io.prestosql.execution.scheduler

Class TopologyAwareNodeSelector

java.lang.Object

io.prestosql.execution.scheduler.TopologyAwareNodeSelector

All Implemented Interfaces:

NodeSelector

public class TopologyAwareNodeSelector
extends Object
implements NodeSelector

Constructor Summary

Constructors

Constructor and Description
TopologyAwareNodeSelector(InternalNodeManager nodeManager, NodeTaskMap nodeTaskMap, boolean includeCoordinator, com.google.common.base.Supplier<NodeMap> nodeMap, int minCandidates, int maxSplitsPerNode, int maxPendingSplitsPerTask, List<io.airlift.stats.CounterStat> topologicalSplitCounters, List<String> networkLocationSegmentNames, NetworkLocationCache networkLocationCache)

Method Summary

Modifier and Type	Method and Description
List<InternalNode>	allNodes()
SplitPlacementResult	computeAssignments(Set<Split> splits, List<RemoteTask> existingTasks) Identifies the nodes for running the specified splits.
SplitPlacementResult	computeAssignments(Set<Split> splits, List<RemoteTask> existingTasks, BucketNodeMap bucketNodeMap) Identifies the nodes for running the specified splits based on a precomputed fixed partitioning.
void	lockDownNodes()
InternalNode	selectCurrentNode()
List<InternalNode>	selectRandomNodes(int limit, Set<InternalNode> excludedNodes)

Methods inherited from class java.lang.Object

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

Methods inherited from interface io.prestosql.execution.scheduler.NodeSelector

selectRandomNodes

Constructor Detail

TopologyAwareNodeSelector

public TopologyAwareNodeSelector(InternalNodeManager nodeManager,
                                 NodeTaskMap nodeTaskMap,
                                 boolean includeCoordinator,
                                 com.google.common.base.Supplier<NodeMap> nodeMap,
                                 int minCandidates,
                                 int maxSplitsPerNode,
                                 int maxPendingSplitsPerTask,
                                 List<io.airlift.stats.CounterStat> topologicalSplitCounters,
                                 List<String> networkLocationSegmentNames,
                                 NetworkLocationCache networkLocationCache)

Method Detail

lockDownNodes

public void lockDownNodes()

Specified by:

lockDownNodes in interface NodeSelector

allNodes

public List<InternalNode> allNodes()

Specified by:

allNodes in interface NodeSelector

selectCurrentNode

public InternalNode selectCurrentNode()

Specified by:

selectCurrentNode in interface NodeSelector

selectRandomNodes

public List<InternalNode> selectRandomNodes(int limit,
                                            Set<InternalNode> excludedNodes)

Specified by:

selectRandomNodes in interface NodeSelector

computeAssignments

public SplitPlacementResult computeAssignments(Set<Split> splits,
                                               List<RemoteTask> existingTasks)

Description copied from interface: NodeSelector

Identifies the nodes for running the specified splits.

Specified by:

computeAssignments in interface NodeSelector

Parameters:

splits - the splits that need to be assigned to nodes

Returns:

a multimap from node to splits only for splits for which we could identify a node to schedule on. If we cannot find an assignment for a split, it is not included in the map. Also returns a future indicating when to reattempt scheduling of this batch of splits, if some of them could not be scheduled.

computeAssignments

public SplitPlacementResult computeAssignments(Set<Split> splits,
                                               List<RemoteTask> existingTasks,
                                               BucketNodeMap bucketNodeMap)

Description copied from interface: NodeSelector

Identifies the nodes for running the specified splits based on a precomputed fixed partitioning.

Specified by:

computeAssignments in interface NodeSelector

Parameters:

splits - the splits that need to be assigned to nodes

Returns:

a multimap from node to splits only for splits for which we could identify a node with free space. If we cannot find an assignment for a split, it is not included in the map. Also returns a future indicating when to reattempt scheduling of this batch of splits, if some of them could not be scheduled.