org.objectweb.cjdbc.controller.recoverylog
Class RecoverThread

java.lang.Object
  java.lang.Thread
      org.objectweb.cjdbc.controller.recoverylog.RecoverThread

All Implemented Interfaces:

java.lang.Runnable

public class RecoverThread

extends java.lang.Thread

This class defines a RecoverThread that is in charge of replaying the recovery log on a given backend to re-synchronize it with the other nodes of the cluster.

Version:

1.0

Author:

Nicolas Modrzyk , Emmanuel Cecchet

Field Summary

private DatabaseBackend	backend
private BackendWorkerThread	bwt
private java.lang.String	checkpointName
private java.sql.SQLException	exception
private AbstractLoadBalancer	loadBalancer
(package private) static Trace	logger
private int	recoveryBatchSize Size of the pendingRecoveryTasks queue used during recovery
private RecoveryLog	recoveryLog
private AbstractScheduler	scheduler
private java.util.ArrayList	tids

Fields inherited from class java.lang.Thread

MAX_PRIORITY, MIN_PRIORITY, NORM_PRIORITY

Constructor Summary

RecoverThread(AbstractScheduler scheduler, RecoveryLog recoveryLog, DatabaseBackend backend, AbstractLoadBalancer loadBalancer, java.lang.String checkpointName)
Creates a new RecoverThread object

Method Summary

private void	addWorkerTask(BackendWorkerThread bwt, AbstractTask task) Add a task to a BackendWorkerThread using the proper synchronization.
void	endRecovery() Properly end the recovery and kill the worker thread used for recovery if it exists.
java.sql.SQLException	getException() Returns the exception value.
private long	recover(long logIdx) Replay the recovery log from the given logIdx index.
private void	recoveryFailed(java.sql.SQLException e) Unset the last known checkpoint and set the backend to disabled state.
void	run()
void	startRecovery() Start the recovery process by forking a BackendWorkerThread.

Methods inherited from class java.lang.Thread

activeCount, checkAccess, countStackFrames, currentThread, destroy, dumpStack, enumerate, getContextClassLoader, getName, getPriority, getThreadGroup, holdsLock, interrupt, interrupted, isAlive, isDaemon, isInterrupted, join, join, join, resume, setContextClassLoader, setDaemon, setName, setPriority, sleep, sleep, start, stop, stop, suspend, toString, yield

Methods inherited from class java.lang.Object

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

Field Detail

logger

static Trace logger

recoveryLog

private RecoveryLog recoveryLog

backend

private DatabaseBackend backend

loadBalancer

private AbstractLoadBalancer loadBalancer

exception

private java.sql.SQLException exception

bwt

private BackendWorkerThread bwt

tids

private java.util.ArrayList tids

scheduler

private AbstractScheduler scheduler

checkpointName

private java.lang.String checkpointName

recoveryBatchSize

private int recoveryBatchSize

Size of the pendingRecoveryTasks queue used during recovery

Constructor Detail

RecoverThread

public RecoverThread(AbstractScheduler scheduler,
                     RecoveryLog recoveryLog,
                     DatabaseBackend backend,
                     AbstractLoadBalancer loadBalancer,
                     java.lang.String checkpointName)

Creates a new RecoverThread object

Parameters:

scheduler - the currently used scheduler

recoveryLog - Recovery log that creates this thread

backend - database backend for logging

loadBalancer - index to start from for recovery

checkpointName - load balancer to use to create a BackendWorkerThread

Method Detail

getException

public java.sql.SQLException getException()

Returns the exception value.

Returns:

Returns the exception.

run

public void run()

See Also:

Runnable.run()

recoveryFailed

private void recoveryFailed(java.sql.SQLException e)

Unset the last known checkpoint and set the backend to disabled state. This should be called when the recovery has failed.

Parameters:

e - cause of the recovery failure

recover

private long recover(long logIdx)
              throws java.sql.SQLException

Replay the recovery log from the given logIdx index. Note that startRecovery() must have been called to fork and start the BackendWorkerThread before calling recover. endRecovery() must be called after recover() to terminate the thread.

Parameters:

logIdx - logIdx used to start the recovery

Returns:

last logIdx that was replayed.

Throws:

java.sql.SQLException - if fails

See Also:

startRecovery(), endRecovery()

addWorkerTask

private void addWorkerTask(BackendWorkerThread bwt,
                           AbstractTask task)

Add a task to a BackendWorkerThread using the proper synchronization.

Parameters:

bwt - BackendWorkerThread to synchronize on

task - the task to add to the thread queue

endRecovery

public void endRecovery()

Properly end the recovery and kill the worker thread used for recovery if it exists.

See Also:

startRecovery()

startRecovery

public void startRecovery()
                   throws java.sql.SQLException

Start the recovery process by forking a BackendWorkerThread. You must call endRecovery() to terminate the thread.

Throws:

java.sql.SQLException - if an error occurs

See Also:

endRecovery()