Package org.apache.hadoop.hbase.tool

Class WriteSinkCoprocessor

java.lang.Object

org.apache.hadoop.hbase.tool.WriteSinkCoprocessor

All Implemented Interfaces:

Coprocessor, RegionCoprocessor, RegionObserver

@Private
public class WriteSinkCoprocessor
extends Object
implements RegionCoprocessor, RegionObserver

This coprocessor 'shallows' all the writes. It allows to test a pure write workload, going through all the communication layers. The reads will work as well, but they as we never write, they will always always return an empty structure. The WAL is also skipped. Obviously, the region will never be split automatically. It's up to the user to split and move it.

For a table created like this: create 'usertable', {NAME => 'f1', VERSIONS => 1}

You can then add the coprocessor with this command: alter 'usertable', 'coprocessor' => '|org.apache.hadoop.hbase.tool.WriteSinkCoprocessor|'

And then put 'usertable', 'f1', 'f1', 'f1'

scan 'usertable' Will return: 0 row(s) in 0.0050 seconds

TODO: It needs tests

Nested Class Summary

Nested classes/interfaces inherited from interface org.apache.hadoop.hbase.Coprocessor

Coprocessor.State

Nested classes/interfaces inherited from interface org.apache.hadoop.hbase.coprocessor.RegionObserver

RegionObserver.MutationType

Field Summary

Fields

Modifier and Type	Field	Description
private static final org.slf4j.Logger	LOG
private final AtomicLong	ops
private String	regionName

Fields inherited from interface org.apache.hadoop.hbase.Coprocessor

PRIORITY_HIGHEST, PRIORITY_LOWEST, PRIORITY_SYSTEM, PRIORITY_USER, VERSION

Constructor Summary

Constructors

Constructor	Description
WriteSinkCoprocessor()

Method Summary

Modifier and Type	Method	Description
Optional<RegionObserver>	getRegionObserver()
void	preBatchMutate(ObserverContext<? extends RegionCoprocessorEnvironment> c, MiniBatchOperationInProgress<Mutation> miniBatchOp)	This will be called for every batch mutation operation happening at the server.
void	preOpen(ObserverContext<? extends RegionCoprocessorEnvironment> e)	Called before the region is reported as open to the master.

Methods inherited from class java.lang.Object

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

Methods inherited from interface org.apache.hadoop.hbase.Coprocessor

getServices, start, stop

Methods inherited from interface org.apache.hadoop.hbase.coprocessor.RegionCoprocessor

getBulkLoadObserver, getEndpointObserver

Methods inherited from interface org.apache.hadoop.hbase.coprocessor.RegionObserver

postAppend, postAppend, postAppendBeforeWAL, postBatchMutate, postBatchMutateIndispensably, postBulkLoadHFile, postCheckAndDelete, postCheckAndDelete, postCheckAndMutate, postCheckAndPut, postCheckAndPut, postClose, postCloseRegionOperation, postCommitStoreFile, postCompact, postCompactSelection, postDelete, postDelete, postExists, postFlush, postFlush, postGetOp, postIncrement, postIncrement, postIncrementBeforeWAL, postInstantiateDeleteTracker, postMemStoreCompaction, postMutationBeforeWAL, postOpen, postPut, postPut, postReplayWALs, postScannerClose, postScannerFilterRow, postScannerNext, postScannerOpen, postStartRegionOperation, postStoreFileReaderOpen, preAppend, preAppend, preAppendAfterRowLock, preBulkLoadHFile, preCheckAndDelete, preCheckAndDelete, preCheckAndDeleteAfterRowLock, preCheckAndDeleteAfterRowLock, preCheckAndMutate, preCheckAndMutateAfterRowLock, preCheckAndPut, preCheckAndPut, preCheckAndPutAfterRowLock, preCheckAndPutAfterRowLock, preClose, preCommitStoreFile, preCompact, preCompactScannerOpen, preCompactSelection, preDelete, preDelete, preExists, preFlush, preFlush, preFlushScannerOpen, preGetOp, preIncrement, preIncrement, preIncrementAfterRowLock, preMemStoreCompaction, preMemStoreCompactionCompact, preMemStoreCompactionCompactScannerOpen, prePrepareTimeStampForDeleteVersion, prePut, prePut, preReplayWALs, preScannerClose, preScannerNext, preScannerOpen, preStoreFileReaderOpen, preStoreScannerOpen, preWALAppend

Field Details

LOG

private static final org.slf4j.Logger LOG

ops

private final AtomicLong ops

regionName

private String regionName

Constructor Details

WriteSinkCoprocessor

public WriteSinkCoprocessor()

Method Details

getRegionObserver

public Optional<RegionObserver> getRegionObserver()

Specified by:

getRegionObserver in interface RegionCoprocessor

preOpen

public void preOpen(ObserverContext<? extends RegionCoprocessorEnvironment> e)
             throws IOException

Description copied from interface: RegionObserver

Called before the region is reported as open to the master.

Specified by:

preOpen in interface RegionObserver

Parameters:

e - the environment provided by the region server

Throws:

IOException

preBatchMutate

public void preBatchMutate(ObserverContext<? extends RegionCoprocessorEnvironment> c,
 MiniBatchOperationInProgress<Mutation> miniBatchOp)
                    throws IOException

Description copied from interface: RegionObserver

This will be called for every batch mutation operation happening at the server. This will be called after acquiring the locks on the mutating rows and after applying the proper timestamp for each Mutation at the server. The batch may contain Put/Delete/Increment/Append. By setting OperationStatus of Mutations (MiniBatchOperationInProgress.setOperationStatus(int, OperationStatus)), RegionObserver can make Region to skip these Mutations.

Note: Do not retain references to any Cells in Mutations beyond the life of this invocation. If need a Cell reference for later use, copy the cell and use that.

Specified by:

preBatchMutate in interface RegionObserver

Parameters:

c - the environment provided by the region server

miniBatchOp - batch of Mutations getting applied to region.

Throws:

IOException