Package io.debezium.pipeline.source.snapshot.incremental

Class AbstractIncrementalSnapshotChangeEventSource<P extends Partition,T extends DataCollectionId>

java.lang.Object

io.debezium.pipeline.source.snapshot.incremental.AbstractIncrementalSnapshotChangeEventSource<P,T>

All Implemented Interfaces:

IncrementalSnapshotChangeEventSource<P,T>

Direct Known Subclasses:

SignalBasedIncrementalSnapshotChangeEventSource

@NotThreadSafe
public abstract class AbstractIncrementalSnapshotChangeEventSource<P extends Partition,T extends DataCollectionId>
extends Object
implements IncrementalSnapshotChangeEventSource<P,T>

An incremental snapshot change event source that emits events from a DB log interleaved with snapshot events.

Field Summary

Fields

Modifier and Type	Field	Description
private final Clock	clock
private final RelationalDatabaseConnectorConfig	connectorConfig
protected IncrementalSnapshotContext<T>	context
private Table	currentTable
private final RelationalDatabaseSchema	databaseSchema
private final DataChangeEventListener<P>	dataListener
protected EventDispatcher<P,T>	dispatcher
protected JdbcConnection	jdbcConnection
private static final org.slf4j.Logger	LOGGER
private final SnapshotProgressListener<P>	progressListener
private long	totalRowsScanned
protected final Map<org.apache.kafka.connect.data.Struct,Object[]>	window

Constructor Summary

Constructors

Constructor	Description
AbstractIncrementalSnapshotChangeEventSource(RelationalDatabaseConnectorConfig config, JdbcConnection jdbcConnection, EventDispatcher<P,T> dispatcher, DatabaseSchema<?> databaseSchema, Clock clock, SnapshotProgressListener<P> progressListener, DataChangeEventListener<P> dataChangeEventListener)

Method Summary

Modifier and Type	Method	Description
void	addDataCollectionNamesToSnapshot(P partition, List<String> dataCollectionIds, OffsetContext offsetContext)
protected void	addKeyColumnsToCondition(Table table, StringBuilder sql, String predicate)
private void	addLowerBound(Table table, StringBuilder sql)
protected String	buildChunkQuery(Table table)
protected String	buildChunkQuery(Table table, int limit)
protected String	buildMaxPrimaryKeyQuery(Table table)
private boolean	checkSchemaChanges(ResultSet rs)
void	closeWindow(P partition, String id, OffsetContext offsetContext)
private boolean	createDataEventsForTable(P partition)	Dispatches the data change events for the records of a single table.
protected void	deduplicateWindow(DataCollectionId dataCollectionId, Object key)
protected abstract void	emitWindowClose(P partition)	Update high watermark for the incremental snapshot chunk
protected abstract void	emitWindowOpen()	Update low watermark for the incremental snapshot chunk
protected ChangeRecordEmitter<P>	getChangeRecordEmitter(P partition, T dataCollectionId, OffsetContext offsetContext, Object[] row)	Returns a ChangeRecordEmitter producing the change records for the given table row.
private Key.KeyMapper	getKeyMapper()
protected String	getSignalTableName(String dataCollectionId)
private Table	getTable(ResultSet rs)
private Threads.Timer	getTableScanLogTimer()
private void	incrementTableRowsScanned(P partition, long rows)
void	init(P partition, OffsetContext offsetContext)
private boolean	isTableInvalid(P partition)
private Object[]	keyFromRow(Object[] row)
private void	nextDataCollection(P partition)
protected void	postIncrementalSnapshotCompleted()
protected void	postReadChunk(IncrementalSnapshotContext<T> context)
protected void	preReadChunk(IncrementalSnapshotContext<T> context)
void	processSchemaChange(P partition, DataCollectionId dataCollectionId)
protected void	readChunk(P partition)
private Table	readSchema()
protected PreparedStatement	readTableChunkStatement(String sql)
protected Table	refreshTableSchema(Table table)
void	rereadChunk(P partition)
private boolean	schemaHistoryIsUpToDate()	Verifies that in-memory representation of the table’s schema is up to date with the table's schema in the database.
protected void	sendEvent(P partition, EventDispatcher<P,T> dispatcher, OffsetContext offsetContext, Object[] row)
protected void	sendWindowEvents(P partition, OffsetContext offsetContext)
protected void	setContext(IncrementalSnapshotContext<T> context)
private void	tableScanCompleted(P partition)
private void	verifySchemaUnchanged()	Verifies that table's schema in the database has not changed since it was captured in the previous window

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Methods inherited from interface io.debezium.pipeline.source.snapshot.incremental.IncrementalSnapshotChangeEventSource

processFilteredEvent, processHeartbeat, processMessage, processTransactionCommittedEvent, processTransactionStartedEvent

Field Details

LOGGER

private static final org.slf4j.Logger LOGGER

connectorConfig

private final RelationalDatabaseConnectorConfig connectorConfig

clock

private final Clock clock

databaseSchema

private final RelationalDatabaseSchema databaseSchema

progressListener

private final SnapshotProgressListener<P extends Partition> progressListener

dataListener

private final DataChangeEventListener<P extends Partition> dataListener

totalRowsScanned

private long totalRowsScanned

currentTable

private Table currentTable

dispatcher

protected EventDispatcher<P extends Partition,T extends DataCollectionId> dispatcher

context

protected IncrementalSnapshotContext<T extends DataCollectionId> context

jdbcConnection

protected JdbcConnection jdbcConnection

window

protected final Map<org.apache.kafka.connect.data.Struct,Object[]> window

Constructor Details

AbstractIncrementalSnapshotChangeEventSource

public AbstractIncrementalSnapshotChangeEventSource(RelationalDatabaseConnectorConfig config,
 JdbcConnection jdbcConnection,
 EventDispatcher<P,T> dispatcher,
 DatabaseSchema<?> databaseSchema,
 Clock clock,
 SnapshotProgressListener<P> progressListener,
 DataChangeEventListener<P> dataChangeEventListener)

Method Details

closeWindow

public void closeWindow(P partition,
 String id,
 OffsetContext offsetContext)
                 throws InterruptedException

Specified by:

closeWindow in interface IncrementalSnapshotChangeEventSource<P extends Partition,T extends DataCollectionId>

Throws:

InterruptedException

processSchemaChange

public void processSchemaChange(P partition,
 DataCollectionId dataCollectionId)
                         throws InterruptedException

Specified by:

processSchemaChange in interface IncrementalSnapshotChangeEventSource<P extends Partition,T extends DataCollectionId>

Throws:

InterruptedException

rereadChunk

public void rereadChunk(P partition)
                 throws InterruptedException

Throws:

InterruptedException

getSignalTableName

protected String getSignalTableName(String dataCollectionId)

sendWindowEvents

protected void sendWindowEvents(P partition,
 OffsetContext offsetContext)
                         throws InterruptedException

Throws:

InterruptedException

sendEvent

protected void sendEvent(P partition,
 EventDispatcher<P,T> dispatcher,
 OffsetContext offsetContext,
 Object[] row)
                  throws InterruptedException

Throws:

InterruptedException

getChangeRecordEmitter

protected ChangeRecordEmitter<P> getChangeRecordEmitter(P partition,
 T dataCollectionId,
 OffsetContext offsetContext,
 Object[] row)

Returns a ChangeRecordEmitter producing the change records for the given table row.

deduplicateWindow

protected void deduplicateWindow(DataCollectionId dataCollectionId,
 Object key)

emitWindowOpen

protected abstract void emitWindowOpen()
                                throws SQLException

Update low watermark for the incremental snapshot chunk

Throws:

SQLException

emitWindowClose

protected abstract void emitWindowClose(P partition)
                                 throws SQLException,
InterruptedException

Update high watermark for the incremental snapshot chunk

Throws:

SQLException

InterruptedException

buildChunkQuery

protected String buildChunkQuery(Table table)

buildChunkQuery

protected String buildChunkQuery(Table table,
 int limit)

addLowerBound

private void addLowerBound(Table table,
 StringBuilder sql)

buildMaxPrimaryKeyQuery

protected String buildMaxPrimaryKeyQuery(Table table)

init

public void init(P partition,
 OffsetContext offsetContext)

Specified by:

init in interface IncrementalSnapshotChangeEventSource<P extends Partition,T extends DataCollectionId>

readChunk

protected void readChunk(P partition)
                  throws InterruptedException

Throws:

InterruptedException

isTableInvalid

private boolean isTableInvalid(P partition)

schemaHistoryIsUpToDate

private boolean schemaHistoryIsUpToDate()

Verifies that in-memory representation of the table’s schema is up to date with the table's schema in the database.

Verification is a two step process:

1. Save table's schema from the database to the context
2. Verify schema hasn't changed in the following window. If schema has changed repeat the process

Two step process allows to wait for the connector to receive the DDL event in the binlog stream and update the in-memory representation of the table’s schema.

Verification is done at the beginning of the incremental snapshot and on every schema change during the snapshotting.

verifySchemaUnchanged

private void verifySchemaUnchanged()

Verifies that table's schema in the database has not changed since it was captured in the previous window

readSchema

private Table readSchema()

nextDataCollection

private void nextDataCollection(P partition)

addDataCollectionNamesToSnapshot

public void addDataCollectionNamesToSnapshot(P partition,
 List<String> dataCollectionIds,
 OffsetContext offsetContext)
                                      throws InterruptedException

Specified by:

addDataCollectionNamesToSnapshot in interface IncrementalSnapshotChangeEventSource<P extends Partition,T extends DataCollectionId>

Throws:

InterruptedException

addKeyColumnsToCondition

protected void addKeyColumnsToCondition(Table table,
 StringBuilder sql,
 String predicate)

createDataEventsForTable

private boolean createDataEventsForTable(P partition)

Dispatches the data change events for the records of a single table.

checkSchemaChanges

private boolean checkSchemaChanges(ResultSet rs)
                            throws SQLException

Throws:

SQLException

getTable

private Table getTable(ResultSet rs)
                throws SQLException

Throws:

SQLException

incrementTableRowsScanned

private void incrementTableRowsScanned(P partition,
 long rows)

tableScanCompleted

private void tableScanCompleted(P partition)

readTableChunkStatement

protected PreparedStatement readTableChunkStatement(String sql)
                                             throws SQLException

Throws:

SQLException

getTableScanLogTimer

private Threads.Timer getTableScanLogTimer()

keyFromRow

private Object[] keyFromRow(Object[] row)

setContext

protected void setContext(IncrementalSnapshotContext<T> context)

preReadChunk

protected void preReadChunk(IncrementalSnapshotContext<T> context)

postReadChunk

protected void postReadChunk(IncrementalSnapshotContext<T> context)

postIncrementalSnapshotCompleted

protected void postIncrementalSnapshotCompleted()

refreshTableSchema

protected Table refreshTableSchema(Table table)
                            throws SQLException

Throws:

SQLException

getKeyMapper

private Key.KeyMapper getKeyMapper()