org.apache.flink.runtime.scheduler.adaptivebatch

类 DefaultVertexParallelismDecider

java.lang.Object

org.apache.flink.runtime.scheduler.adaptivebatch.DefaultVertexParallelismDecider

所有已实现的接口:

VertexParallelismDecider

public class DefaultVertexParallelismDecider
extends Object
implements VertexParallelismDecider

Default implementation of VertexParallelismDecider. Currently, in order to make the number of subpartitions evenly consumed by downstream tasks, we will normalize the decided parallelism to a power of 2.

方法概要

限定符和类型	方法和说明
int	decideParallelismForVertex(JobVertexID jobVertexId, List<BlockingResultInfo> consumedResults, int vertexMaxParallelism) Computing the parallelism.

从类继承的方法 java.lang.Object

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

方法详细资料

decideParallelismForVertex

public int decideParallelismForVertex(JobVertexID jobVertexId,
                                      List<BlockingResultInfo> consumedResults,
                                      int vertexMaxParallelism)

从接口复制的说明: VertexParallelismDecider

Computing the parallelism.

指定者:

decideParallelismForVertex 在接口中 VertexParallelismDecider

参数:

jobVertexId - The job vertex id.

consumedResults - The information of consumed blocking results.

vertexMaxParallelism - The max parallelism of the job vertex.

返回:

the parallelism of the job vertex.