org.apache.flink.connector.kafka.source.enumerator

类 KafkaSourceEnumerator

java.lang.Object

org.apache.flink.connector.kafka.source.enumerator.KafkaSourceEnumerator

所有已实现的接口:

AutoCloseable, org.apache.flink.api.common.state.CheckpointListener, org.apache.flink.api.connector.source.SplitEnumerator<KafkaPartitionSplit,KafkaSourceEnumState>

@Internal
public class KafkaSourceEnumerator
extends Object
implements org.apache.flink.api.connector.source.SplitEnumerator<KafkaPartitionSplit,KafkaSourceEnumState>

The enumerator class for Kafka source.

嵌套类概要

嵌套类

限定符和类型	类和说明
static class	KafkaSourceEnumerator.PartitionOffsetsRetrieverImpl The implementation for offsets retriever with a consumer and an admin client.

构造器概要

构造器

构造器和说明
KafkaSourceEnumerator(KafkaSubscriber subscriber, OffsetsInitializer startingOffsetInitializer, OffsetsInitializer stoppingOffsetInitializer, Properties properties, org.apache.flink.api.connector.source.SplitEnumeratorContext<KafkaPartitionSplit> context, org.apache.flink.api.connector.source.Boundedness boundedness)
KafkaSourceEnumerator(KafkaSubscriber subscriber, OffsetsInitializer startingOffsetInitializer, OffsetsInitializer stoppingOffsetInitializer, Properties properties, org.apache.flink.api.connector.source.SplitEnumeratorContext<KafkaPartitionSplit> context, org.apache.flink.api.connector.source.Boundedness boundedness, Set<org.apache.kafka.common.TopicPartition> assignedPartitions)

方法概要

限定符和类型	方法和说明
void	addReader(int subtaskId)
void	addSplitsBack(List<KafkaPartitionSplit> splits, int subtaskId)
void	close()
void	handleSplitRequest(int subtaskId, String requesterHostname)
KafkaSourceEnumState	snapshotState(long checkpointId)
void	start() Start the enumerator.

从类继承的方法 java.lang.Object

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

从接口继承的方法 org.apache.flink.api.connector.source.SplitEnumerator

handleSourceEvent, notifyCheckpointComplete

从接口继承的方法 org.apache.flink.api.common.state.CheckpointListener

notifyCheckpointAborted

构造器详细资料

KafkaSourceEnumerator

public KafkaSourceEnumerator(KafkaSubscriber subscriber,
                             OffsetsInitializer startingOffsetInitializer,
                             OffsetsInitializer stoppingOffsetInitializer,
                             Properties properties,
                             org.apache.flink.api.connector.source.SplitEnumeratorContext<KafkaPartitionSplit> context,
                             org.apache.flink.api.connector.source.Boundedness boundedness)

KafkaSourceEnumerator

public KafkaSourceEnumerator(KafkaSubscriber subscriber,
                             OffsetsInitializer startingOffsetInitializer,
                             OffsetsInitializer stoppingOffsetInitializer,
                             Properties properties,
                             org.apache.flink.api.connector.source.SplitEnumeratorContext<KafkaPartitionSplit> context,
                             org.apache.flink.api.connector.source.Boundedness boundedness,
                             Set<org.apache.kafka.common.TopicPartition> assignedPartitions)

方法详细资料

start

public void start()

Start the enumerator.

Depending on partitionDiscoveryIntervalMs, the enumerator will trigger a one-time partition discovery, or schedule a callable for discover partitions periodically.

The invoking chain of partition discovery would be:

1. getSubscribedTopicPartitions() in worker thread
2. checkPartitionChanges(java.util.Set<org.apache.kafka.common.TopicPartition>, java.lang.Throwable) in coordinator thread
3. initializePartitionSplits(org.apache.flink.connector.kafka.source.enumerator.KafkaSourceEnumerator.PartitionChange) in worker thread
4. handlePartitionSplitChanges(org.apache.flink.connector.kafka.source.enumerator.KafkaSourceEnumerator.PartitionSplitChange, java.lang.Throwable) in coordinator thread

指定者:

start 在接口中 org.apache.flink.api.connector.source.SplitEnumerator<KafkaPartitionSplit,KafkaSourceEnumState>

handleSplitRequest

public void handleSplitRequest(int subtaskId,
                               @Nullable
                               String requesterHostname)

指定者:

handleSplitRequest 在接口中 org.apache.flink.api.connector.source.SplitEnumerator<KafkaPartitionSplit,KafkaSourceEnumState>

addSplitsBack

public void addSplitsBack(List<KafkaPartitionSplit> splits,
                          int subtaskId)

指定者:

addSplitsBack 在接口中 org.apache.flink.api.connector.source.SplitEnumerator<KafkaPartitionSplit,KafkaSourceEnumState>

addReader

public void addReader(int subtaskId)

指定者:

addReader 在接口中 org.apache.flink.api.connector.source.SplitEnumerator<KafkaPartitionSplit,KafkaSourceEnumState>

snapshotState

public KafkaSourceEnumState snapshotState(long checkpointId)
                                   throws Exception

指定者:

snapshotState 在接口中 org.apache.flink.api.connector.source.SplitEnumerator<KafkaPartitionSplit,KafkaSourceEnumState>

抛出:

Exception

close

public void close()

指定者:

close 在接口中 AutoCloseable

指定者:

close 在接口中 org.apache.flink.api.connector.source.SplitEnumerator<KafkaPartitionSplit,KafkaSourceEnumState>