Package io.trino.execution.executor.timesharing

Class MultilevelSplitQueue

java.lang.Object

io.trino.execution.executor.timesharing.MultilevelSplitQueue

@ThreadSafe
public class MultilevelSplitQueue
extends Object

Constructor Summary

Constructors

Constructor	Description
MultilevelSplitQueue(double levelTimeMultiplier)
MultilevelSplitQueue(TaskManagerConfig taskManagerConfig)

Method Summary

Modifier and Type	Method	Description
static int	computeLevel(long threadUsageNanos)
long	getLevel0Time()
long	getLevel1Time()
long	getLevel2Time()
long	getLevel3Time()
long	getLevel4Time()
long	getLevelMinPriority(int level, long taskThreadUsageNanos)
io.airlift.stats.CounterStat	getSelectedCountLevel0()
io.airlift.stats.CounterStat	getSelectedCountLevel1()
io.airlift.stats.CounterStat	getSelectedCountLevel2()
io.airlift.stats.CounterStat	getSelectedCountLevel3()
io.airlift.stats.CounterStat	getSelectedCountLevel4()
void	offer(PrioritizedSplitRunner split)	During periods of time when a level has no waiting splits, it will not accumulate scheduled time and will fall behind relative to other levels.
void	remove(PrioritizedSplitRunner split)
void	removeAll(Collection<PrioritizedSplitRunner> splits)
int	size()
PrioritizedSplitRunner	take()
Priority	updatePriority(Priority oldPriority, long quantaNanos, long scheduledNanos)	Trino 'charges' the quanta run time to the task and the level it belongs to in an effort to maintain the target thread utilization ratios between levels and to maintain fairness within a level.

Methods inherited from class java.lang.Object

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

Constructor Details

MultilevelSplitQueue

@Inject
public MultilevelSplitQueue(TaskManagerConfig taskManagerConfig)

MultilevelSplitQueue

public MultilevelSplitQueue(double levelTimeMultiplier)

Method Details

offer

public void offer(PrioritizedSplitRunner split)

During periods of time when a level has no waiting splits, it will not accumulate scheduled time and will fall behind relative to other levels.

This can cause temporary starvation for other levels when splits do reach the previously-empty level.

To prevent this we set the scheduled time for levels which were empty to the expected scheduled time.

take

public PrioritizedSplitRunner take()
                            throws InterruptedException

Throws:

InterruptedException

updatePriority

public Priority updatePriority(Priority oldPriority,
 long quantaNanos,
 long scheduledNanos)

Trino 'charges' the quanta run time to the task and the level it belongs to in an effort to maintain the target thread utilization ratios between levels and to maintain fairness within a level.

Consider an example split where a read hung for several minutes. This is either a bug or a failing dependency. In either case we do not want to charge the task too much, and we especially do not want to charge the level too much - i.e. cause other queries in this level to starve.

Returns:

the new priority for the task

remove

public void remove(PrioritizedSplitRunner split)

removeAll

public void removeAll(Collection<PrioritizedSplitRunner> splits)

getLevelMinPriority

public long getLevelMinPriority(int level,
 long taskThreadUsageNanos)

size

public int size()

computeLevel

public static int computeLevel(long threadUsageNanos)

getLevel0Time

public long getLevel0Time()

getLevel1Time

public long getLevel1Time()

getLevel2Time

public long getLevel2Time()

getLevel3Time

public long getLevel3Time()

getLevel4Time

public long getLevel4Time()

getSelectedCountLevel0

public io.airlift.stats.CounterStat getSelectedCountLevel0()

getSelectedCountLevel1

public io.airlift.stats.CounterStat getSelectedCountLevel1()

getSelectedCountLevel2

public io.airlift.stats.CounterStat getSelectedCountLevel2()

getSelectedCountLevel3

public io.airlift.stats.CounterStat getSelectedCountLevel3()

getSelectedCountLevel4

public io.airlift.stats.CounterStat getSelectedCountLevel4()