Package org.apache.hadoop.hbase.regionserver

Class HStore.StoreFlusherImpl

java.lang.Object

org.apache.hadoop.hbase.regionserver.HStore.StoreFlusherImpl

All Implemented Interfaces:

StoreFlushContext

Enclosing class:

HStore

private final class HStore.StoreFlusherImpl
extends Object
implements StoreFlushContext

Field Summary

Fields

Modifier and Type	Field	Description
private long	cacheFlushCount
private final long	cacheFlushSeqNum
private long	cacheFlushSize
private List<org.apache.hadoop.fs.Path>	committedFiles
private long	outputFileSize
private MemStoreSnapshot	snapshot
private List<org.apache.hadoop.fs.Path>	tempFiles
private final FlushLifeCycleTracker	tracker
private final HStore.StoreFileWriterCreationTracker	writerCreationTracker

Constructor Summary

Constructors

Modifier	Constructor	Description
private	StoreFlusherImpl(long cacheFlushSeqNum, FlushLifeCycleTracker tracker)

Method Summary

Modifier and Type	Method	Description
void	abort()	Abort the snapshot preparation.
boolean	commit(MonitoredTask status)	Commit the flush - add the store file to the store and clear the memstore snapshot.
void	flushCache(MonitoredTask status)	Flush the cache (create the new store file) A length operation which doesn't require locking out any function of the store.
List<org.apache.hadoop.fs.Path>	getCommittedFiles()	Returns the newly committed files from the flush.
long	getOutputFileSize()	Returns the total file size for flush output files, in bytes
MemStoreSize	prepare()	This is not thread safe.
void	replayFlush(List<String> fileNames, boolean dropMemstoreSnapshot)	Similar to commit, but called in secondary region replicas for replaying the flush cache from primary region.

Methods inherited from class java.lang.Object

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

Field Details

tracker

private final FlushLifeCycleTracker tracker

writerCreationTracker

private final HStore.StoreFileWriterCreationTracker writerCreationTracker

cacheFlushSeqNum

private final long cacheFlushSeqNum

snapshot

private MemStoreSnapshot snapshot

tempFiles

private List<org.apache.hadoop.fs.Path> tempFiles

committedFiles

private List<org.apache.hadoop.fs.Path> committedFiles

cacheFlushCount

private long cacheFlushCount

cacheFlushSize

private long cacheFlushSize

outputFileSize

private long outputFileSize

Constructor Details

StoreFlusherImpl

private StoreFlusherImpl(long cacheFlushSeqNum,
 FlushLifeCycleTracker tracker)

Method Details

prepare

public MemStoreSize prepare()

This is not thread safe. The caller should have a lock on the region or the store. If necessary, the lock can be added with the patch provided in HBASE-10087

Specified by:

prepare in interface StoreFlushContext

Returns:

The size of snapshot to flush

flushCache

public void flushCache(MonitoredTask status)
                throws IOException

Description copied from interface: StoreFlushContext

Flush the cache (create the new store file) A length operation which doesn't require locking out any function of the store.

Specified by:

flushCache in interface StoreFlushContext

Throws:

IOException - in case the flush fails

commit

public boolean commit(MonitoredTask status)
               throws IOException

Description copied from interface: StoreFlushContext

Commit the flush - add the store file to the store and clear the memstore snapshot. Requires pausing scans. A very short operation

Specified by:

commit in interface StoreFlushContext

Returns:

whether compaction is required

Throws:

IOException

getOutputFileSize

public long getOutputFileSize()

Description copied from interface: StoreFlushContext

Returns the total file size for flush output files, in bytes

Specified by:

getOutputFileSize in interface StoreFlushContext

getCommittedFiles

public List<org.apache.hadoop.fs.Path> getCommittedFiles()

Description copied from interface: StoreFlushContext

Returns the newly committed files from the flush. Called only if commit returns true

Specified by:

getCommittedFiles in interface StoreFlushContext

Returns:

a list of Paths for new files

replayFlush

public void replayFlush(List<String> fileNames,
 boolean dropMemstoreSnapshot)
                 throws IOException

Similar to commit, but called in secondary region replicas for replaying the flush cache from primary region. Adds the new files to the store, and drops the snapshot depending on dropMemstoreSnapshot argument.

Specified by:

replayFlush in interface StoreFlushContext

Parameters:

fileNames - names of the flushed files

dropMemstoreSnapshot - whether to drop the prepared memstore snapshot

Throws:

IOException

abort

public void abort()
           throws IOException

Abort the snapshot preparation. Drops the snapshot if any.

Specified by:

abort in interface StoreFlushContext

Throws:

IOException