org.apache.cassandra.repair

Class RepairSession

java.lang.Object

com.google.common.util.concurrent.AbstractFuture<RepairSessionResult>

org.apache.cassandra.repair.RepairSession

All Implemented Interfaces:

com.google.common.util.concurrent.ListenableFuture<RepairSessionResult>, java.util.concurrent.Future<RepairSessionResult>, IEndpointStateChangeSubscriber, IFailureDetectionEventListener

public class RepairSession
extends com.google.common.util.concurrent.AbstractFuture<RepairSessionResult>
implements IEndpointStateChangeSubscriber, IFailureDetectionEventListener

Coordinates the (active) repair of a list of non overlapping token ranges. A given RepairSession repairs a set of replicas for a given set of ranges on a list of column families. For each of the column family to repair, RepairSession creates a RepairJob that handles the repair of that CF. A given RepairJob has the 2 main phases:

1. Validation phase: the job requests merkle trees from each of the replica involves (ValidationTask) and waits until all trees are received (in validationComplete()).
2. Synchronization phase: once all trees are received, the job compares each tree with all the other using a so-called SyncTask. If there is difference between 2 trees, the concerned SyncTask will start a streaming of the difference between the 2 endpoint concerned.

The job is done once all its SyncTasks are done (i.e. have either computed no differences or the streaming they started is done (syncComplete())). A given session will execute the first phase (validation phase) of each of it's job sequentially. In other words, it will start the first job and only start the next one once that first job validation phase is complete. This is done so that the replica only create one merkle tree per range at a time, which is our way to ensure that such creation starts roughly at the same time on every node (see CASSANDRA-2816). However the synchronization phases are allowed to run concurrently (with each other and with validation phases). A given RepairJob has 2 modes: either sequential or not (RepairParallelism). If sequential, it will requests merkle tree creation from each replica in sequence (though in that case we still first send a message to each node to flush and snapshot data so each merkle tree creation is still done on similar data, even if the actual creation is not done simulatneously). If not sequential, all merkle tree are requested in parallel. Similarly, if a job is sequential, it will handle one SyncTask at a time, but will handle all of them in parallel otherwise.

Field Summary

Fields

Modifier and Type	Field and Description
java.util.Set<java.net.InetAddress>	endpoints
java.lang.String	keyspace
RepairParallelism	parallelismDegree
java.util.UUID	parentRepairSession
java.util.Collection<Range<Token>>	ranges Range to repair
long	repairedAt
com.google.common.util.concurrent.ListeningExecutorService	taskExecutor

Constructor Summary

Constructors

Constructor and Description
RepairSession(java.util.UUID parentRepairSession, java.util.UUID id, java.util.Collection<Range<Token>> ranges, java.lang.String keyspace, RepairParallelism parallelismDegree, java.util.Set<java.net.InetAddress> endpoints, long repairedAt, java.lang.String... cfnames) Create new repair session.

Method Summary

Modifier and Type	Method and Description
void	beforeChange(java.net.InetAddress endpoint, EndpointState currentState, ApplicationState newStateKey, VersionedValue newValue)
void	convict(java.net.InetAddress endpoint, double phi) Convict the specified endpoint.
void	forceShutdown(java.lang.Throwable reason) clear all RepairJobs and terminate this session.
java.util.UUID	getId()
java.util.Collection<Range<Token>>	getRanges()
void	onAlive(java.net.InetAddress endpoint, EndpointState state)
void	onChange(java.net.InetAddress endpoint, ApplicationState state, VersionedValue value)
void	onDead(java.net.InetAddress endpoint, EndpointState state)
void	onJoin(java.net.InetAddress endpoint, EndpointState epState) Use to inform interested parties about the change in the state for specified endpoint
void	onRemove(java.net.InetAddress endpoint)
void	onRestart(java.net.InetAddress endpoint, EndpointState epState) Called whenever a node is restarted.
void	start(com.google.common.util.concurrent.ListeningExecutorService executor) Start RepairJob on given ColumnFamilies.
void	syncComplete(RepairJobDesc desc, NodePair nodes, boolean success) Notify this session that sync completed/failed with given NodePair.
void	terminate()
void	validationComplete(RepairJobDesc desc, java.net.InetAddress endpoint, MerkleTrees trees) Receive merkle tree response or failed response from endpoint for current repair job.
void	waitForSync(Pair<RepairJobDesc,NodePair> key, RemoteSyncTask task)
void	waitForValidation(Pair<RepairJobDesc,java.net.InetAddress> key, ValidationTask task)

Methods inherited from class com.google.common.util.concurrent.AbstractFuture

addListener, cancel, get, get, interruptTask, isCancelled, isDone, set, setException, wasInterrupted

Methods inherited from class java.lang.Object

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

Field Detail

parentRepairSession

public final java.util.UUID parentRepairSession

keyspace

public final java.lang.String keyspace

parallelismDegree

public final RepairParallelism parallelismDegree

ranges

public final java.util.Collection<Range<Token>> ranges

Range to repair

endpoints

public final java.util.Set<java.net.InetAddress> endpoints

repairedAt

public final long repairedAt

taskExecutor

public final com.google.common.util.concurrent.ListeningExecutorService taskExecutor

Constructor Detail

RepairSession

public RepairSession(java.util.UUID parentRepairSession,
                     java.util.UUID id,
                     java.util.Collection<Range<Token>> ranges,
                     java.lang.String keyspace,
                     RepairParallelism parallelismDegree,
                     java.util.Set<java.net.InetAddress> endpoints,
                     long repairedAt,
                     java.lang.String... cfnames)

Create new repair session.

Parameters:

parentRepairSession - the parent sessions id

id - this sessions id

ranges - ranges to repair

keyspace - name of keyspace

parallelismDegree - specifies the degree of parallelism when calculating the merkle trees

endpoints - the data centers that should be part of the repair; null for all DCs

repairedAt - when the repair occurred (millis)

cfnames - names of columnfamilies

Method Detail

getId

public java.util.UUID getId()

getRanges

public java.util.Collection<Range<Token>> getRanges()

waitForValidation

public void waitForValidation(Pair<RepairJobDesc,java.net.InetAddress> key,
                              ValidationTask task)

waitForSync

public void waitForSync(Pair<RepairJobDesc,NodePair> key,
                        RemoteSyncTask task)

validationComplete

public void validationComplete(RepairJobDesc desc,
                               java.net.InetAddress endpoint,
                               MerkleTrees trees)

Receive merkle tree response or failed response from endpoint for current repair job.

Parameters:

desc - repair job description

endpoint - endpoint that sent merkle tree

trees - calculated merkle trees, or null if validation failed

syncComplete

public void syncComplete(RepairJobDesc desc,
                         NodePair nodes,
                         boolean success)

Notify this session that sync completed/failed with given NodePair.

Parameters:

desc - synced repair job

nodes - nodes that completed sync

success - true if sync succeeded

start

public void start(com.google.common.util.concurrent.ListeningExecutorService executor)

Start RepairJob on given ColumnFamilies. This first validates if all replica are available, and if they are, creates RepairJobs and submit to run on given executor.

Parameters:

executor - Executor to run validation

terminate

public void terminate()

forceShutdown

public void forceShutdown(java.lang.Throwable reason)

clear all RepairJobs and terminate this session.

Parameters:

reason - Cause of error for shutdown

onJoin

public void onJoin(java.net.InetAddress endpoint,
                   EndpointState epState)

Description copied from interface: IEndpointStateChangeSubscriber

Use to inform interested parties about the change in the state for specified endpoint

Specified by:

onJoin in interface IEndpointStateChangeSubscriber

Parameters:

endpoint - endpoint for which the state change occurred.

epState - state that actually changed for the above endpoint.

beforeChange

public void beforeChange(java.net.InetAddress endpoint,
                         EndpointState currentState,
                         ApplicationState newStateKey,
                         VersionedValue newValue)

Specified by:

beforeChange in interface IEndpointStateChangeSubscriber

onChange

public void onChange(java.net.InetAddress endpoint,
                     ApplicationState state,
                     VersionedValue value)

Specified by:

onChange in interface IEndpointStateChangeSubscriber

onAlive

public void onAlive(java.net.InetAddress endpoint,
                    EndpointState state)

Specified by:

onAlive in interface IEndpointStateChangeSubscriber

onDead

public void onDead(java.net.InetAddress endpoint,
                   EndpointState state)

Specified by:

onDead in interface IEndpointStateChangeSubscriber

onRemove

public void onRemove(java.net.InetAddress endpoint)

Specified by:

onRemove in interface IEndpointStateChangeSubscriber

onRestart

public void onRestart(java.net.InetAddress endpoint,
                      EndpointState epState)

Description copied from interface: IEndpointStateChangeSubscriber

Called whenever a node is restarted. Note that there is no guarantee when that happens that the node was previously marked down. It will have only if state.isAlive() == false as state is from before the restarted node is marked up.

Specified by:

onRestart in interface IEndpointStateChangeSubscriber

convict

public void convict(java.net.InetAddress endpoint,
                    double phi)

Description copied from interface: IFailureDetectionEventListener

Convict the specified endpoint.

Specified by:

convict in interface IFailureDetectionEventListener

Parameters:

endpoint - endpoint to be convicted

phi - the value of phi with with ep was convicted