org.apache.hadoop.hbase.client

Interface AsyncTableBuilder<T extends AsyncTableBase>

@InterfaceAudience.Public
public interface AsyncTableBuilder<T extends AsyncTableBase>

For creating AsyncTable or RawAsyncTable.

The implementation should have default configurations set before returning the builder to user. So users are free to only set the configs they care about to create a new AsyncTable/RawAsyncTable instance.

Method Summary

Modifier and Type	Method and Description
T	build() Create the AsyncTable or RawAsyncTable instance.
AsyncTableBuilder<T>	setMaxAttempts(int maxAttempts) Set the max attempt times for an operation.
default AsyncTableBuilder<T>	setMaxRetries(int maxRetries) Set the max retry times for an operation.
AsyncTableBuilder<T>	setOperationTimeout(long timeout, TimeUnit unit) Set timeout for a whole operation such as get, put or delete.
AsyncTableBuilder<T>	setReadRpcTimeout(long timeout, TimeUnit unit) Set timeout for each read(get, scan) rpc request.
AsyncTableBuilder<T>	setRetryPause(long pause, TimeUnit unit) Set the base pause time for retrying.
AsyncTableBuilder<T>	setRpcTimeout(long timeout, TimeUnit unit) Set timeout for each rpc request.
AsyncTableBuilder<T>	setScanTimeout(long timeout, TimeUnit unit) As now we have heartbeat support for scan, ideally a scan will never timeout unless the RS is crash.
AsyncTableBuilder<T>	setStartLogErrorsCnt(int startLogErrorsCnt) Set the number of retries that are allowed before we start to log.
AsyncTableBuilder<T>	setWriteRpcTimeout(long timeout, TimeUnit unit) Set timeout for each write(put, delete) rpc request.

Method Detail

setOperationTimeout

AsyncTableBuilder<T> setOperationTimeout(long timeout,
                                         TimeUnit unit)

Set timeout for a whole operation such as get, put or delete. Notice that scan will not be effected by this value, see scanTimeoutNs.

Operation timeout and max attempt times(or max retry times) are both limitations for retrying, we will stop retrying when we reach any of the limitations.

See Also:

setMaxAttempts(int), setMaxRetries(int), setScanTimeout(long, TimeUnit)

setScanTimeout

AsyncTableBuilder<T> setScanTimeout(long timeout,
                                    TimeUnit unit)

As now we have heartbeat support for scan, ideally a scan will never timeout unless the RS is crash. The RS will always return something before the rpc timed out or scan timed out to tell the client that it is still alive. The scan timeout is used as operation timeout for every operation in a scan, such as openScanner or next.

See Also:

setScanTimeout(long, TimeUnit)

setRpcTimeout

AsyncTableBuilder<T> setRpcTimeout(long timeout,
                                   TimeUnit unit)

Set timeout for each rpc request.

Notice that this will NOT change the rpc timeout for read(get, scan) request and write request(put, delete).

setReadRpcTimeout

AsyncTableBuilder<T> setReadRpcTimeout(long timeout,
                                       TimeUnit unit)

Set timeout for each read(get, scan) rpc request.

setWriteRpcTimeout

AsyncTableBuilder<T> setWriteRpcTimeout(long timeout,
                                        TimeUnit unit)

Set timeout for each write(put, delete) rpc request.

setRetryPause

AsyncTableBuilder<T> setRetryPause(long pause,
                                   TimeUnit unit)

Set the base pause time for retrying. We use an exponential policy to generate sleep time when retrying.

setMaxRetries

default AsyncTableBuilder<T> setMaxRetries(int maxRetries)

Set the max retry times for an operation. Usually it is the max attempt times minus 1.

Operation timeout and max attempt times(or max retry times) are both limitations for retrying, we will stop retrying when we reach any of the limitations.

See Also:

setMaxAttempts(int), setOperationTimeout(long, TimeUnit)

setMaxAttempts

AsyncTableBuilder<T> setMaxAttempts(int maxAttempts)

Set the max attempt times for an operation. Usually it is the max retry times plus 1. Operation timeout and max attempt times(or max retry times) are both limitations for retrying, we will stop retrying when we reach any of the limitations.

See Also:

setMaxRetries(int), setOperationTimeout(long, TimeUnit)

setStartLogErrorsCnt

AsyncTableBuilder<T> setStartLogErrorsCnt(int startLogErrorsCnt)

Set the number of retries that are allowed before we start to log.

build

T build()

Create the AsyncTable or RawAsyncTable instance.