Package org.apache.kafka.clients.consumer.internals

Class OffsetFetcher

java.lang.Object

org.apache.kafka.clients.consumer.internals.OffsetFetcher

public class OffsetFetcher
extends Object

OffsetFetcher is responsible for fetching the offsets for a given set of topic and partition pairs and for validation and resetting of positions, as needed.

Constructor Summary

Constructors

Constructor	Description
OffsetFetcher(org.apache.kafka.common.utils.LogContext logContext, ConsumerNetworkClient client, ConsumerMetadata metadata, SubscriptionState subscriptions, org.apache.kafka.common.utils.Time time, long retryBackoffMs, long requestTimeoutMs, org.apache.kafka.common.IsolationLevel isolationLevel, org.apache.kafka.clients.ApiVersions apiVersions)

Method Summary

Modifier and Type	Method	Description
Map<org.apache.kafka.common.TopicPartition,Long>	beginningOffsets(Collection<org.apache.kafka.common.TopicPartition> partitions, org.apache.kafka.common.utils.Timer timer)
Map<org.apache.kafka.common.TopicPartition,Long>	endOffsets(Collection<org.apache.kafka.common.TopicPartition> partitions, org.apache.kafka.common.utils.Timer timer)
Map<org.apache.kafka.common.TopicPartition,OffsetAndTimestamp>	offsetsForTimes(Map<org.apache.kafka.common.TopicPartition,Long> timestampsToSearch, org.apache.kafka.common.utils.Timer timer)
void	resetPositionsIfNeeded()	Reset offsets for all assigned partitions that require it.
void	validatePositionsIfNeeded()	Validate offsets for all assigned partitions for which a leader change has been detected.
void	validatePositionsOnMetadataChange()	If we have seen new metadata (as tracked by Metadata.updateVersion()), then we should check that all the assignments have a valid position.

Methods inherited from class java.lang.Object

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

Constructor Details

OffsetFetcher

public OffsetFetcher(org.apache.kafka.common.utils.LogContext logContext,
 ConsumerNetworkClient client,
 ConsumerMetadata metadata,
 SubscriptionState subscriptions,
 org.apache.kafka.common.utils.Time time,
 long retryBackoffMs,
 long requestTimeoutMs,
 org.apache.kafka.common.IsolationLevel isolationLevel,
 org.apache.kafka.clients.ApiVersions apiVersions)

Method Details

resetPositionsIfNeeded

public void resetPositionsIfNeeded()

Reset offsets for all assigned partitions that require it.

Throws:

NoOffsetForPartitionException - If no offset reset strategy is defined and one or more partitions aren't awaiting a seekToBeginning() or seekToEnd().

validatePositionsIfNeeded

public void validatePositionsIfNeeded()

Validate offsets for all assigned partitions for which a leader change has been detected.

offsetsForTimes

public Map<org.apache.kafka.common.TopicPartition,OffsetAndTimestamp> offsetsForTimes(Map<org.apache.kafka.common.TopicPartition,Long> timestampsToSearch,
 org.apache.kafka.common.utils.Timer timer)

beginningOffsets

public Map<org.apache.kafka.common.TopicPartition,Long> beginningOffsets(Collection<org.apache.kafka.common.TopicPartition> partitions,
 org.apache.kafka.common.utils.Timer timer)

endOffsets

public Map<org.apache.kafka.common.TopicPartition,Long> endOffsets(Collection<org.apache.kafka.common.TopicPartition> partitions,
 org.apache.kafka.common.utils.Timer timer)

validatePositionsOnMetadataChange

public void validatePositionsOnMetadataChange()

If we have seen new metadata (as tracked by Metadata.updateVersion()), then we should check that all the assignments have a valid position.