org.apache.flink.streaming.connectors.kafka.internal

类 KafkaConsumerThread

java.lang.Object

java.lang.Thread

org.apache.flink.streaming.connectors.kafka.internal.KafkaConsumerThread

所有已实现的接口:

Runnable

@Internal
public class KafkaConsumerThread
extends Thread

The thread the runs the KafkaConsumer, connecting to the brokers and polling records. The thread pushes the data into a Handover to be picked up by the fetcher that will deserialize and emit the records.

IMPORTANT: This thread must not be interrupted when attempting to shut it down. The Kafka consumer code was found to not always handle interrupts well, and to even deadlock in certain situations.

Implementation Note: This code is written to be reusable in later versions of the KafkaConsumer. Because Kafka is not maintaining binary compatibility, we use a "call bridge" as an indirection to the KafkaConsumer calls that change signature.

嵌套类概要

从类继承的嵌套类/接口 java.lang.Thread

Thread.State, Thread.UncaughtExceptionHandler

字段概要

从类继承的字段 java.lang.Thread

MAX_PRIORITY, MIN_PRIORITY, NORM_PRIORITY

构造器概要

构造器

构造器和说明

KafkaConsumerThread(org.slf4j.Logger log, Handover handover, Properties kafkaProperties, org.apache.flink.streaming.connectors.kafka.internals.ClosableBlockingQueue<org.apache.flink.streaming.connectors.kafka.internals.KafkaTopicPartitionState<org.apache.kafka.common.TopicPartition>> unassignedPartitionsQueue, KafkaConsumerCallBridge09 consumerCallBridge, String threadName, long pollTimeout, boolean useMetrics, org.apache.flink.metrics.MetricGroup consumerMetricGroup, org.apache.flink.metrics.MetricGroup subtaskMetricGroup, org.apache.flink.api.common.io.ratelimiting.FlinkConnectorRateLimiter rateLimiter)

方法概要

限定符和类型	方法和说明
protected org.apache.kafka.clients.consumer.ConsumerRecords<byte[],byte[]>	getRecordsFromKafka() Get records from Kafka.
void	run()
void	shutdown() Shuts this thread down, waking up the thread gracefully if blocked (without Thread.interrupt() calls).

从类继承的方法 java.lang.Thread

activeCount, checkAccess, clone, countStackFrames, currentThread, destroy, dumpStack, enumerate, getAllStackTraces, getContextClassLoader, getDefaultUncaughtExceptionHandler, getId, getName, getPriority, getStackTrace, getState, getThreadGroup, getUncaughtExceptionHandler, holdsLock, interrupt, interrupted, isAlive, isDaemon, isInterrupted, join, join, join, resume, setContextClassLoader, setDaemon, setDefaultUncaughtExceptionHandler, setName, setPriority, setUncaughtExceptionHandler, sleep, sleep, start, stop, stop, suspend, toString, yield

从类继承的方法 java.lang.Object

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

构造器详细资料

KafkaConsumerThread

public KafkaConsumerThread(org.slf4j.Logger log,
                           Handover handover,
                           Properties kafkaProperties,
                           org.apache.flink.streaming.connectors.kafka.internals.ClosableBlockingQueue<org.apache.flink.streaming.connectors.kafka.internals.KafkaTopicPartitionState<org.apache.kafka.common.TopicPartition>> unassignedPartitionsQueue,
                           KafkaConsumerCallBridge09 consumerCallBridge,
                           String threadName,
                           long pollTimeout,
                           boolean useMetrics,
                           org.apache.flink.metrics.MetricGroup consumerMetricGroup,
                           org.apache.flink.metrics.MetricGroup subtaskMetricGroup,
                           org.apache.flink.api.common.io.ratelimiting.FlinkConnectorRateLimiter rateLimiter)

方法详细资料

run

public void run()

指定者:

run 在接口中 Runnable

覆盖:

run 在类中 Thread

shutdown

public void shutdown()

Shuts this thread down, waking up the thread gracefully if blocked (without Thread.interrupt() calls).

getRecordsFromKafka

@VisibleForTesting
protected org.apache.kafka.clients.consumer.ConsumerRecords<byte[],byte[]> getRecordsFromKafka()

Get records from Kafka. If the rate-limiting feature is turned on, this method is called at a rate specified by the rateLimiter.

返回:

ConsumerRecords