Flink - Scheduler

Job资源分配的过程，

在submitJob中，会生成ExecutionGraph
最终调用到，

executionGraph.scheduleForExecution(scheduler)

接着，ExecutionGraph

public void scheduleForExecution(SlotProvider slotProvider) throws JobException {

// simply take the vertices without inputs.
for (ExecutionJobVertex ejv : this.tasks.values()) {
   if (ejv.getJobVertex().isInputVertex()) {
      ejv.scheduleAll(slotProvider, allowQueuedScheduling);
   }
}

然后，ExecutionJobVertex

public void scheduleAll(SlotProvider slotProvider, boolean queued) throws NoResourceAvailableException {
   
   ExecutionVertex[] vertices = this.taskVertices;

   // kick off the tasks
   for (ExecutionVertex ev : vertices) {
      ev.scheduleForExecution(slotProvider, queued);
   }
}

再，ExecutionVertex

public boolean scheduleForExecution(SlotProvider slotProvider, boolean queued) throws NoResourceAvailableException {
   return this.currentExecution.scheduleForExecution(slotProvider, queued);
}

最终，Execution

public boolean scheduleForExecution(SlotProvider slotProvider, boolean queued) throws NoResourceAvailableException {

    final SlotSharingGroup sharingGroup = vertex.getJobVertex().getSlotSharingGroup();
    final CoLocationConstraint locationConstraint = vertex.getLocationConstraint();

    if (transitionState(CREATED, SCHEDULED)) {

        ScheduledUnit toSchedule = locationConstraint == null ?
            new ScheduledUnit(this, sharingGroup) :
            new ScheduledUnit(this, sharingGroup, locationConstraint);

        // IMPORTANT: To prevent leaks of cluster resources, we need to make sure that slots are returned
        //     in all cases where the deployment failed. we use many try {} finally {} clauses to assure that
        final Future<SimpleSlot> slotAllocationFuture = slotProvider.allocateSlot(toSchedule, queued); //异步去申请资源

        // IMPORTANT: We have to use the synchronous handle operation (direct executor) here so
        // that we directly deploy the tasks if the slot allocation future is completed. This is
        // necessary for immediate deployment.
        final Future<Void> deploymentFuture = slotAllocationFuture.handle(new BiFunction<SimpleSlot, Throwable, Void>() {
            @Override
            public Void apply(SimpleSlot simpleSlot, Throwable throwable) {
                if (simpleSlot != null) {
                    try {
                        deployToSlot(simpleSlot); //如果申请到，去部署
                    } catch (Throwable t) {
                        try {
                            simpleSlot.releaseSlot();
                        } finally {
                            markFailed(t);
                        }
                    }
                }
                else {
                    markFailed(throwable);
                }
                return null;
            }
        });
        
        return true;
    }

 

调用到，slotProvider.allocateSlot， slotProvider即Scheduler

@Override
public Future<SimpleSlot> allocateSlot(ScheduledUnit task, boolean allowQueued)
        throws NoResourceAvailableException {

    final Object ret = scheduleTask(task, allowQueued);
    if (ret instanceof SimpleSlot) {
        return FlinkCompletableFuture.completed((SimpleSlot) ret); //如果是SimpleSlot，即已经分配成功，表示future结束
    }
    else if (ret instanceof Future) {
        return (Future) ret; //Future说明没有足够资源，申请还在异步中，继续future
    }
    else {
        throw new RuntimeException();
    }
}

 

scheduleTask

   /**
     * Returns either a {@link SimpleSlot}, or a {@link Future}.
     */
    private Object scheduleTask(ScheduledUnit task, boolean queueIfNoResource) throws NoResourceAvailableException {

        final ExecutionVertex vertex = task.getTaskToExecute().getVertex();
        
        final Iterable<TaskManagerLocation> preferredLocations = vertex.getPreferredLocations();
        final boolean forceExternalLocation = vertex.isScheduleLocalOnly() &&
                                    preferredLocations != null && preferredLocations.iterator().hasNext(); //如果preferredLocations不为空，且vertex仅能local schedule
    
        synchronized (globalLock) { //全局锁
            
            SlotSharingGroup sharingUnit = task.getSlotSharingGroup();
            
            if (sharingUnit != null) { //如果有SlotSharingGroup

                // 1)  === If the task has a slot sharing group, schedule with shared slots ===
                
                final SlotSharingGroupAssignment assignment = sharingUnit.getTaskAssignment();
                final CoLocationConstraint constraint = task.getLocationConstraint();
                
                // get a slot from the group, if the group has one for us (and can fulfill the constraint)
                final SimpleSlot slotFromGroup;
                if (constraint == null) {
                    slotFromGroup = assignment.getSlotForTask(vertex); //通过SlotSharingGroupAssignment来分配slot
                }
                else {
                    slotFromGroup = assignment.getSlotForTask(vertex, constraint);
                }

                SimpleSlot newSlot = null;
                SimpleSlot toUse = null;

                // the following needs to make sure any allocated slot is released in case of an error
                try {
                    
                    // check whether the slot from the group is already what we want.
                    // any slot that is local, or where the assignment was unconstrained is good!
                    if (slotFromGroup != null && slotFromGroup.getLocality() != Locality.NON_LOCAL) { //如果找到local slot
                        updateLocalityCounters(slotFromGroup, vertex);
                        return slotFromGroup; //已经找到合适的slot，返回
                    }
                    
                    
                    // the group did not have a local slot for us. see if we can one (or a better one)
                    // our location preference is either determined by the location constraint, or by the
                    // vertex's preferred locations
                    final Iterable<TaskManagerLocation> locations;
                    final boolean localOnly;
                    if (constraint != null && constraint.isAssigned()) { //如果有constraint
                        locations = Collections.singleton(constraint.getLocation());
                        localOnly = true;
                    }
                    else {
                        locations = vertex.getPreferredLocationsBasedOnInputs(); //否则，以输入节点所分配的slot的location信息，作为Preferred
                        localOnly = forceExternalLocation;
                    }
                    // the group did not have a local slot for us. see if we can one (or a better one)
                    newSlot = getNewSlotForSharingGroup(vertex, locations, assignment, constraint, localOnly); //试图为SharingGroup申请一个新的slot

                    
                    if (slotFromGroup == null || !slotFromGroup.isAlive() || newSlot.getLocality() == Locality.LOCAL) {//如果newSlot是local的，那么就是使用newSlot
                        // if there is no slot from the group, or the new slot is local,
                        // then we use the new slot
                        if (slotFromGroup != null) {
                            slotFromGroup.releaseSlot();
                        }
                        toUse = newSlot; //使用新new的slot
                    }
                    else {
                        // both are available and usable. neither is local. in that case, we may
                        // as well use the slot from the sharing group, to minimize the number of
                        // instances that the job occupies
                        newSlot.releaseSlot();
                        toUse = slotFromGroup;
                    }

                    // if this is the first slot for the co-location constraint, we lock
                    // the location, because we are going to use that slot
                    if (constraint != null && !constraint.isAssigned()) {
                        constraint.lockLocation();
                    }
                    
                    updateLocalityCounters(toUse, vertex);
                }
                
                return toUse; //返回申请的slot
            }
            else { //如果不是共享slot，比较简单
                
                // 2) === schedule without hints and sharing ===
                
                SimpleSlot slot = getFreeSlotForTask(vertex, preferredLocations, forceExternalLocation); //直接申请slot
                if (slot != null) {
                    updateLocalityCounters(slot, vertex);
                    return slot; //申请到了就返回slot
                }
                else {
                    // no resource available now, so queue the request
                    if (queueIfNoResource) { //如果可以queue
                        CompletableFuture<SimpleSlot> future = new FlinkCompletableFuture<>();
                        this.taskQueue.add(new QueuedTask(task, future)); //把task缓存起来，并把future对象返回，表示异步申请
                        return future;
                    }
                }
            }
        }
    }

 

如果有SlotSharingGroup

首先试图从SlotSharingGroupAssignment中分配slot

slotFromGroup = assignment.getSlotForTask(vertex)， 参考，Flink – SlotSharingGroup

如果没有发现local的slot，试图为该vertex创建一个新的slot，

newSlot = getNewSlotForSharingGroup(vertex, locations, assignment, constraint, localOnly); //试图为SharingGroup申请一个新的slot

    protected SimpleSlot getNewSlotForSharingGroup(ExecutionVertex vertex,
                                                    Iterable<TaskManagerLocation> requestedLocations,
                                                    SlotSharingGroupAssignment groupAssignment,
                                                    CoLocationConstraint constraint,
                                                    boolean localOnly)
    {
        // we need potentially to loop multiple times, because there may be false positives
        // in the set-with-available-instances
        while (true) {
            Pair<Instance, Locality> instanceLocalityPair = findInstance(requestedLocations, localOnly); //根据locations信息找到local的instance
            
            if (instanceLocalityPair == null) { //如果没有可用的instance，返回null
                // nothing is available
                return null;
            }

            final Instance instanceToUse = instanceLocalityPair.getLeft();
            final Locality locality = instanceLocalityPair.getRight();

            try {
                JobVertexID groupID = vertex.getJobvertexId();
                
                // allocate a shared slot from the instance
                SharedSlot sharedSlot = instanceToUse.allocateSharedSlot(vertex.getJobId(), groupAssignment); //从instance申请一个SharedSlot

                // if the instance has further available slots, re-add it to the set of available resources.
                if (instanceToUse.hasResourcesAvailable()) { //如果这个instance还有多余的资源，再加入instancesWithAvailableResources，下次还能继续用来分配
                    this.instancesWithAvailableResources.put(instanceToUse.getTaskManagerID(), instanceToUse);
                }

                if (sharedSlot != null) {
                    // add the shared slot to the assignment group and allocate a sub-slot
                    SimpleSlot slot = constraint == null ?
                            groupAssignment.addSharedSlotAndAllocateSubSlot(sharedSlot, locality, groupID) : //把分配的SharedSlot加到SlotSharingGroup的SlotSharingGroupAssignment中，并返回SharedSlot所持有的slot
                            groupAssignment.addSharedSlotAndAllocateSubSlot(sharedSlot, locality, constraint);

                    if (slot != null) {
                        return slot;
                    }
                    else {
                        // could not add and allocate the sub-slot, so release shared slot
                        sharedSlot.releaseSlot();
                    }
                }
            }
            catch (InstanceDiedException e) {
                // the instance died it has not yet been propagated to this scheduler
                // remove the instance from the set of available instances
                removeInstance(instanceToUse);
            }

            // if we failed to get a slot, fall through the loop
        }
    }

findInstance

    private Pair<Instance, Locality> findInstance(Iterable<TaskManagerLocation> requestedLocations, boolean localOnly) {
        
        // drain the queue of newly available instances
        while (this.newlyAvailableInstances.size() > 0) { //BlockingQueue<Instance> newlyAvailableInstances
            Instance queuedInstance = this.newlyAvailableInstances.poll();
            if (queuedInstance != null) {
                this.instancesWithAvailableResources.put(queuedInstance.getTaskManagerID(), queuedInstance); // Map<ResourceID, Instance> instancesWithAvailableResources
            }
        }
        
        // if nothing is available at all, return null
        if (this.instancesWithAvailableResources.isEmpty()) {
            return null;
        }

        Iterator<TaskManagerLocation> locations = requestedLocations == null ? null : requestedLocations.iterator();

        if (locations != null && locations.hasNext()) { //如果有prefered locations，优先找相对应的Instance
            // we have a locality preference

            while (locations.hasNext()) {
                TaskManagerLocation location = locations.next();
                if (location != null) {
                    Instance instance = instancesWithAvailableResources.remove(location.getResourceID()); //找到对应于perfer location的Instance
                    if (instance != null) {
                        return new ImmutablePair<Instance, Locality>(instance, Locality.LOCAL);
                    }
                }
            }
            
            // no local instance available
            if (localOnly) { //如果localOnly，而前面又没有找到local的，所以只能返回null
                return null;
            }
            else {
                // take the first instance from the instances with resources
                Iterator<Instance> instances = instancesWithAvailableResources.values().iterator();
                Instance instanceToUse = instances.next();
                instances.remove();

                return new ImmutablePair<>(instanceToUse, Locality.NON_LOCAL); //由于前面没有找到local的，所以返回第一个instance，locality为non_local
            }
        }
        else {
            // no location preference, so use some instance
            Iterator<Instance> instances = instancesWithAvailableResources.values().iterator();
            Instance instanceToUse = instances.next();
            instances.remove();

            return new ImmutablePair<>(instanceToUse, Locality.UNCONSTRAINED); //没有约束，也是取第一个instance，locality为UNCONSTRAINED
        }
    }

Instance.allocateSharedSlot

    public SharedSlot allocateSharedSlot(JobID jobID, SlotSharingGroupAssignment sharingGroupAssignment)
            throws InstanceDiedException
    {
        synchronized (instanceLock) {
            if (isDead) {
                throw new InstanceDiedException(this);
            }

            Integer nextSlot = availableSlots.poll(); //Queue<Integer> availableSlots;
            if (nextSlot == null) {
                return null;
            }
            else {
                SharedSlot slot = new SharedSlot(
                        jobID, this, location, nextSlot, taskManagerGateway, sharingGroupAssignment);
                allocatedSlots.add(slot); //Set<Slot> allocatedSlots
                return slot;
            }
        }
    }

如果新分配的slot是local的，就用newSlot；如果不是并且当前SlotSharingGroup中是有non-local的slot，就用现成的slot，没必要使用新的slot，这时需要把newSlot释放掉

如果没有SlotSharingGroup

简单的调用

SimpleSlot slot = getFreeSlotForTask(vertex, preferredLocations, forceExternalLocation);

     protected SimpleSlot getFreeSlotForTask(ExecutionVertex vertex,
                                            Iterable<TaskManagerLocation> requestedLocations,
                                            boolean localOnly) {
        // we need potentially to loop multiple times, because there may be false positives
        // in the set-with-available-instances
        while (true) {
            Pair<Instance, Locality> instanceLocalityPair = findInstance(requestedLocations, localOnly); //找到一个合适的instance

            if (instanceLocalityPair == null){
                return null;
            }

            Instance instanceToUse = instanceLocalityPair.getLeft();
            Locality locality = instanceLocalityPair.getRight();

            try {
                SimpleSlot slot = instanceToUse.allocateSimpleSlot(vertex.getJobId()); //分配一个simpleSlot
                
                // if the instance has further available slots, re-add it to the set of available resources.
                if (instanceToUse.hasResourcesAvailable()) {
                    this.instancesWithAvailableResources.put(instanceToUse.getTaskManagerID(), instanceToUse);
                }
                
                if (slot != null) {
                    slot.setLocality(locality);
                    return slot;
                }
            }
            catch (InstanceDiedException e) {
                // the instance died it has not yet been propagated to this scheduler
                // remove the instance from the set of available instances
                removeInstance(instanceToUse);
            }
            
            // if we failed to get a slot, fall through the loop
        }
    }

逻辑和分配SharedSlot基本相同，只是会调用，

    public SimpleSlot allocateSimpleSlot(JobID jobID) throws InstanceDiedException {
        if (jobID == null) {
            throw new IllegalArgumentException();
        }

        synchronized (instanceLock) {
            if (isDead) {
                throw new InstanceDiedException(this);
            }

            Integer nextSlot = availableSlots.poll();
            if (nextSlot == null) {
                return null;
            }
            else {
                SimpleSlot slot = new SimpleSlot(jobID, this, location, nextSlot, taskManagerGateway);
                allocatedSlots.add(slot);
                return slot;
            }
        }
    }

 

Instance

Scheduler中的Instance怎么来的？

Scheduler实现InstanceListener接口的

newInstanceAvailable

    @Override
    public void newInstanceAvailable(Instance instance) {
        
        // synchronize globally for instance changes
        synchronized (this.globalLock) {
            
            // check we do not already use this instance
            if (!this.allInstances.add(instance)) { //看看是否已经有了这个instance
                throw new IllegalArgumentException("The instance is already contained.");
            }
            
            try {
                // make sure we get notifications about slots becoming available
                instance.setSlotAvailabilityListener(this); //加上SlotAvailabilityListener，当slot ready的时候，可以被通知
                
                // store the instance in the by-host-lookup
                String instanceHostName = instance.getTaskManagerLocation().getHostname(); 
                Set<Instance> instanceSet = allInstancesByHost.get(instanceHostName); // HashMap<String, Set<Instance>> allInstancesByHost
                if (instanceSet == null) {
                    instanceSet = new HashSet<Instance>();
                    allInstancesByHost.put(instanceHostName, instanceSet);
                }
                instanceSet.add(instance);

                // add it to the available resources and let potential waiters know
                this.instancesWithAvailableResources.put(instance.getTaskManagerID(), instance); // Map<ResourceID, Instance> instancesWithAvailableResources

                // add all slots as available
                for (int i = 0; i < instance.getNumberOfAvailableSlots(); i++) { //多次触发newSlotAvailable
                    newSlotAvailable(instance);
                }
            }
            catch (Throwable t) {
                LOG.error("Scheduler could not add new instance " + instance, t);
                removeInstance(instance);
            }
        }
    }

 

newInstanceAvailable，何时被调用，

 

JobManager

      case msg @ RegisterTaskManager(
          resourceId,
          connectionInfo,
          hardwareInformation,
          numberOfSlots) =>
          
          
          val instanceID = instanceManager.registerTaskManager(
            taskManagerGateway,
            connectionInfo,
            hardwareInformation,
            numberOfSlots)

 

InstanceManager

    public InstanceID registerTaskManager(
            TaskManagerGateway taskManagerGateway,
            TaskManagerLocation taskManagerLocation,
            HardwareDescription resources,
            int numberOfSlots) {
        
        synchronized (this.lock) {

            InstanceID instanceID = new InstanceID();

            Instance host = new Instance(
                taskManagerGateway,
                taskManagerLocation,
                instanceID,
                resources,
                numberOfSlots);

            // notify all listeners (for example the scheduler)
            notifyNewInstance(host);

            return instanceID;
        }
    }

 

    private void notifyNewInstance(Instance instance) {
        synchronized (this.instanceListeners) {
            for (InstanceListener listener : this.instanceListeners) {
                try {
                    listener.newInstanceAvailable(instance);
                }
                catch (Throwable t) {
                    LOG.error("Notification of new instance availability failed.", t);
                }
            }
        }
    }

 

Scheduler还是实现SlotAvailabilityListener

会调用newSlotAvailable

逻辑只是check是否有待分配的task，当有新的slot ready的时候，把queuedTask的future complete掉

@Override
    public void newSlotAvailable(final Instance instance) {
        
        // WARNING: The asynchrony here is necessary, because  we cannot guarantee the order
        // of lock acquisition (global scheduler, instance) and otherwise lead to potential deadlocks:
        // 
        // -> The scheduler needs to grab them (1) global scheduler lock
        //                                     (2) slot/instance lock
        // -> The slot releasing grabs (1) slot/instance (for releasing) and
        //                             (2) scheduler (to check whether to take a new task item
        // 
        // that leads with a high probability to deadlocks, when scheduling fast

        this.newlyAvailableInstances.add(instance);

        Futures.future(new Callable<Object>() {
            @Override
            public Object call() throws Exception {
                handleNewSlot();
                return null;
            }
        }, executionContext);
    }
    
    private void handleNewSlot() {
        
        synchronized (globalLock) {
            Instance instance = this.newlyAvailableInstances.poll();
            if (instance == null || !instance.hasResourcesAvailable()) {
                // someone else took it
                return;
            }
            
            QueuedTask queued = taskQueue.peek(); //如果有待分配的task
            
            // the slot was properly released, we can allocate a new one from that instance
            
            if (queued != null) {
                ScheduledUnit task = queued.getTask();
                ExecutionVertex vertex = task.getTaskToExecute().getVertex();
                
                try {
                    SimpleSlot newSlot = instance.allocateSimpleSlot(vertex.getJobId()); //从instance分配一个simpleSlot
                    if (newSlot != null) {
                        
                        // success, remove from the task queue and notify the future
                        taskQueue.poll();
                        if (queued.getFuture() != null) {
                            try {
                                queued.getFuture().complete(newSlot); //complete该task的future，有slot了，你不用继续等了
                            }
                            catch (Throwable t) {
                                LOG.error("Error calling allocation future for task " + vertex.getSimpleName(), t);
                                task.getTaskToExecute().fail(t);
                            }
                        }
                    }
                }
                catch (InstanceDiedException e) {
                    if (LOG.isDebugEnabled()) {
                        LOG.debug("Instance " + instance + " was marked dead asynchronously.");
                    }
                    
                    removeInstance(instance);
                }
            }
            else { //如果没有排队的task，直接把instance放到instancesWithAvailableResources就好
                this.instancesWithAvailableResources.put(instance.getTaskManagerID(), instance);
            }
        }
    }

 

newSlotAvailable除了当new instance注册时被调用外，还会在Instance.returnAllocatedSlot，即有人释放AllocatedSlot时，会被调用