Package oracle.kv

Class BulkWriteOptions

java.lang.Object

oracle.kv.table.WriteOptions

oracle.kv.BulkWriteOptions

All Implemented Interfaces:

Cloneable

public class BulkWriteOptions
extends WriteOptions

BulkWriteOptions is used to configure bulk write operations. The default values, documented in the setter methods, should be a good choice for a wide range of conditions. If you should wish to fine tune the bulk load operation further, you can use these values as a starting point for a benchmark using your own application data and actual hardware.

Since:

4.0

Constructor Summary

Constructors

Constructor	Description
BulkWriteOptions()	Create a BulkWriteOptions with default values.
BulkWriteOptions(BulkWriteOptions options)
BulkWriteOptions(Durability durability, long timeout, TimeUnit timeoutUnit)	The options used to configure the bulk put operation.

Method Summary

Modifier and Type	Method	Description
int	getBulkHeapPercent()	Returns the percentage of Runtime.maxMemory() that can be used for the operation.
int	getMaxRequestSize()	Returns the max number of bytes of records in a single bulk put request.
int	getPerShardParallelism()	The maximum number of threads that can concurrently write a batch of entries to a single shard in the store.
int	getStreamParallelism()	Returns the maximum number of streams that can be read concurrently.
void	setBulkHeapPercent(int bulkHeapPercent)	The percentage of Runtime.maxMemory() that can be used for the operation.
void	setMaxRequestSize(int maxRequestSize)	The max request size is used to limit the total number of bytes of records in a single bulk put request.
void	setPerShardParallelism(int perShardParallelism)	Sets the maximum number of threads that can concurrently write it's batch of entries to a single shard in the store.
void	setStreamParallelism(int streamParallelism)	Sets the maximum number of streams that can be read concurrently.

Methods inherited from class oracle.kv.table.WriteOptions

clone, getDurability, getIdentityCacheSize, getTimeout, getTimeoutUnit, getUpdateTTL, setDurability, setIdentityCacheSize, setTimeout, setUpdateTTL, toString

Methods inherited from class java.lang.Object

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

Constructor Details

BulkWriteOptions

public BulkWriteOptions(Durability durability,
 long timeout,
 TimeUnit timeoutUnit)

The options used to configure the bulk put operation.

Parameters:

durability - the durability to be used by the underlying write operations that make up the bulk put.

timeout - the timeout associated with the underlying write operations that make up the bulk put.

timeoutUnit - the units associated with the timeout

BulkWriteOptions

public BulkWriteOptions(BulkWriteOptions options)

BulkWriteOptions

public BulkWriteOptions()

Create a BulkWriteOptions with default values.

Method Details

getBulkHeapPercent

public int getBulkHeapPercent()

Returns the percentage of Runtime.maxMemory() that can be used for the operation.

setBulkHeapPercent

public void setBulkHeapPercent(int bulkHeapPercent)

The percentage of Runtime.maxMemory() that can be used for the operation. This heap is used to assemble batches of entries associated with specific shards and partitions.

The default is 40%.

getMaxRequestSize

public int getMaxRequestSize()

Returns the max number of bytes of records in a single bulk put request.

setMaxRequestSize

public void setMaxRequestSize(int maxRequestSize)

The max request size is used to limit the total number of bytes of records in a single bulk put request.

The default is 512K.

getPerShardParallelism

public int getPerShardParallelism()

The maximum number of threads that can concurrently write a batch of entries to a single shard in the store.

setPerShardParallelism

public void setPerShardParallelism(int perShardParallelism)

Sets the maximum number of threads that can concurrently write it's batch of entries to a single shard in the store.

The default value is 3 and allows for overlapping the reading of the next batch with processing of the current batch at a server node. Higher capacity networks and and storage nodes can allow for higher parallelism.

getStreamParallelism

public int getStreamParallelism()

Returns the maximum number of streams that can be read concurrently. Each stream is read by a dedicated thread from a thread pool. This setting determines the size of the thread pool used for reading streams.

setStreamParallelism

public void setStreamParallelism(int streamParallelism)

Sets the maximum number of streams that can be read concurrently. Each stream is read by a dedicated thread from a thread pool. This setting determines the size of the thread pool used for reading streams.

The default parallelism is 1. For streams with high overheads, say because the I/O device underlying the stream is slow and there are different I/O devices underlying each stream, a higher value would be appropriate.