoIncremented": false,
  59. "generated": false
  60. },
  61. {
  62. "name": "LAST_NAME",
  63. "jdbcType": 12,
  64. "nativeType": null,
  65. "typeName": "VARCHAR2",
  66. "typeExpression": "VARCHAR2",
  67. "charsetName": null,
  68. "length": 255,
  69. "scale": null,
  70. "position": 3,
  71. "optional": false,
  72. "autoIncremented": false,
  73. "generated": false
  74. },
  75. {
  76. "name": "EMAIL",
  77. "jdbcType": 12,
  78. "nativeType": null,
  79. "typeName": "VARCHAR2",
  80. "typeExpression": "VARCHAR2",
  81. "charsetName": null,
  82. "length": 255,
  83. "scale": null,
  84. "position": 4,
  85. "optional": false,
  86. "autoIncremented": false,
  87. "generated": false
  88. }
  89. ]
  90. }
  91. }
  92. ]
  93. }
  94. }

The fields

  • databaseName and schemaName describe which database/schema has been affected

  • ddl contains the DDL responsible for the schema change

  • tableChanges array contains one or more schema changes generated by the DDL command

  • type describes the kind of the change: CREATE - table created, ALTER - table modified, DROP - table deleted

  • id is the full identifier of the table

  • table represents table metadata after the applied change

  • primaryKeyColumnNames is a list of columns that compose the table primary key

  • columns are column metadata of each of the columns

The schema change messages use as the key the name of the database to which the changes apply:

  1. {
  2. "schema": {
  3. "type": "struct",
  4. "fields": [
  5. {
  6. "type": "string",
  7. "optional": false,
  8. "field": "databaseName"
  9. }
  10. ],
  11. "optional": false,
  12. "name": "io.debezium.connector.oracle.SchemaChangeKey"
  13. },
  14. "payload": {
  15. "databaseName": "ORCLPDB1"
  16. }
  17. }

Events

All data change events produced by the Oracle connector have a key and a value, although the structure of the key and value depend on the table from which the change events originated (see Topic names).

The Debezium Oracle connector ensures that all Kafka Connect schema names are valid Avro schema names. This means that the logical server name must start with Latin letters or an underscore (e.g., [a-z,A-Z,]), and the remaining characters in the logical server name and all characters in the schema and table names must be Latin letters, digits, or an underscore (e.g., [a-z,A-Z,0-9,\]). If not, then all invalid characters will automatically be replaced with an underscore character.

This can lead to unexpected conflicts when the logical server name, schema names, and table names contain other characters, and the only distinguishing characters between table full names are invalid and thus replaced with underscores.

Debezium and Kafka Connect are designed around continuous streams of event messages, and the structure of these events may change over time. This could be difficult for consumers to deal with, so to make it easy Kafka Connect makes each event self-contained. Every message key and value has two parts: a schema and payload. The schema describes the structure of the payload, while the payload contains the actual data.

Change Event Keys

For a given table, the change event’s key will have a structure that contains a field for each column in the primary key (or unique key constraint) of the table at the time the event was created.

Consider a customers table defined in the inventory database schema:

  1. CREATE TABLE customers (
  2. id NUMBER(9) GENERATED BY DEFAULT ON NULL AS IDENTITY (START WITH 1001) NOT NULL PRIMARY KEY,
  3. first_name VARCHAR2(255) NOT NULL,
  4. last_name VARCHAR2(255) NOT NULL,
  5. email VARCHAR2(255) NOT NULL UNIQUE
  6. );

If the database.server.name configuration property has the value server1, every change event for the customers table while it has this definition will feature the same key structure, which in JSON looks like this:

  1. {
  2. "schema": {
  3. "type": "struct",
  4. "fields": [
  5. {
  6. "type": "int32",
  7. "optional": false,
  8. "field": "ID"
  9. }
  10. ],
  11. "optional": false,
  12. "name": "server1.INVENTORY.CUSTOMERS.Key"
  13. },
  14. "payload": {
  15. "ID": 1004
  16. }
  17. }

The schema portion of the key contains a Kafka Connect schema describing what is in the key portion, and in our case that means that the payload value is not optional, is a structure defined by a schema named server1.DEBEZIUM.CUSTOMERS.Key, and has one required field named id of type int32. If you look at the value of the key’s payload field, you can see that it is indeed a structure (which in JSON is just an object) with a single id field, whose value is 1004.

Therefore, you can interpret this key as describing the row in the inventory.customers table (output from the connector named server1) whose id primary key column had a value of 1004.

Change Event Values

Like the message key, the value of a change event message has a schema section and payload section. The payload section of every change event value produced by the Oracle connector has an envelope structure with the following fields:

  • op is a mandatory field that contains a string value describing the type of operation. Values for the Oracle connector are c for create (or insert), u for update, d for delete, and r for read (in the case of a snapshot).

  • before is an optional field that if present contains the state of the row before the event occurred. The structure will be described by the server1.INVENTORY.CUSTOMERS.Value Kafka Connect schema, which the server1 connector uses for all rows in the inventory.customers table.

Whether or not this field and its elements are available is highly dependent on the Supplemental Logging configuration applying to the table.

  • after is an optional field that if present contains the state of the row after the event occurred. The structure is described by the same server1.INVENTORY.CUSTOMERS.Value Kafka Connect schema used in before.

  • source is a mandatory field that contains a structure describing the source metadata for the event, which in the case of Oracle contains these fields: the Debezium version, the connector name, whether the event is part of an ongoing snapshot or not, the transaction id (not while snapshotting), the SCN of the change, and a timestamp representing the point in time when the record was changed in the source database (during snapshotting, this is the point in time of snapshotting).

The commit_scn field is optional and describes the SCN of the transaction commit that the change event participates within. This field is only present when using the LogMiner connection adapter.

  • ts_ms is optional and if present contains the time (using the system clock in the JVM running the Kafka Connect task) at which the connector processed the event.

And of course, the schema portion of the event message’s value contains a schema that describes this envelope structure and the nested fields within it.

Create events

Let’s look at what a create event value might look like for our customers table:

  1. {
  2. "schema": {
  3. "type": "struct",
  4. "fields": [
  5. {
  6. "type": "struct",
  7. "fields": [
  8. {
  9. "type": "int32",
  10. "optional": false,
  11. "field": "ID"
  12. },
  13. {
  14. "type": "string",
  15. "optional": false,
  16. "field": "FIRST_NAME"
  17. },
  18. {
  19. "type": "string",
  20. "optional": false,
  21. "field": "LAST_NAME"
  22. },
  23. {
  24. "type": "string",
  25. "optional": false,
  26. "field": "EMAIL"
  27. }
  28. ],
  29. "optional": true,
  30. "name": "server1.DEBEZIUM.CUSTOMERS.Value",
  31. "field": "before"
  32. },
  33. {
  34. "type": "struct",
  35. "fields": [
  36. {
  37. "type": "int32",
  38. "optional": false,
  39. "field": "ID"
  40. },
  41. {
  42. "type": "string",
  43. "optional": false,
  44. "field": "FIRST_NAME"
  45. },
  46. {
  47. "type": "string",
  48. "optional": false,
  49. "field": "LAST_NAME"
  50. },
  51. {
  52. "type": "string",
  53. "optional": false,
  54. "field": "EMAIL"
  55. }
  56. ],
  57. "optional": true,
  58. "name": "server1.DEBEZIUM.CUSTOMERS.Value",
  59. "field": "after"
  60. },
  61. {
  62. "type": "struct",
  63. "fields": [
  64. {
  65. "type": "string",
  66. "optional": true,
  67. "field": "version"
  68. },
  69. {
  70. "type": "string",
  71. "optional": false,
  72. "field": "name"
  73. },
  74. {
  75. "type": "int64",
  76. "optional": true,
  77. "field": "ts_ms"
  78. },
  79. {
  80. "type": "string",
  81. "optional": true,
  82. "field": "txId"
  83. },
  84. {
  85. "type": "int64",
  86. "optional": true,
  87. "field": "scn"
  88. },
  89. {
  90. "type": "int64",
  91. "optional": true,
  92. "field": "commit_scn"
  93. },
  94. {
  95. "type": "boolean",
  96. "optional": true,
  97. "field": "snapshot"
  98. }
  99. ],
  100. "optional": false,
  101. "name": "io.debezium.connector.oracle.Source",
  102. "field": "source"
  103. },
  104. {
  105. "type": "string",
  106. "optional": false,
  107. "field": "op"
  108. },
  109. {
  110. "type": "int64",
  111. "optional": true,
  112. "field": "ts_ms"
  113. }
  114. ],
  115. "optional": false,
  116. "name": "server1.DEBEZIUM.CUSTOMERS.Envelope"
  117. },
  118. "payload": {
  119. "before": null,
  120. "after": {
  121. "ID": 1004,
  122. "FIRST_NAME": "Anne",
  123. "LAST_NAME": "Kretchmar",
  124. "EMAIL": "annek@noanswer.org"
  125. },
  126. "source": {
  127. "version": "0.9.0.Alpha1",
  128. "name": "server1",
  129. "ts_ms": 1520085154000,
  130. "txId": "6.28.807",
  131. "scn": 2122185,
  132. "commit_scn": 2122185,
  133. "snapshot": false
  134. },
  135. "op": "c",
  136. "ts_ms": 1532592105975
  137. }
  138. }

If we look at the schema portion of this event’s value, we can see the schema for the envelope, the schema for the source structure (which is specific to the Oracle connector and reused across all events), and the table-specific schemas for the before and after fields.

The names of the schemas for the before and after fields are of the form logicalName.schemaName.tableName.Value, and thus are entirely independent from all other schemas for all other tables. This means that when using the Avro Converter, the resulting Avro schems for each table in each logical source have their own evolution and history.

If we look at the payload portion of this event’s value, we can see the information in the event, namely that it is describing that the row was created (since op=c), and that the after field value contains the values of the new inserted row’s’ ID, FIRST_NAME, LAST_NAME, and EMAIL columns.

It may appear that the JSON representations of the events are much larger than the rows they describe. This is true, because the JSON representation must include the schema and the payload portions of the message. It is possible and even recommended to use the Avro Converter to dramatically decrease the size of the actual messages written to the Kafka topics.

Update events

The value of an update change event on this table will actually have the exact same schema, and its payload will be structured the same but will hold different values. Here’s an example:

  1. {
  2. "schema": { ... },
  3. "payload": {
  4. "before": {
  5. "ID": 1004,
  6. "FIRST_NAME": "Anne",
  7. "LAST_NAME": "Kretchmar",
  8. "EMAIL": "annek@noanswer.org"
  9. },
  10. "after": {
  11. "ID": 1004,
  12. "FIRST_NAME": "Anne",
  13. "LAST_NAME": "Kretchmar",
  14. "EMAIL": "anne@example.com"
  15. },
  16. "source": {
  17. "version": "0.9.0.Alpha1",
  18. "name": "server1",
  19. "ts_ms": 1520085811000,
  20. "txId": "6.9.809",
  21. "scn": 2125544,
  22. "commit_scn": 2125544,
  23. "snapshot": false
  24. },
  25. "op": "u",
  26. "ts_ms": 1532592713485
  27. }
  28. }

When we compare this to the value in the insert event, we see a couple of differences in the payload section:

  • The op field value is now u, signifying that this row changed because of an update

  • The before field now has the state of the row with the values before the database commit

  • The after field now has the updated state of the row, and here was can see that the EMAIL value is now anne@example.com.

  • The source field structure has the same fields as before, but the values are different since this event is from a different position in the redo log.

  • The ts_ms shows the timestamp that Debezium processed this event.

There are several things we can learn by just looking at this payload section. We can compare the before and after structures to determine what actually changed in this row because of the commit. The source structure tells us information about Oracle’s record of this change (providing traceability), but more importantly this has information we can compare to other events in this and other topics to know whether this event occurred before, after, or as part of the same Oracle commit as other events.

When the columns for a row’s primary/unique key are updated, the value of the row’s key has changed so Debezium will output three events: a DELETE event and a tombstone event with the old key for the row, followed by an INSERT event with the new key for the row.

Delete events

So far we’ve seen samples of create and update events. Now, let’s look at the value of a delete event for the same table. Once again, the schema portion of the value will be exactly the same as with the create and update events:

  1. {
  2. "schema": { ... },
  3. "payload": {
  4. "before": {
  5. "ID": 1004,
  6. "FIRST_NAME": "Anne",
  7. "LAST_NAME": "Kretchmar",
  8. "EMAIL": "anne@example.com"
  9. },
  10. "after": null,
  11. "source": {
  12. "version": "0.9.0.Alpha1",
  13. "name": "server1",
  14. "ts_ms": 1520085153000,
  15. "txId": "6.28.807",
  16. "scn": 2122184,
  17. "commit_scn": 2122184,
  18. "snapshot": false
  19. },
  20. "op": "d",
  21. "ts_ms": 1532592105960
  22. }
  23. }

If we look at the payload portion, we see a number of differences compared with the create or update event payloads:

  • The op field value is now d, signifying that this row was deleted

  • The before field now has the state of the row that was deleted with the database commit.

  • The after field is null, signifying that the row no longer exists

  • The source field structure has many of the same values as before, except the ts_ms, scn and txId fields have changed

  • The ts_ms shows the timestamp that Debezium processed this event.

This event gives a consumer all kinds of information that it can use to process the removal of this row.

The Oracle connector’s events are designed to work with Kafka log compaction, which allows for the removal of some older messages as long as at least the most recent message for every key is kept. This allows Kafka to reclaim storage space while ensuring the topic contains a complete dataset and can be used for reloading key-based state.

When a row is deleted, the delete event value listed above still works with log compaction, since Kafka can still remove all earlier messages with that same key. But only if the message value is null will Kafka know that it can remove all messages with that same key. To make this possible, Debezium’s Oracle connector always follows the delete event with a special tombstone event that has the same key but null value.

Transaction Metadata

Debezium can generate events that represents tranaction metadata boundaries and enrich data messages.

Transaction boundaries

Debezium generates events for every transaction BEGIN and END. Every event contains

  • status - BEGIN or END

  • id - string representation of unique transaction identifier

  • event_count (for END events) - total number of events emmitted by the transaction

  • data_collections (for END events) - an array of pairs of data_collection and event_count that provides number of events emitted by changes originating from given data collection

Following is an example of what a message looks like:

  1. {
  2. "status": "BEGIN",
  3. "id": "5.6.641",
  4. "event_count": null,
  5. "data_collections": null
  6. }
  7. {
  8. "status": "END",
  9. "id": "5.6.641",
  10. "event_count": 2,
  11. "data_collections": [
  12. {
  13. "data_collection": "ORCLPDB1.DEBEZIUM.CUSTOMER",
  14. "event_count": 1
  15. },
  16. {
  17. "data_collection": "ORCLPDB1.DEBEZIUM.ORDER",
  18. "event_count": 1
  19. }
  20. ]
  21. }

The transaction events are written to the topic named <database.server.name>.transaction.

Data events 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 - the 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 that were emitted by the transaction

Following is an example of what a message looks like:

  1. {
  2. "before": null,
  3. "after": {
  4. "pk": "2",
  5. "aa": "1"
  6. },
  7. "source": {
  8. ...
  9. },
  10. "op": "c",
  11. "ts_ms": "1580390884335",
  12. "transaction": {
  13. "id": "5.6.641",
  14. "total_order": "1",
  15. "data_collection_order": "1"
  16. }
  17. }

Data Types

As described above, the Debezium Oracle connector represents the changes to rows with events that are structured like the table in which the row exist. The event contains a field for each column value, and how that value is represented in the event depends on the Oracle data type of the column. This section describes this mapping from Oracle’s data types to a literal type and semantic type within the events’ fields.

Here, the literal type describes how the value is literally represented using Kafka Connect schema types, namely 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.

Support for further data types will be added in subsequent releases. Please file a JIRA issue for any specific types you are missing.

Character Values

Oracle Data Type

Literal type (schema type)

Semantic type (schema name)

Notes

CHAR[(M)]

STRING

n/a

NCHAR[(M)]

STRING

n/a

VARCHAR[(M)]

STRING

n/a

VARCHAR2[(M)]

STRING

n/a

NVARCHAR2[(M)]

STRING

n/a

Numeric Values

Oracle Data Type

Literal type (schema type)

Semantic type (schema name)

Notes

NUMBER[(P[, *])]

STRUCT

io.debezium.data.VariableScaleDecimal

Contains a structure with two fields: scale of type INT32 that contains the scale of the transferred value and value of type BYTES containing the original value in an unscaled form.

NUMBER(P, S > 0)

BYTES

org.apache.kafka.connect.data.Decimal

NUMBER(P, S ⇐ 0)

INT8 / INT16 / INT32 / INT64

n/a

NUMBER columns with a scale of 0 represent integer numbers; a negative scale indicates rounding in Oracle, e.g. a scale of -2 will cause rounding to hundreds.
Depending on the precision and scale, a matching Kafka Connect integer type will be chosen: INT8 if P - S < 3, INT16 if P - S < 5, INT32 if P - S < 10 and INT64 if P - S < 19.
If P - S >= 19, the column will be mapped to BYTES (org.apache.kafka.connect​.data.Decimal).

SMALLINT

BYTES

org.apache.kafka.connect.data.Decimal

SMALLINT is mapped in Oracle to NUMBER(38,0) and hence can hold values larger than any of the INT types could store

INTEGER, INT

BYTES

org.apache.kafka.connect.data.Decimal

INTEGER is mapped in Oracle to NUMBER(38,0) and hence can hold values larger than any of the INT types could store

NUMERIC[(P, S)]

BYTES / INT8 / INT16 / INT32 / INT64

org.apache.kafka.connect.data.Decimal if using BYTES

Handled equivalently to NUMBER (note that S defaults to 0 for NUMERIC).

DECIMAL[(P, S)]

BYTES / INT8 / INT16 / INT32 / INT64

org.apache.kafka.connect.data.Decimal if using BYTES

Handled equivalently to NUMBER (note that S defaults to 0 for DECIMAL).

BINARY_FLOAT

FLOAT32

n/a

BINARY_DOUBLE

FLOAT64

n/a

FLOAT[(P)]

STRUCT

io.debezium.data.VariableScaleDecimal

Contains a structure with two fields: scale of type INT32 that contains the scale of the transferred value and value of type BYTES containing the original value in an unscaled form.

DOUBLE PRECISION

STRUCT

io.debezium.data.VariableScaleDecimal

Contains a structure with two fields: scale of type INT32 that contains the scale of the transferred value and value of type BYTES containing the original value in an unscaled form.

REAL

STRUCT

io.debezium.data.VariableScaleDecimal

Contains a structure with two fields: scale of type INT32 that contains the scale of the transferred value and value of type BYTES containing the original value in an unscaled form.

Decimal Values

When decimal.handling.mode configuration property is set to precise, then the connector will use the predefined Kafka Connect org.apache.kafka.connect.data.Decimal or io.debezium.data.VariableScaleDecimal logical types for numeric columns as described above. This is the default mode.

However, when decimal.handling.mode configuration property is set to double, then the connector will represent the values as Java double values with schema type FLOAT64. The last option for decimal.handling.mode configuration property is string. In this case the connector will represent the values as their formatted string representation with schema type STRING.

Temporal Values

Oracle Data Type

Literal type (schema type)

Semantic type (schema name)

Notes

DATE

INT64

io.debezium.time.Timestamp

Represents the number of milliseconds past epoch, and does not include timezone information.

TIMESTAMP(0 - 3)

INT64

io.debezium.time.Timestamp

Represents the number of milliseconds past epoch, and does not include timezone information.

TIMESTAMP, TIMESTAMP(4 - 6)

INT64

io.debezium.time.MicroTimestamp

Represents the number of microseconds past epoch, and does not include timezone information.

TIMESTAMP(7 - 9)

INT64

io.debezium.time.NanoTimestamp

Represents the number of nanoseconds past epoch, and does not include timezone information.

TIMESTAMP WITH TIME ZONE

STRING

io.debezium.time.ZonedTimestamp

A string representation of a timestamp with timezone information

INTERVAL

FLOAT64

io.debezium.time.MicroDuration

The number of micro seconds for a time interval using the 365.25 / 12.0 formula for days per month average

Deploying a Connector

Due to licensing requirements, the Debezium Oracle Connector does not ship with the Oracle JDBC driver and the XStream API JAR. You can obtain them for free by downloading the Oracle Instant Client.

Extract the archive into a directory, e.g. /path/to/instant_client/. Copy the files _ojdbc8.jar and xstreams.jar from the Instant Client into Kafka’s libs directory. Create the environment variable LD_LIBRARY_PATH, pointing to the Instant Client directory:

  1. LD_LIBRARY_PATH=/path/to/instant_client/

Example Configuration

The following shows an example JSON request for registering an instance of the Debezium Oracle connector:

  1. {
  2. "name": "inventory-connector",
  3. "config": {
  4. "connector.class" : "io.debezium.connector.oracle.OracleConnector",
  5. "tasks.max" : "1",
  6. "database.server.name" : "server1",
  7. "database.hostname" : "<oracle ip>",
  8. "database.port" : "1521",
  9. "database.user" : "c##xstrm",
  10. "database.password" : "xs",
  11. "database.dbname" : "ORCLCDB",
  12. "database.pdb.name" : "ORCLPDB1",
  13. "database.out.server.name" : "dbzxout",
  14. "database.history.kafka.bootstrap.servers" : "kafka:9092",
  15. "database.history.kafka.topic": "schema-changes.inventory"
  16. }
  17. }

Selecting the adapter

Debezium provides multiple ways to ingest change events from Oracle. By default Debezium uses the XStream API but this isn’t always applicable for every installation.

The following example configuration illustrates that by adding the database.connection.adapter, the connector can be toggled to use the LogMiner implementation.

  1. {
  2. "name": "inventory-connector",
  3. "config": {
  4. "connector.class" : "io.debezium.connector.oracle.OracleConnector",
  5. "tasks.max" : "1",
  6. "database.server.name" : "server1",
  7. "database.hostname" : "<oracle ip>",
  8. "database.port" : "1521",
  9. "database.user" : "c##xstrm",
  10. "database.password" : "xs",
  11. "database.dbname" : "ORCLCDB",
  12. "database.pdb.name" : "ORCLPDB1",
  13. "database.out.server.name" : "dbzxout",
  14. "database.history.kafka.bootstrap.servers" : "kafka:9092",
  15. "database.history.kafka.topic": "schema-changes.inventory"
  16. "database.connection.adapter": "logminer"
  17. }
  18. }

We do encourage the use of the LogMiner implementation for testing purposes and providing us with your feedback, but we do not yet recommend its use in production as its still under active development.

Monitoring

The Debezium Oracle connector has three metric types in addition to the built-in support for JMX metrics that Zookeeper, Kafka, and Kafka Connect have.

Please refer to the monitoring documentation for details of how to expose these metrics via JMX.

Snapshot Metrics

The MBean is debezium.oracle:type=connector-metrics,context=snapshot,server=*<database.server.name>*.

AttributesTypeDescription

string

The last snapshot event that the connector has read.

long

The number of milliseconds since the connector has read and processed the most recent event.

long

The total number of events that this connector has seen since last started or reset.

long

The number of events that have been filtered by include/exclude list filtering rules configured on the connector.

string[]

The list of tables that are monitored by the connector.

int

The length the queue used to pass events between the snapshotter and the main Kafka Connect loop.

int

The free capacity of the queue used to pass events between the snapshotter and the main Kafka Connect loop.

int

The total number of tables that are being included in the snapshot.

int

The number of tables that the snapshot has yet to copy.

boolean

Whether the snapshot was started.

boolean

Whether the snapshot was aborted.

boolean

Whether the snapshot completed.

long

The total number of seconds that the snapshot has taken so far, even if not complete.

Map<String, Long>

Map containing the number of rows scanned for each table in the snapshot. Tables are incrementally added to the Map during processing. Updates every 10,000 rows scanned and upon completing a table.

Streaming Metrics

The MBean is debezium.oracle:type=connector-metrics,context=streaming,server=*<database.server.name>*.

AttributesTypeDescription

string

The last streaming event that the connector has read.

long

The number of milliseconds since the connector has read and processed the most recent event.

long

The total number of events that this connector has seen since last started or reset.

long

The number of events that have been filtered by include/exclude list filtering rules configured on the connector.

string[]

The list of tables that are monitored by the connector.

int

The length the queue used to pass events between the streamer and the main Kafka Connect loop.

int

The free capacity of the queue used to pass events between the streamer and the main Kafka Connect loop.

boolean

Flag that denotes whether the connector is currently connected to the database server.

long

The number of milliseconds between the last change event’s timestamp and the connector processing it. The values will incoporate any differences between the clocks on the machines where the database server and the connector are running.

long

The number of processed transactions that were committed.

Map<String, String>

The coordinates of the last received event.

string

Transaction identifier of the last processed transaction.

LogMiner Metrics

The MBean is debezium.oracle:type=connector-metrics,context=log-miner,server=*<database.server.name>*.

AttributesTypeDescription

long

The most recent SCN has has been processed.

int

The number of DML operations observed.

string[]

The current redo log filename.

string[]

The current status of the redo logs.

long

The number of times the redo logs have switched.

Duration

The duration it took for the last log mining query to prepare results for processing.

Duration

The average duration it has taken for log mining queries to prepare results for processing.

int

The number of log mining queries executed.

Duration

The duration it took for the last log mining query results to be processed.

Duration

The average duration it has taken for the log mining query results to be processed.

int

The number of log mining query results processed.

int

The number of entries fetched by the log mining query per database round-trip.

int

The number of milliseconds the connector sleeps before fetching another batch of results from the log mining view.

Schema History Metrics

The MBean is debezium.mysql:type=connector-metrics,context=schema-history,server=*<database.server.name>*.

AttributesTypeDescription

string

One of STOPPED, RECOVERING (recovering history from the storage), RUNNING describing the state of the database history.

long

The time in epoch seconds at what recovery has started.

long

The number of changes that were read during recovery phase.

long

the total number of schema changes applied during recovery and runtime.

long

The number of milliseconds that elapsed since the last change was recovered from the history store.

long

The number of milliseconds that elapsed since the last change was applied.

string

The string representation of the last change recovered from the history store.

string

The string representation of the last applied change.

Connector Properties

The following configuration properties are required unless a default value is available.

Property

Default

Description

Unique name for the connector. Attempting to register again with the same name will fail. (This property is required by all Kafka Connect connectors.)

The name of the Java class for the connector. Always use a value of io.debezium​.connector.oracle.OracleConnector for the Oracle connector.

1

The maximum number of tasks that should be created for this connector. The Oracle connector always uses a single task and therefore does not use this value, so the default is always acceptable.

IP address or hostname of the Oracle database server.

Integer port number of the Oracle database server.

Name of the user to use when connecting to the Oracle database server.

Password to use when connecting to the Oracle database server.

Name of the database to connect to. Must be the CDB name when working with the CDB + PDB model.

Name of the PDB to connect to, when working with the CDB + PDB model.

Name of the XStream outbound server configured in the database.

Logical name that identifies and provides a namespace for the particular Oracle database server being monitored. The logical name should be unique across all other connectors, since it is used as a prefix for all Kafka topic names emanating from this connector. Only alphanumeric characters and underscores should be used.

xstream

The adapter implementation to use. xstream uses the Oracle XStreams API. logminer uses the native Oracle LogMiner API.

The full name of the Kafka topic where the connector will store the database schema history.

A list of host/port pairs that the connector will use for establishing an initial connection to the Kafka cluster. This connection will be used for retrieving database schema history previously stored by the connector, and for writing each DDL statement read from the source database. This should point to the same Kafka cluster used by the Kafka Connect process.

initial

A mode for taking an initial snapshot of the structure and optionally data of captured tables. Supported values are initial (will take a snapshot of structure and data of captured tables; useful if topics should be populated with a complete representation of the data from the captured tables) and schema_only (will take a snapshot of the structure of captured tables only; useful if only changes happening from now onwards should be propagated to topics). Once the snapshot is complete, the connector will continue reading change events from the database’s redo logs.

empty string

An optional comma-separated list of regular expressions that match fully-qualified table identifiers for tables to be monitored; any table not included in the include list will be excluded from monitoring. Each identifier is of the form schemaName.tableName. By default the connector will monitor every non-system table in each monitored database. May not be used with table.exclude.list.

empty string

An optional comma-separated list of regular expressions that match fully-qualified table identifiers for tables to be excluded from monitoring; any table not included in the exclude list will be monitored. Each identifier is of the form schemaName.tableName. May not be used with table.include.list.

n/a

An optional comma-separated list of regular expressions that match the fully-qualified names of character-based columns whose values should be pseudonyms in the change event message values with a field value consisting of the hashed value using the algorithm hashAlgorithm and salt salt. Based on the used hash function referential integrity is kept while data is pseudonymized. Supported hash functions are described in the MessageDigest section of the Java Cryptography Architecture Standard Algorithm Name Documentation. The hash is automatically shortened to the length of the column.

Multiple properties with different lengths can be used in a single configuration, although in each the length must be a positive integer or zero. Fully-qualified names for columns are of the form pdbName.schemaName.tableName.columnName.

Example:

  1. column.mask.hash.SHA-256.with.salt.CzQMA0cB5K = inventory.orders.customerName, inventory.shipment.customerName

where CzQMA0cB5K is a randomly selected salt.

Note: Depending on the hashAlgorithm used, the salt selected and the actual data set, the resulting masked data set may not be completely anonymized.

precise

Specifies how the connector should handle floating point values for NUMBER, DECIMAL and NUMERIC columns: precise (the default) represents them precisely using java.math.BigDecimal values represented in change events in a binary form; or double represents them using double values, which may result in a loss of precision but will be far easier to use. string option encodes values as formatted string which is easy to consume but a semantic information about the real type is lost. See [decimal-values].

fail

Specifies how the connector should react to exceptions during processing of events. fail will propagate the exception (indicating the offset of the problematic event), causing the connector to stop.
warn will cause the problematic event to be skipped and the offset of the problematic event to be logged.
skip will cause the problematic event to be skipped.

8192

Positive integer value that specifies the maximum size of the blocking queue into which change events read from the database log are placed before they are written to Kafka. This queue can provide backpressure to the binlog reader when, for example, writes to Kafka are slower or if Kafka is not available. Events that appear in the queue are not included in the offsets periodically recorded by this connector. Defaults to 8192, and should always be larger than the maximum batch size specified in the max.batch.size property.

2048

Positive integer value that specifies the maximum size of each batch of events that should be processed during each iteration of this connector. Defaults to 2048.

1000

Positive integer value that specifies the number of milliseconds the connector should wait during each iteration for new change events to appear. Defaults to 1000 milliseconds, or 1 second.

true

Controls whether a tombstone event should be generated after a delete event.
When true the delete operations are represented by a delete event and a subsequent tombstone event. When false only a delete event is sent.
Emitting the tombstone event (the default behavior) allows Kafka to completely delete all events pertaining to the given key once the source record got deleted.

empty string

A semi-colon list of regular expressions that match fully-qualified tables and columns to map a primary key.
Each item (regular expression) must match the <fully-qualified table>:<a comma-separated list of columns> representing the custom key.
Fully-qualified tables could be defined as pdbName.schemaName.tableName.

n/a

An optional comma-separated list of regular expressions that match the fully-qualified names of character-based columns whose values should be truncated in the change event message values if the field values are longer than the specified number of characters. Multiple properties with different lengths can be used in a single configuration, although in each the length must be a positive integer. Fully-qualified names for columns are of the form pdbName.schemaName.tableName.columnName.

n/a

An optional comma-separated list of regular expressions that match the fully-qualified names of character-based columns whose values should be replaced in the change event message values with a field value consisting of the specified number of asterisk (*) characters. Multiple properties with different lengths can be used in a single configuration, although in each the length must be a positive integer or zero. Fully-qualified names for columns are of the form pdbName.schemaName.tableName.columnName.

n/a

An optional comma-separated list of regular expressions that match the fully-qualified names of columns whose original type and length should be added as a parameter to the corresponding field schemas in the emitted change messages. The schema parameters debezium.source.column.type, debezium.source.column.length and debezium.source.column.scale will be used to propagate the original type name and length (for variable-width types), respectively. Useful to properly size corresponding columns in sink databases. Fully-qualified names for columns are of the form tableName.columnName, or schemaName.tableName.columnName.

n/a

An optional comma-separated list of regular expressions that match the database-specific data type name of columns whose original type and length should be added as a parameter to the corresponding field schemas in the emitted change messages. The schema parameters debezium.source.column.type, debezium.source.column.length and debezium.source.column.scale will be used to propagate the original type name and length (for variable-width types), respectively. Useful to properly size corresponding columns in sink databases. Fully-qualified data type names are of the form tableName.typeName, or schemaName.tableName.typeName. See the list of Oracle-specific data type names.

0

Controls how frequently heartbeat messages are sent.
This property contains an interval in milli-seconds that defines how frequently the connector sends messages into a heartbeat topic. This can be used to monitor whether the connector is still receiving change events from the database. You also should leverage heartbeat messages in cases where only records in non-captured tables are changed for a longer period of time. In such situation the connector would proceed to read the log from the database but never emit any change messages into Kafka, which in turn means that no offset updates will be committed to Kafka. This will cause the redo log files to be retained by the database longer than needed (as the connector actually has processed them already but never got a chance to flush the latest retrieved SCN to the database) and also may result in more change events to be re-sent after a connector restart. Set this parameter to 0 to not send heartbeat messages at all.
Disabled by default.

__debezium-heartbeat

Controls the naming of the topic to which heartbeat messages are sent.
The topic is named according to the pattern <heartbeat.topics.prefix>.<server.name>.

An interval in milli-seconds that the connector should wait before taking a snapshot after starting up;
Can be used to avoid snapshot interruptions when starting multiple connectors in a cluster, which may cause re-balancing of connectors.

2000

Specifies the maximum number of rows that should be read in one go from each table while taking a snapshot. The connector will read the table contents in multiple batches of this size. Defaults to 2000.

true when connector configuration explicitly specifies the key.converter or value.converter parameters to use Avro, otherwise defaults to false.

Whether field names will be sanitized to adhere to Avro naming requirements. See Avro naming for more details.

false

When set to true Debezium generates events with transaction boundaries and enriches data events envelope with transaction metadata.

See Transaction Metadata for additional details.