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

类 JobMasterPartitionTrackerImpl

java.lang.Object

org.apache.flink.runtime.io.network.partition.AbstractPartitionTracker<ResourceID,ResultPartitionDeploymentDescriptor>

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

所有已实现的接口:

JobMasterPartitionTracker, PartitionTracker<ResourceID,ResultPartitionDeploymentDescriptor>

public class JobMasterPartitionTrackerImpl
extends AbstractPartitionTracker<ResourceID,ResultPartitionDeploymentDescriptor>
implements JobMasterPartitionTracker

Utility for tracking partitions and issuing release calls to task executors and shuffle masters.

字段概要

从类继承的字段 org.apache.flink.runtime.io.network.partition.AbstractPartitionTracker

partitionInfos, partitionTable

构造器概要

构造器

构造器和说明
JobMasterPartitionTrackerImpl(org.apache.flink.api.common.JobID jobId, ShuffleMaster<?> shuffleMaster, PartitionTrackerFactory.TaskExecutorGatewayLookup taskExecutorGatewayLookup)

方法概要

限定符和类型	方法和说明
void	connectToResourceManager(ResourceManagerGateway resourceManagerGateway)
Collection<ResultPartitionDeploymentDescriptor>	getAllTrackedPartitions() Gets all the partitions under tracking.
List<ShuffleDescriptor>	getClusterPartitionShuffleDescriptors(IntermediateDataSetID intermediateDataSetID) Get the shuffle descriptors of the cluster partitions ordered by partition number.
void	startTrackingPartition(ResourceID producingTaskExecutorId, ResultPartitionDeploymentDescriptor resultPartitionDeploymentDescriptor) Starts the tracking of the given partition for the given task executor ID.
CompletableFuture<Void>	stopTrackingAndPromotePartitions(Collection<ResultPartitionID> resultPartitionIds) Promotes the given partitions, and stops the tracking of partitions that were promoted.
void	stopTrackingAndReleasePartitions(Collection<ResultPartitionID> resultPartitionIds, boolean releaseOnShuffleMaster) Releases the given partitions and stop the tracking of partitions that were released.

从类继承的方法 org.apache.flink.runtime.io.network.partition.AbstractPartitionTracker

isPartitionTracked, isTrackingPartitionsFor, stopTrackingPartitions, stopTrackingPartitionsFor

从类继承的方法 java.lang.Object

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

从接口继承的方法 org.apache.flink.runtime.io.network.partition.JobMasterPartitionTracker

getAllTrackedClusterPartitions, getAllTrackedNonClusterPartitions, stopTrackingAndReleasePartitions

从接口继承的方法 org.apache.flink.runtime.io.network.partition.PartitionTracker

isPartitionTracked, isTrackingPartitionsFor, stopTrackingPartitions, stopTrackingPartitionsFor

构造器详细资料

JobMasterPartitionTrackerImpl

public JobMasterPartitionTrackerImpl(org.apache.flink.api.common.JobID jobId,
                                     ShuffleMaster<?> shuffleMaster,
                                     PartitionTrackerFactory.TaskExecutorGatewayLookup taskExecutorGatewayLookup)

方法详细资料

startTrackingPartition

public void startTrackingPartition(ResourceID producingTaskExecutorId,
                                   ResultPartitionDeploymentDescriptor resultPartitionDeploymentDescriptor)

从接口复制的说明: JobMasterPartitionTracker

Starts the tracking of the given partition for the given task executor ID.

指定者:

startTrackingPartition 在接口中 JobMasterPartitionTracker

参数:

producingTaskExecutorId - ID of task executor on which the partition is produced

resultPartitionDeploymentDescriptor - deployment descriptor of the partition

stopTrackingAndReleasePartitions

public void stopTrackingAndReleasePartitions(Collection<ResultPartitionID> resultPartitionIds,
                                             boolean releaseOnShuffleMaster)

从接口复制的说明: JobMasterPartitionTracker

Releases the given partitions and stop the tracking of partitions that were released. The boolean flag indicates whether we need to notify the ShuffleMaster to release all external resources or not.

指定者:

stopTrackingAndReleasePartitions 在接口中 JobMasterPartitionTracker

stopTrackingAndPromotePartitions

public CompletableFuture<Void> stopTrackingAndPromotePartitions(Collection<ResultPartitionID> resultPartitionIds)

从接口复制的说明: JobMasterPartitionTracker

Promotes the given partitions, and stops the tracking of partitions that were promoted.

指定者:

stopTrackingAndPromotePartitions 在接口中 JobMasterPartitionTracker

参数:

resultPartitionIds - ID of the partition containing both job partitions and cluster partitions.

返回:

Future that will be completed if the partitions are promoted.

getAllTrackedPartitions

public Collection<ResultPartitionDeploymentDescriptor> getAllTrackedPartitions()

从接口复制的说明: JobMasterPartitionTracker

Gets all the partitions under tracking.

指定者:

getAllTrackedPartitions 在接口中 JobMasterPartitionTracker

connectToResourceManager

public void connectToResourceManager(ResourceManagerGateway resourceManagerGateway)

指定者:

connectToResourceManager 在接口中 JobMasterPartitionTracker

getClusterPartitionShuffleDescriptors

public List<ShuffleDescriptor> getClusterPartitionShuffleDescriptors(IntermediateDataSetID intermediateDataSetID)

从接口复制的说明: JobMasterPartitionTracker

Get the shuffle descriptors of the cluster partitions ordered by partition number.

指定者:

getClusterPartitionShuffleDescriptors 在接口中 JobMasterPartitionTracker