Package org.apache.hadoop.yarn.server.resourcemanager.scheduler

Interface ResourceScheduler

All Superinterfaces:

org.apache.hadoop.yarn.event.EventHandler<SchedulerEvent>, Recoverable, YarnScheduler

All Known Subinterfaces:

MutableConfScheduler, PreemptableResourceScheduler

All Known Implementing Classes:

AbstractYarnScheduler, CapacityScheduler, FairScheduler, FifoScheduler

@LimitedPrivate("yarn")
@Evolving
public interface ResourceScheduler
extends YarnScheduler, Recoverable

This interface is the one implemented by the schedulers. It mainly extends YarnScheduler.

Method Summary

Modifier and Type	Method	Description
boolean	attemptAllocationOnNode(SchedulerApplicationAttempt appAttempt, org.apache.hadoop.yarn.api.records.SchedulingRequest schedulingRequest, SchedulerNode schedulerNode)	Attempts to allocate a SchedulerRequest on a Node.
java.util.List<org.apache.hadoop.yarn.api.records.NodeId>	getNodeIds(java.lang.String resourceName)	Get the NodeId available in the cluster by resource name.
void	reinitialize(org.apache.hadoop.conf.Configuration conf, RMContext rmContext)	Re-initialize the ResourceScheduler.
void	resetSchedulerMetrics()	Reset scheduler metrics.
void	setRMContext(RMContext rmContext)	Set RMContext for ResourceScheduler.

Methods inherited from interface org.apache.hadoop.yarn.event.EventHandler

handle

Methods inherited from interface org.apache.hadoop.yarn.server.resourcemanager.recovery.Recoverable

recover

Methods inherited from interface org.apache.hadoop.yarn.server.resourcemanager.scheduler.YarnScheduler

addQueue, allocate, checkAccess, checkAndGetApplicationLifetime, checkAndGetApplicationPriority, getAppResourceUsageReport, getAppsInQueue, getClusterResource, getMaxClusterLevelAppPriority, getMaximumApplicationLifetime, getMaximumResourceCapability, getMaximumResourceCapability, getMinimumResourceCapability, getNodeReport, getNormalizedResource, getNumClusterNodes, getPendingResourceRequestsForAttempt, getPendingSchedulingRequestsForAttempt, getPlanQueues, getQueueInfo, getQueueUserAclInfo, getResourceCalculator, getRMContainer, getRootQueueMetrics, getSchedulerAppInfo, getSchedulerNode, getSchedulingResourceTypes, getTransferredContainers, killAllAppsInQueue, moveAllApps, moveApplication, preValidateMoveApplication, removeQueue, setClusterMaxPriority, setEntitlement, updateApplicationPriority

Method Detail

setRMContext

void setRMContext(RMContext rmContext)

Set RMContext for ResourceScheduler. This method should be called immediately after instantiating a scheduler once.

Parameters:

rmContext - created by ResourceManager

reinitialize

void reinitialize(org.apache.hadoop.conf.Configuration conf,
                  RMContext rmContext)
           throws java.io.IOException

Re-initialize the ResourceScheduler.

Parameters:

conf - configuration

rmContext - RMContext.

Throws:

java.io.IOException - an I/O exception has occurred.

getNodeIds

java.util.List<org.apache.hadoop.yarn.api.records.NodeId> getNodeIds(java.lang.String resourceName)

Get the NodeId available in the cluster by resource name.

Parameters:

resourceName - resource name

Returns:

the number of available NodeId by resource name.

attemptAllocationOnNode

boolean attemptAllocationOnNode(SchedulerApplicationAttempt appAttempt,
                                org.apache.hadoop.yarn.api.records.SchedulingRequest schedulingRequest,
                                SchedulerNode schedulerNode)

Attempts to allocate a SchedulerRequest on a Node. NOTE: This ignores the numAllocations in the resource sizing and tries to allocate a SINGLE container only.

Parameters:

appAttempt - ApplicationAttempt.

schedulingRequest - SchedulingRequest.

schedulerNode - SchedulerNode.

Returns:

true if proposal was accepted.

resetSchedulerMetrics

void resetSchedulerMetrics()

Reset scheduler metrics.