Package org.apache.bookkeeper.client

Class RackawareEnsemblePlacementPolicy

java.lang.Object

org.apache.bookkeeper.client.TopologyAwareEnsemblePlacementPolicy

org.apache.bookkeeper.client.RackawareEnsemblePlacementPolicyImpl

org.apache.bookkeeper.client.RackawareEnsemblePlacementPolicy

All Implemented Interfaces:

EnsemblePlacementPolicy, ITopologyAwareEnsemblePlacementPolicy<BookieNode>

Direct Known Subclasses:

RegionAwareEnsemblePlacementPolicy

public class RackawareEnsemblePlacementPolicy
extends RackawareEnsemblePlacementPolicyImpl
implements ITopologyAwareEnsemblePlacementPolicy<BookieNode>

A placement policy implementation use rack information for placing ensembles.

See Also:

EnsemblePlacementPolicy

Nested Class Summary

Nested classes/interfaces inherited from class org.apache.bookkeeper.client.TopologyAwareEnsemblePlacementPolicy

TopologyAwareEnsemblePlacementPolicy.DefaultResolver, TopologyAwareEnsemblePlacementPolicy.DNSResolverDecorator, TopologyAwareEnsemblePlacementPolicy.EnsembleForReplacementWithNoConstraints, TopologyAwareEnsemblePlacementPolicy.RRTopologyAwareCoverageEnsemble, TopologyAwareEnsemblePlacementPolicy.TruePredicate

Nested classes/interfaces inherited from interface org.apache.bookkeeper.client.EnsemblePlacementPolicy

EnsemblePlacementPolicy.PlacementPolicyAdherence, EnsemblePlacementPolicy.PlacementResult<T>

Nested classes/interfaces inherited from interface org.apache.bookkeeper.client.ITopologyAwareEnsemblePlacementPolicy

ITopologyAwareEnsemblePlacementPolicy.Ensemble<T extends Node>, ITopologyAwareEnsemblePlacementPolicy.Predicate<T extends Node>

Field Summary

Fields

Modifier and Type	Field	Description
(package private) RackawareEnsemblePlacementPolicyImpl	slave

Fields inherited from class org.apache.bookkeeper.client.RackawareEnsemblePlacementPolicyImpl

enforceDurability, enforceMinNumRacksPerWriteQuorum, failedToResolveNetworkLocationCounter, ignoreLocalNodeInPlacementPolicy, LOCAL_FAIL_MASK, LOCAL_MASK, localNode, LOG, MASK_BITS, maxWeightMultiple, minNumRacksPerWriteQuorum, numWritableBookiesInDefaultRack, RACKNAME_DISTANCE_FROM_LEAVES, READ_ONLY_MASK, readReorderedCounter, REMOTE_FAIL_MASK, REMOTE_MASK, reorderReadsRandom, reorderThresholdPendingRequests, REPP_RANDOM_READ_REORDERING, SLOW_MASK, slowBookies, stabilizePeriodSeconds, statsLogger, timer, UNAVAIL_MASK, useHostnameResolveLocalNodePlacementPolicy

Fields inherited from class org.apache.bookkeeper.client.TopologyAwareEnsemblePlacementPolicy

bookieAddressResolver, bookieInfoMap, bookiesJoinedCounter, bookiesLeftCounter, dnsResolver, historyBookies, isWeighted, knownBookies, readOnlyBookies, REPP_DNS_RESOLVER_CLASS, rwLock, topology, weightedSelection

Constructor Summary

Constructors

Constructor	Description
RackawareEnsemblePlacementPolicy()
RackawareEnsemblePlacementPolicy(boolean enforceDurability)

Method Summary

Modifier and Type	Method	Description
void	handleBookiesThatJoined(java.util.Set<BookieId> joinedBookies)	Handle bookies that joined.
void	handleBookiesThatLeft(java.util.Set<BookieId> leftBookies)	Handle bookies that left.
protected RackawareEnsemblePlacementPolicy	initialize(DNSToSwitchMapping dnsResolver, io.netty.util.HashedWheelTimer timer, boolean reorderReadsRandom, int stabilizePeriodSeconds, int reorderThresholdPendingRequests, boolean isWeighted, int maxWeightMultiple, int minNumRacksPerWriteQuorum, boolean enforceMinNumRacksPerWriteQuorum, boolean ignoreLocalNodeInPlacementPolicy, boolean useHostnameResolveLocalNodePlacementPolicy, StatsLogger statsLogger, BookieAddressResolver bookieAddressResolver)	Initialize the policy.
protected RackawareEnsemblePlacementPolicy	initialize(DNSToSwitchMapping dnsResolver, io.netty.util.HashedWheelTimer timer, boolean reorderReadsRandom, int stabilizePeriodSeconds, int reorderThresholdPendingRequests, boolean isWeighted, int maxWeightMultiple, int minNumRacksPerWriteQuorum, boolean enforceMinNumRacksPerWriteQuorum, boolean ignoreLocalNodeInPlacementPolicy, StatsLogger statsLogger, BookieAddressResolver bookieAddressResolver)	Initialize the policy.
EnsemblePlacementPolicy.PlacementResult<java.util.List<BookieId>>	newEnsemble(int ensembleSize, int writeQuorumSize, int ackQuorumSize, java.util.Map<java.lang.String,byte[]> customMetadata, java.util.Set<BookieId> excludeBookies)	Choose numBookies bookies for ensemble.
EnsemblePlacementPolicy.PlacementResult<java.util.List<BookieId>>	newEnsemble(int ensembleSize, int writeQuorumSize, int ackQuorumSize, java.util.Set<BookieId> excludeBookies, ITopologyAwareEnsemblePlacementPolicy.Ensemble<BookieNode> parentEnsemble, ITopologyAwareEnsemblePlacementPolicy.Predicate<BookieNode> parentPredicate)	Create an ensemble with parent ensemble.
java.util.Set<BookieId>	onClusterChanged(java.util.Set<BookieId> writableBookies, java.util.Set<BookieId> readOnlyBookies)	A consistent view of the cluster (what bookies are available as writable, what bookies are available as readonly) is updated when any changes happen in the cluster.
DistributionSchedule.WriteSet	reorderReadLACSequence(java.util.List<BookieId> ensemble, BookiesHealthInfo bookiesHealthInfo, DistributionSchedule.WriteSet writeSet)	Reorder the read last add confirmed sequence of a given write quorum writeSet.
DistributionSchedule.WriteSet	reorderReadSequence(java.util.List<BookieId> ensemble, BookiesHealthInfo bookiesHealthInfo, DistributionSchedule.WriteSet writeSet)	Reorder the read sequence of a given write quorum writeSet.
EnsemblePlacementPolicy.PlacementResult<BookieId>	replaceBookie(int ensembleSize, int writeQuorumSize, int ackQuorumSize, java.util.Map<java.lang.String,byte[]> customMetadata, java.util.List<BookieId> currentEnsemble, BookieId bookieToReplace, java.util.Set<BookieId> excludeBookies)	Choose a new bookie to replace bookieToReplace.
EnsemblePlacementPolicy.PlacementResult<java.util.List<BookieId>>	replaceToAdherePlacementPolicy(int ensembleSize, int writeQuorumSize, int ackQuorumSize, java.util.Set<BookieId> excludeBookies, java.util.List<BookieId> currentEnsemble)	Returns placement result.
BookieNode	selectFromNetworkLocation(java.lang.String networkLoc, java.util.Set<java.lang.String> excludeRacks, java.util.Set<Node> excludeBookies, ITopologyAwareEnsemblePlacementPolicy.Predicate<BookieNode> predicate, ITopologyAwareEnsemblePlacementPolicy.Ensemble<BookieNode> ensemble, boolean fallbackToRandom)	Select a node from networkLoc rack excluding excludeBookies.
BookieNode	selectFromNetworkLocation(java.lang.String networkLoc, java.util.Set<Node> excludeBookies, ITopologyAwareEnsemblePlacementPolicy.Predicate<BookieNode> predicate, ITopologyAwareEnsemblePlacementPolicy.Ensemble<BookieNode> ensemble, boolean fallbackToRandom)	Select a node from a given network location.
BookieNode	selectFromNetworkLocation(java.util.Set<java.lang.String> excludeRacks, java.util.Set<Node> excludeBookies, ITopologyAwareEnsemblePlacementPolicy.Predicate<BookieNode> predicate, ITopologyAwareEnsemblePlacementPolicy.Ensemble<BookieNode> ensemble, boolean fallbackToRandom)	It randomly selects a BookieNode that is not on the excludeRacks set, excluding the nodes in excludeBookies set.
void	uninitalize()	Uninitialize the policy.

Methods inherited from class org.apache.bookkeeper.client.RackawareEnsemblePlacementPolicyImpl

addDefaultRackBookiesIfMinNumRacksIsEnforced, areAckedBookiesAdheringToPlacementPolicy, getDefaultRack, initialize, isEnsembleAdheringToPlacementPolicy, newEnsembleInternal, registerSlowBookie, reorderReadSequenceWithRegion, selectRandom, selectRandomFromRack, selectRandomInternal, withDefaultRack

Methods inherited from class org.apache.bookkeeper.client.TopologyAwareEnsemblePlacementPolicy

convertBookiesToNodes, convertBookieToNode, createBookieNode, createDummyLocalBookieNode, differBetweenBookies, getNetworkLocations, onBookieRackChange, resolveNetworkLocation, shuffleWithMask, updateBookieInfo

Methods inherited from class java.lang.Object

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

Methods inherited from interface org.apache.bookkeeper.client.EnsemblePlacementPolicy

areAckedBookiesAdheringToPlacementPolicy, getStickyReadBookieIndex, initialize, isEnsembleAdheringToPlacementPolicy, registerSlowBookie, updateBookieInfo

Methods inherited from interface org.apache.bookkeeper.client.ITopologyAwareEnsemblePlacementPolicy

onBookieRackChange

Field Detail

slave

RackawareEnsemblePlacementPolicyImpl slave

Constructor Detail

RackawareEnsemblePlacementPolicy

public RackawareEnsemblePlacementPolicy()

RackawareEnsemblePlacementPolicy

public RackawareEnsemblePlacementPolicy(boolean enforceDurability)

Method Detail

initialize

protected RackawareEnsemblePlacementPolicy initialize(DNSToSwitchMapping dnsResolver,
                                                      io.netty.util.HashedWheelTimer timer,
                                                      boolean reorderReadsRandom,
                                                      int stabilizePeriodSeconds,
                                                      int reorderThresholdPendingRequests,
                                                      boolean isWeighted,
                                                      int maxWeightMultiple,
                                                      int minNumRacksPerWriteQuorum,
                                                      boolean enforceMinNumRacksPerWriteQuorum,
                                                      boolean ignoreLocalNodeInPlacementPolicy,
                                                      StatsLogger statsLogger,
                                                      BookieAddressResolver bookieAddressResolver)

Description copied from class: RackawareEnsemblePlacementPolicyImpl

Initialize the policy.

Overrides:

initialize in class RackawareEnsemblePlacementPolicyImpl

Returns:

initialized ensemble placement policy

initialize

protected RackawareEnsemblePlacementPolicy initialize(DNSToSwitchMapping dnsResolver,
                                                      io.netty.util.HashedWheelTimer timer,
                                                      boolean reorderReadsRandom,
                                                      int stabilizePeriodSeconds,
                                                      int reorderThresholdPendingRequests,
                                                      boolean isWeighted,
                                                      int maxWeightMultiple,
                                                      int minNumRacksPerWriteQuorum,
                                                      boolean enforceMinNumRacksPerWriteQuorum,
                                                      boolean ignoreLocalNodeInPlacementPolicy,
                                                      boolean useHostnameResolveLocalNodePlacementPolicy,
                                                      StatsLogger statsLogger,
                                                      BookieAddressResolver bookieAddressResolver)

Description copied from class: RackawareEnsemblePlacementPolicyImpl

Initialize the policy.

Overrides:

initialize in class RackawareEnsemblePlacementPolicyImpl

Parameters:

dnsResolver - the object used to resolve addresses to their network address

Returns:

initialized ensemble placement policy

uninitalize

public void uninitalize()

Description copied from interface: EnsemblePlacementPolicy

Uninitialize the policy.

Specified by:

uninitalize in interface EnsemblePlacementPolicy

Overrides:

uninitalize in class RackawareEnsemblePlacementPolicyImpl

onClusterChanged

public java.util.Set<BookieId> onClusterChanged(java.util.Set<BookieId> writableBookies,
                                                java.util.Set<BookieId> readOnlyBookies)

Description copied from interface: EnsemblePlacementPolicy

A consistent view of the cluster (what bookies are available as writable, what bookies are available as readonly) is updated when any changes happen in the cluster.

The implementation should take actions when the cluster view is changed. So subsequent EnsemblePlacementPolicy.newEnsemble(int, int, int, Map, Set) and EnsemblePlacementPolicy.replaceBookie(int, int, int, java.util.Map, java.util.List, BookieId, java.util.Set) can choose proper bookies.

Specified by:

onClusterChanged in interface EnsemblePlacementPolicy

Overrides:

onClusterChanged in class TopologyAwareEnsemblePlacementPolicy

Parameters:

writableBookies - All the bookies in the cluster available for write/read.

readOnlyBookies - All the bookies in the cluster available for readonly.

Returns:

the dead bookies during this cluster change.

newEnsemble

public EnsemblePlacementPolicy.PlacementResult<java.util.List<BookieId>> newEnsemble(int ensembleSize,
                                                                                     int writeQuorumSize,
                                                                                     int ackQuorumSize,
                                                                                     java.util.Map<java.lang.String,byte[]> customMetadata,
                                                                                     java.util.Set<BookieId> excludeBookies)
                                                                              throws BKException.BKNotEnoughBookiesException

Description copied from interface: EnsemblePlacementPolicy

Choose numBookies bookies for ensemble. If the count is more than the number of available nodes, BKException.BKNotEnoughBookiesException is thrown.

The implementation should respect to the replace settings. The size of the returned bookie list should be equal to the provide ensembleSize.

customMetadata is the same user defined data that user provides when BookKeeper.createLedger(int, int, int, BookKeeper.DigestType, byte[], Map).

If 'enforceMinNumRacksPerWriteQuorum' config is enabled then the bookies belonging to default faultzone (rack) will be excluded while selecting bookies.

Specified by:

newEnsemble in interface EnsemblePlacementPolicy

Overrides:

newEnsemble in class RackawareEnsemblePlacementPolicyImpl

Parameters:

ensembleSize - Ensemble Size

writeQuorumSize - Write Quorum Size

ackQuorumSize - the value of ackQuorumSize (added since 4.5)

customMetadata - the value of customMetadata. it is the same user defined metadata that user provides in BookKeeper.createLedger(int, int, int, BookKeeper.DigestType, byte[])

excludeBookies - Bookies that should not be considered as targets.

Returns:

a placement result containing list of bookie addresses for the ensemble.

Throws:

BKException.BKNotEnoughBookiesException - if not enough bookies available.

replaceBookie

public EnsemblePlacementPolicy.PlacementResult<BookieId> replaceBookie(int ensembleSize,
                                                                       int writeQuorumSize,
                                                                       int ackQuorumSize,
                                                                       java.util.Map<java.lang.String,byte[]> customMetadata,
                                                                       java.util.List<BookieId> currentEnsemble,
                                                                       BookieId bookieToReplace,
                                                                       java.util.Set<BookieId> excludeBookies)
                                                                throws BKException.BKNotEnoughBookiesException

Description copied from interface: EnsemblePlacementPolicy

Choose a new bookie to replace bookieToReplace. If no bookie available in the cluster, BKException.BKNotEnoughBookiesException is thrown.

If 'enforceMinNumRacksPerWriteQuorum' config is enabled then the bookies belonging to default faultzone (rack) will be excluded while selecting bookies.

Specified by:

replaceBookie in interface EnsemblePlacementPolicy

Overrides:

replaceBookie in class RackawareEnsemblePlacementPolicyImpl

Parameters:

ensembleSize - the value of ensembleSize

writeQuorumSize - the value of writeQuorumSize

ackQuorumSize - the value of ackQuorumSize (added since 4.5)

customMetadata - the value of customMetadata. it is the same user defined metadata that user provides in BookKeeper.createLedger(int, int, int, BookKeeper.DigestType, byte[])

currentEnsemble - the value of currentEnsemble

bookieToReplace - bookie to replace

excludeBookies - bookies that should not be considered as candidate.

Returns:

a placement result containing the new bookie address.

Throws:

BKException.BKNotEnoughBookiesException

reorderReadSequence

public DistributionSchedule.WriteSet reorderReadSequence(java.util.List<BookieId> ensemble,
                                                         BookiesHealthInfo bookiesHealthInfo,
                                                         DistributionSchedule.WriteSet writeSet)

Description copied from interface: EnsemblePlacementPolicy

Reorder the read sequence of a given write quorum writeSet.

Specified by:

reorderReadSequence in interface EnsemblePlacementPolicy

Overrides:

reorderReadSequence in class RackawareEnsemblePlacementPolicyImpl

Parameters:

ensemble - Ensemble to read entries.

bookiesHealthInfo - Health info for bookies

writeSet - Write quorum to read entries. This will be modified, rather than allocating a new WriteSet.

Returns:

The read sequence. This will be the same object as the passed in writeSet.

reorderReadLACSequence

public DistributionSchedule.WriteSet reorderReadLACSequence(java.util.List<BookieId> ensemble,
                                                            BookiesHealthInfo bookiesHealthInfo,
                                                            DistributionSchedule.WriteSet writeSet)

Description copied from interface: EnsemblePlacementPolicy

Reorder the read last add confirmed sequence of a given write quorum writeSet.

Specified by:

reorderReadLACSequence in interface EnsemblePlacementPolicy

Overrides:

reorderReadLACSequence in class TopologyAwareEnsemblePlacementPolicy

Parameters:

ensemble - Ensemble to read entries.

bookiesHealthInfo - Health info for bookies

writeSet - Write quorum to read entries. This will be modified, rather than allocating a new WriteSet.

Returns:

The read sequence. This will be the same object as the passed in writeSet.

newEnsemble

public EnsemblePlacementPolicy.PlacementResult<java.util.List<BookieId>> newEnsemble(int ensembleSize,
                                                                                     int writeQuorumSize,
                                                                                     int ackQuorumSize,
                                                                                     java.util.Set<BookieId> excludeBookies,
                                                                                     ITopologyAwareEnsemblePlacementPolicy.Ensemble<BookieNode> parentEnsemble,
                                                                                     ITopologyAwareEnsemblePlacementPolicy.Predicate<BookieNode> parentPredicate)
                                                                              throws BKException.BKNotEnoughBookiesException

Description copied from interface: ITopologyAwareEnsemblePlacementPolicy

Create an ensemble with parent ensemble.

Specified by:

newEnsemble in interface ITopologyAwareEnsemblePlacementPolicy<BookieNode>

Overrides:

newEnsemble in class RackawareEnsemblePlacementPolicyImpl

Parameters:

ensembleSize - ensemble size

writeQuorumSize - write quorum size

ackQuorumSize - ack quorum size

excludeBookies - exclude bookies

parentEnsemble - parent ensemble

Returns:

list of bookies forming the ensemble

Throws:

BKException.BKNotEnoughBookiesException

selectFromNetworkLocation

public BookieNode selectFromNetworkLocation(java.lang.String networkLoc,
                                            java.util.Set<Node> excludeBookies,
                                            ITopologyAwareEnsemblePlacementPolicy.Predicate<BookieNode> predicate,
                                            ITopologyAwareEnsemblePlacementPolicy.Ensemble<BookieNode> ensemble,
                                            boolean fallbackToRandom)
                                     throws BKException.BKNotEnoughBookiesException

Description copied from interface: ITopologyAwareEnsemblePlacementPolicy

Select a node from a given network location.

Specified by:

selectFromNetworkLocation in interface ITopologyAwareEnsemblePlacementPolicy<BookieNode>

Overrides:

selectFromNetworkLocation in class RackawareEnsemblePlacementPolicyImpl

Parameters:

networkLoc - network location

excludeBookies - exclude bookies set

predicate - predicate to apply

ensemble - ensemble

fallbackToRandom - fallbackToRandom

Returns:

the selected bookie.

Throws:

BKException.BKNotEnoughBookiesException

selectFromNetworkLocation

public BookieNode selectFromNetworkLocation(java.util.Set<java.lang.String> excludeRacks,
                                            java.util.Set<Node> excludeBookies,
                                            ITopologyAwareEnsemblePlacementPolicy.Predicate<BookieNode> predicate,
                                            ITopologyAwareEnsemblePlacementPolicy.Ensemble<BookieNode> ensemble,
                                            boolean fallbackToRandom)
                                     throws BKException.BKNotEnoughBookiesException

Description copied from class: RackawareEnsemblePlacementPolicyImpl

It randomly selects a BookieNode that is not on the excludeRacks set, excluding the nodes in excludeBookies set. If it fails to find one, it selects a random BookieNode from the whole cluster.

Specified by:

selectFromNetworkLocation in interface ITopologyAwareEnsemblePlacementPolicy<BookieNode>

Overrides:

selectFromNetworkLocation in class RackawareEnsemblePlacementPolicyImpl

Returns:

Throws:

BKException.BKNotEnoughBookiesException

selectFromNetworkLocation

public BookieNode selectFromNetworkLocation(java.lang.String networkLoc,
                                            java.util.Set<java.lang.String> excludeRacks,
                                            java.util.Set<Node> excludeBookies,
                                            ITopologyAwareEnsemblePlacementPolicy.Predicate<BookieNode> predicate,
                                            ITopologyAwareEnsemblePlacementPolicy.Ensemble<BookieNode> ensemble,
                                            boolean fallbackToRandom)
                                     throws BKException.BKNotEnoughBookiesException

Description copied from interface: ITopologyAwareEnsemblePlacementPolicy

Select a node from networkLoc rack excluding excludeBookies. If there isn't any node in 'networkLoc', then it will try to get a node from cluster excluding excludeRacks and excludeBookies. If fallbackToRandom is set to true then it will get a random bookie from cluster excluding excludeBookies if it couldn't find a bookie

Specified by:

selectFromNetworkLocation in interface ITopologyAwareEnsemblePlacementPolicy<BookieNode>

Overrides:

selectFromNetworkLocation in class RackawareEnsemblePlacementPolicyImpl

Returns:

Throws:

BKException.BKNotEnoughBookiesException

replaceToAdherePlacementPolicy

public EnsemblePlacementPolicy.PlacementResult<java.util.List<BookieId>> replaceToAdherePlacementPolicy(int ensembleSize,
                                                                                                        int writeQuorumSize,
                                                                                                        int ackQuorumSize,
                                                                                                        java.util.Set<BookieId> excludeBookies,
                                                                                                        java.util.List<BookieId> currentEnsemble)

Description copied from interface: EnsemblePlacementPolicy

Returns placement result. If the currentEnsemble is not adhering placement policy, returns new ensemble that adheres placement policy. It should be implemented so as to minify the number of bookies replaced.

Specified by:

replaceToAdherePlacementPolicy in interface EnsemblePlacementPolicy

Overrides:

replaceToAdherePlacementPolicy in class RackawareEnsemblePlacementPolicyImpl

Parameters:

ensembleSize - ensemble size

writeQuorumSize - writeQuorumSize of the ensemble

ackQuorumSize - ackQuorumSize of the ensemble

excludeBookies - bookies that should not be considered as targets

currentEnsemble - current ensemble

Returns:

a placement result

handleBookiesThatLeft

public void handleBookiesThatLeft(java.util.Set<BookieId> leftBookies)

Description copied from interface: ITopologyAwareEnsemblePlacementPolicy

Handle bookies that left.

Specified by:

handleBookiesThatLeft in interface ITopologyAwareEnsemblePlacementPolicy<BookieNode>

Overrides:

handleBookiesThatLeft in class TopologyAwareEnsemblePlacementPolicy

Parameters:

leftBookies - bookies that left

handleBookiesThatJoined

public void handleBookiesThatJoined(java.util.Set<BookieId> joinedBookies)

Description copied from interface: ITopologyAwareEnsemblePlacementPolicy

Handle bookies that joined.

Specified by:

handleBookiesThatJoined in interface ITopologyAwareEnsemblePlacementPolicy<BookieNode>

Overrides:

handleBookiesThatJoined in class TopologyAwareEnsemblePlacementPolicy

Parameters:

joinedBookies - bookies that joined.