Package org.apache.flink.runtime.scheduler.adapter

Class DefaultExecutionTopology

java.lang.Object

org.apache.flink.runtime.scheduler.adapter.DefaultExecutionTopology

All Implemented Interfaces:

SchedulingTopology, BaseTopology<ExecutionVertexID,IntermediateResultPartitionID,SchedulingExecutionVertex,SchedulingResultPartition>, Topology<ExecutionVertexID,IntermediateResultPartitionID,SchedulingExecutionVertex,SchedulingResultPartition,SchedulingPipelinedRegion>

public class DefaultExecutionTopology
extends Object
implements SchedulingTopology

Adapter of ExecutionGraph to SchedulingTopology.

Method Summary

Modifier and Type	Method	Description
static Map<JobVertexID,DefaultLogicalPipelinedRegion>	computeLogicalPipelinedRegionsByJobVertexId(List<JobVertex> topologicallySortedJobVertices)
static DefaultExecutionTopology	fromExecutionGraph(DefaultExecutionGraph executionGraph)
Iterable<DefaultSchedulingPipelinedRegion>	getAllPipelinedRegions()	Returns all pipelined regions in this topology.
EdgeManager	getEdgeManager()
DefaultSchedulingPipelinedRegion	getPipelinedRegionOfVertex(ExecutionVertexID vertexId)	The pipelined region for a specified vertex.
org.apache.flink.runtime.scheduler.adapter.DefaultResultPartition	getResultPartition(IntermediateResultPartitionID intermediateResultPartitionId)	Looks up the SchedulingResultPartition for the given IntermediateResultPartitionID.
org.apache.flink.runtime.scheduler.adapter.DefaultExecutionVertex	getVertex(ExecutionVertexID executionVertexId)	Looks up the SchedulingExecutionVertex for the given ExecutionVertexID.
Iterable<org.apache.flink.runtime.scheduler.adapter.DefaultExecutionVertex>	getVertices()	Returns an iterable over all vertices, topologically sorted.
void	notifyExecutionGraphUpdatedWithInitializedJobVertices(DefaultExecutionGraph executionGraph, List<ExecutionJobVertex> newlyInitializedJobVertices)
void	notifyExecutionGraphUpdatedWithNewJobVertices(List<JobVertex> topologicallySortedJobVertices)
void	registerSchedulingTopologyListener(SchedulingTopologyListener listener)	Register a scheduling topology listener.

Methods inherited from class java.lang.Object

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

Method Details

getVertices

public Iterable<org.apache.flink.runtime.scheduler.adapter.DefaultExecutionVertex> getVertices()

Description copied from interface: BaseTopology

Returns an iterable over all vertices, topologically sorted.

Specified by:

getVertices in interface BaseTopology<ExecutionVertexID,IntermediateResultPartitionID,SchedulingExecutionVertex,SchedulingResultPartition>

Returns:

topologically sorted iterable over all vertices

getVertex

public org.apache.flink.runtime.scheduler.adapter.DefaultExecutionVertex getVertex(ExecutionVertexID executionVertexId)

Description copied from interface: SchedulingTopology

Looks up the SchedulingExecutionVertex for the given ExecutionVertexID.

Specified by:

getVertex in interface SchedulingTopology

Parameters:

executionVertexId - identifying the respective scheduling vertex

Returns:

The respective scheduling vertex

getResultPartition

public org.apache.flink.runtime.scheduler.adapter.DefaultResultPartition getResultPartition(IntermediateResultPartitionID intermediateResultPartitionId)

Description copied from interface: SchedulingTopology

Looks up the SchedulingResultPartition for the given IntermediateResultPartitionID.

Specified by:

getResultPartition in interface SchedulingTopology

Parameters:

intermediateResultPartitionId - identifying the respective scheduling result partition

Returns:

The respective scheduling result partition

registerSchedulingTopologyListener

public void registerSchedulingTopologyListener(SchedulingTopologyListener listener)

Description copied from interface: SchedulingTopology

Register a scheduling topology listener. The listener will be notified by SchedulingTopologyListener.notifySchedulingTopologyUpdated(SchedulingTopology, List) when the scheduling topology is updated.

Specified by:

registerSchedulingTopologyListener in interface SchedulingTopology

Parameters:

listener - the registered listener.

getAllPipelinedRegions

public Iterable<DefaultSchedulingPipelinedRegion> getAllPipelinedRegions()

Description copied from interface: Topology

Returns all pipelined regions in this topology.

Specified by:

getAllPipelinedRegions in interface Topology<ExecutionVertexID,IntermediateResultPartitionID,SchedulingExecutionVertex,SchedulingResultPartition,SchedulingPipelinedRegion>

Returns:

Iterable over pipelined regions in this topology

getPipelinedRegionOfVertex

public DefaultSchedulingPipelinedRegion getPipelinedRegionOfVertex(ExecutionVertexID vertexId)

Description copied from interface: Topology

The pipelined region for a specified vertex.

Specified by:

getPipelinedRegionOfVertex in interface Topology<ExecutionVertexID,IntermediateResultPartitionID,SchedulingExecutionVertex,SchedulingResultPartition,SchedulingPipelinedRegion>

Parameters:

vertexId - the vertex id identifying the vertex for which the pipelined region should be returned

Returns:

the pipelined region of the vertex

getEdgeManager

public EdgeManager getEdgeManager()

computeLogicalPipelinedRegionsByJobVertexId

public static Map<JobVertexID,DefaultLogicalPipelinedRegion> computeLogicalPipelinedRegionsByJobVertexId(List<JobVertex> topologicallySortedJobVertices)

notifyExecutionGraphUpdatedWithNewJobVertices

public void notifyExecutionGraphUpdatedWithNewJobVertices(List<JobVertex> topologicallySortedJobVertices)

notifyExecutionGraphUpdatedWithInitializedJobVertices

public void notifyExecutionGraphUpdatedWithInitializedJobVertices(DefaultExecutionGraph executionGraph,
 List<ExecutionJobVertex> newlyInitializedJobVertices)

fromExecutionGraph

public static DefaultExecutionTopology fromExecutionGraph(DefaultExecutionGraph executionGraph)