org.apache.flink.runtime.io.network.partition

Class PipelinedApproximateSubpartitionView

java.lang.Object

org.apache.flink.runtime.io.network.partition.PipelinedSubpartitionView

org.apache.flink.runtime.io.network.partition.PipelinedApproximateSubpartitionView

All Implemented Interfaces:

ResultSubpartitionView

public class PipelinedApproximateSubpartitionView
extends PipelinedSubpartitionView

View over a pipelined in-memory only subpartition allowing reconnecting.

Nested Class Summary

Nested classes/interfaces inherited from interface org.apache.flink.runtime.io.network.partition.ResultSubpartitionView

ResultSubpartitionView.AvailabilityWithBacklog

Method Summary

Modifier and Type	Method and Description
void	releaseAllResources() Pipelined ResultPartition relies on its subpartition view's release to decide whether the partition is ready to release.

Methods inherited from class org.apache.flink.runtime.io.network.partition.PipelinedSubpartitionView

acknowledgeAllDataProcessed, getAvailabilityAndBacklog, getFailureCause, getNextBuffer, getNumberOfQueuedBuffers, isReleased, notifyDataAvailable, notifyNewBufferSize, notifyPriorityEvent, resumeConsumption, toString, unsynchronizedGetNumberOfQueuedBuffers

Methods inherited from class java.lang.Object

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

Method Detail

releaseAllResources

public void releaseAllResources()

Pipelined ResultPartition relies on its subpartition view's release to decide whether the partition is ready to release. In contrast, Approximate Pipelined ResultPartition is put into the JobMaster's Partition Tracker and relies on the tracker to release partitions after the job is finished. Hence in the approximate pipelined case, no resource related to view is needed to be released.

Specified by:

releaseAllResources in interface ResultSubpartitionView

Overrides:

releaseAllResources in class PipelinedSubpartitionView