org.apache.cassandra.service.reads.repair

Class AbstractReadRepair<E extends Endpoints<E>,P extends ReplicaPlan.ForRead<E>>

java.lang.Object

org.apache.cassandra.service.reads.repair.AbstractReadRepair<E,P>

All Implemented Interfaces:

ReadRepair<E,P>

Direct Known Subclasses:

BlockingReadRepair, ReadOnlyReadRepair

public abstract class AbstractReadRepair<E extends Endpoints<E>,P extends ReplicaPlan.ForRead<E>>
extends java.lang.Object
implements ReadRepair<E,P>

Nested Class Summary

Nested classes/interfaces inherited from interface org.apache.cassandra.service.reads.repair.ReadRepair

ReadRepair.Factory

Field Summary

Fields

Modifier and Type	Field and Description
protected ColumnFamilyStore	cfs
protected ReadCommand	command
protected long	queryStartNanoTime
protected ReplicaPlan.Shared<E,P>	replicaPlan

Constructor Summary

Constructors

Constructor and Description
AbstractReadRepair(ReadCommand command, ReplicaPlan.Shared<E,P> replicaPlan, long queryStartNanoTime)

Method Summary

Modifier and Type	Method and Description
void	awaitReads() Block on the reads (or timeout) sent out in ReadRepair.startRepair(org.apache.cassandra.service.reads.DigestResolver<E, P>, java.util.function.Consumer<org.apache.cassandra.db.partitions.PartitionIterator>)
void	maybeSendAdditionalReads() if it looks like we might not receive data requests from everyone in time, send additional requests to additional replicas not contacted in the initial full data read.
protected P	replicaPlan()
void	startRepair(DigestResolver<E,P> digestResolver, java.util.function.Consumer<PartitionIterator> resultConsumer) Called when the digests from the initial read don't match.

Methods inherited from class java.lang.Object

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

Methods inherited from interface org.apache.cassandra.service.reads.repair.ReadRepair

awaitWrites, create, getMergeListener, maybeSendAdditionalWrites, repairPartition

Field Detail

command

protected final ReadCommand command

queryStartNanoTime

protected final long queryStartNanoTime

replicaPlan

protected final ReplicaPlan.Shared<E extends Endpoints<E>,P extends ReplicaPlan.ForRead<E>> replicaPlan

cfs

protected final ColumnFamilyStore cfs

Constructor Detail

AbstractReadRepair

public AbstractReadRepair(ReadCommand command,
                          ReplicaPlan.Shared<E,P> replicaPlan,
                          long queryStartNanoTime)

Method Detail

replicaPlan

protected P replicaPlan()

startRepair

public void startRepair(DigestResolver<E,P> digestResolver,
                        java.util.function.Consumer<PartitionIterator> resultConsumer)

Description copied from interface: ReadRepair

Called when the digests from the initial read don't match. Reads may block on the repair started by this method.

Specified by:

startRepair in interface ReadRepair<E extends Endpoints<E>,P extends ReplicaPlan.ForRead<E>>

Parameters:

digestResolver - supplied so we can get the original data response

resultConsumer - hook for the repair to set it's result on completion

awaitReads

public void awaitReads()
                throws ReadTimeoutException

Description copied from interface: ReadRepair

Block on the reads (or timeout) sent out in ReadRepair.startRepair(org.apache.cassandra.service.reads.DigestResolver<E, P>, java.util.function.Consumer<org.apache.cassandra.db.partitions.PartitionIterator>)

Specified by:

awaitReads in interface ReadRepair<E extends Endpoints<E>,P extends ReplicaPlan.ForRead<E>>

Throws:

ReadTimeoutException

maybeSendAdditionalReads

public void maybeSendAdditionalReads()

Description copied from interface: ReadRepair

if it looks like we might not receive data requests from everyone in time, send additional requests to additional replicas not contacted in the initial full data read. If the collection of nodes that end up responding in time end up agreeing on the data, and we don't consider the response from the disagreeing replica that triggered the read repair, that's ok, since the disagreeing data would not have been successfully written and won't be included in the response the the client, preserving the expectation of monotonic quorum reads

Specified by:

maybeSendAdditionalReads in interface ReadRepair<E extends Endpoints<E>,P extends ReplicaPlan.ForRead<E>>