org.apache.jackrabbit.core.gc

Class GarbageCollector

java.lang.Object

org.apache.jackrabbit.core.gc.GarbageCollector

All Implemented Interfaces:

DataStoreGarbageCollector

public class GarbageCollector
extends Object
implements DataStoreGarbageCollector

Garbage collector for DataStore. This implementation iterates through all nodes and reads the binary properties. To detect nodes that are moved while the scan runs, event listeners are started. Like the well known garbage collection in Java, the items that are still in use are marked. Currently this is achieved by updating the modified date of the entries. Newly added entries are detected because the modified date is changed when they are added.

Example code to run the data store garbage collection:

 JackrabbitRepositoryFactory jf = (JackrabbitRepositoryFactory) factory;
 RepositoryManager m = jf.getRepositoryManager((JackrabbitRepository) repository);
 GarbageCollector gc = m.createDataStoreGarbageCollector();
 try {
     gc.mark();
     gc.sweep();
 } finally {
     gc.close();
 }
 

Field Summary

Fields

Modifier and Type	Field and Description
protected int	testDelay

Constructor Summary

Constructors

Constructor and Description
GarbageCollector(RepositoryContext context, DataStore dataStore, IterablePersistenceManager[] list, SessionImpl[] sessionList) Create a new garbage collector.

Method Summary

Modifier and Type	Method and Description
void	close() Cleanup resources used internally by this instance.
protected void	finalize() Auto-close in case the application didn't call it explicitly.
int	getConcurrentThreadSize()
DataStore	getDataStore() Get the data store if one is used.
long	getMinSplitSize()
long	getSleepBetweenNodes() Get the delay between scanning items.
boolean	isPersistenceManagerScan() Check if using the IterablePersistenceManager interface is allowed.
void	mark() Scan the repository.
void	setConcurrentThreadSize(int concurrentThreadSize)
void	setMarkEventListener(MarkEventListener callback) Set the event listener.
void	setMinSplitSize(long minSplitSize)
void	setPersistenceManagerScan(boolean allow) Enable or disable using the IterablePersistenceManager interface to scan the items.
void	setSleepBetweenNodes(long millis) Set the delay between scanning items.
void	setTestDelay(int testDelay) When testing the garbage collection, a delay is used instead of simulating concurrent access.
void	stopScan() Reset modifiedDateOnAccess to 0 and stop the observation listener if any are installed.
int	sweep() Delete all unused items in the data store.

Methods inherited from class java.lang.Object

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

Field Detail

testDelay

protected int testDelay

Constructor Detail

GarbageCollector

public GarbageCollector(RepositoryContext context,
                        DataStore dataStore,
                        IterablePersistenceManager[] list,
                        SessionImpl[] sessionList)

Create a new garbage collector. This method is usually not called by the application, it is called by SessionImpl.createDataStoreGarbageCollector().

Parameters:

context - repository context

dataStore - the data store to be garbage-collected

list - the persistence managers

sessionList - the sessions to access the workspaces

Method Detail

setSleepBetweenNodes

public void setSleepBetweenNodes(long millis)

Description copied from interface: DataStoreGarbageCollector

Set the delay between scanning items. The main scan loop sleeps this many milliseconds after scanning a node. The default is 0, meaning the scan should run at full speed.

Specified by:

setSleepBetweenNodes in interface DataStoreGarbageCollector

Parameters:

millis - the number of milliseconds to sleep

getSleepBetweenNodes

public long getSleepBetweenNodes()

Description copied from interface: DataStoreGarbageCollector

Get the delay between scanning items.

Specified by:

getSleepBetweenNodes in interface DataStoreGarbageCollector

Returns:

the number of milliseconds to sleep

getMinSplitSize

public long getMinSplitSize()

setMinSplitSize

public void setMinSplitSize(long minSplitSize)

getConcurrentThreadSize

public int getConcurrentThreadSize()

setConcurrentThreadSize

public void setConcurrentThreadSize(int concurrentThreadSize)

setTestDelay

public void setTestDelay(int testDelay)

When testing the garbage collection, a delay is used instead of simulating concurrent access.

Parameters:

testDelay - the delay in milliseconds

setMarkEventListener

public void setMarkEventListener(MarkEventListener callback)

Description copied from interface: DataStoreGarbageCollector

Set the event listener. If set, the event listener will be called for each item that is scanned. This mechanism can be used to display the progress.

Specified by:

setMarkEventListener in interface DataStoreGarbageCollector

Parameters:

callback - if set, this is called while scanning

mark

public void mark()
          throws RepositoryException

Description copied from interface: DataStoreGarbageCollector

Scan the repository. The garbage collector will iterate over all nodes in the repository and update the last modified date. If all persistence managers implement the IterablePersistenceManager interface, this mechanism is used; if not, the garbage collector scans the repository using the JCR API starting from the root node.

Specified by:

mark in interface DataStoreGarbageCollector

Throws:

RepositoryException

setPersistenceManagerScan

public void setPersistenceManagerScan(boolean allow)

Description copied from interface: DataStoreGarbageCollector

Enable or disable using the IterablePersistenceManager interface to scan the items. This is important for clients that need the complete Node implementation in the ScanEventListener callback.

Specified by:

setPersistenceManagerScan in interface DataStoreGarbageCollector

Parameters:

allow - true if using the IterablePersistenceManager interface is allowed

isPersistenceManagerScan

public boolean isPersistenceManagerScan()

Description copied from interface: DataStoreGarbageCollector

Check if using the IterablePersistenceManager interface is allowed.

Specified by:

isPersistenceManagerScan in interface DataStoreGarbageCollector

Returns:

true if using IterablePersistenceManager is possible.

stopScan

public void stopScan()
              throws RepositoryException

Reset modifiedDateOnAccess to 0 and stop the observation listener if any are installed.

Throws:

RepositoryException

sweep

public int sweep()
          throws RepositoryException

Description copied from interface: DataStoreGarbageCollector

Delete all unused items in the data store.

Specified by:

sweep in interface DataStoreGarbageCollector

Returns:

the number of deleted items

Throws:

RepositoryException

getDataStore

public DataStore getDataStore()

Get the data store if one is used.

Returns:

the data store, or null

close

public void close()

Description copied from interface: DataStoreGarbageCollector

Cleanup resources used internally by this instance.

Specified by:

close in interface DataStoreGarbageCollector

finalize

protected void finalize()
                 throws Throwable

Auto-close in case the application didn't call it explicitly.

Overrides:

finalize in class Object

Throws:

Throwable