Package org.apache.flink.client.deployment.application

Class EmbeddedJobClient

java.lang.Object

org.apache.flink.client.deployment.application.EmbeddedJobClient

All Implemented Interfaces:

JobClient, CoordinationRequestGateway

@Internal
public class EmbeddedJobClient
extends Object
implements JobClient, CoordinationRequestGateway

A JobClient with the ability to also submit jobs which uses directly the DispatcherGateway.

Constructor Summary

Constructors

Constructor	Description
EmbeddedJobClient(JobID jobId, DispatcherGateway dispatcherGateway, ScheduledExecutor retryExecutor, Duration rpcTimeout, ClassLoader classLoader)

Method Summary

Modifier and Type	Method	Description
CompletableFuture<Void>	cancel()	Cancels the associated job.
CompletableFuture<Map<String,Object>>	getAccumulators()	Requests the accumulators of the associated job.
CompletableFuture<JobExecutionResult>	getJobExecutionResult()	Returns the result of the job execution of the submitted job.
JobID	getJobID()	Returns the JobID that uniquely identifies the job this client is scoped to.
CompletableFuture<JobStatus>	getJobStatus()	Requests the JobStatus of the associated job.
CompletableFuture<CoordinationResponse>	sendCoordinationRequest(String operatorUid, CoordinationRequest request)	Send out a request to a specified coordinator and return the response.
CompletableFuture<String>	stopWithSavepoint(boolean advanceToEndOfEventTime, String savepointDirectory, SavepointFormatType formatType)	Stops the associated job on Flink cluster.
CompletableFuture<String>	triggerSavepoint(String savepointDirectory, SavepointFormatType formatType)	Triggers a savepoint for the associated job.

Methods inherited from class java.lang.Object

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

Methods inherited from interface org.apache.flink.core.execution.JobClient

reportHeartbeat

Constructor Detail

EmbeddedJobClient

public EmbeddedJobClient(JobID jobId,
                         DispatcherGateway dispatcherGateway,
                         ScheduledExecutor retryExecutor,
                         Duration rpcTimeout,
                         ClassLoader classLoader)

Method Detail

getJobID

public JobID getJobID()

Description copied from interface: JobClient

Returns the JobID that uniquely identifies the job this client is scoped to.

Specified by:

getJobID in interface JobClient

getJobStatus

public CompletableFuture<JobStatus> getJobStatus()

Description copied from interface: JobClient

Requests the JobStatus of the associated job.

Specified by:

getJobStatus in interface JobClient

cancel

public CompletableFuture<Void> cancel()

Description copied from interface: JobClient

Cancels the associated job.

Specified by:

cancel in interface JobClient

stopWithSavepoint

public CompletableFuture<String> stopWithSavepoint(boolean advanceToEndOfEventTime,
                                                   @Nullable
                                                   String savepointDirectory,
                                                   SavepointFormatType formatType)

Description copied from interface: JobClient

Stops the associated job on Flink cluster.

Stopping works only for streaming programs. Be aware, that the job might continue to run for a while after sending the stop command, because after sources stopped to emit data all operators need to finish processing.

Specified by:

stopWithSavepoint in interface JobClient

Parameters:

advanceToEndOfEventTime - flag indicating if the source should inject a MAX_WATERMARK in the pipeline

savepointDirectory - directory the savepoint should be written to

formatType - binary format of the savepoint

Returns:

a CompletableFuture containing the path where the savepoint is located

triggerSavepoint

public CompletableFuture<String> triggerSavepoint(@Nullable
                                                  String savepointDirectory,
                                                  SavepointFormatType formatType)

Description copied from interface: JobClient

Triggers a savepoint for the associated job. The savepoint will be written to the given savepoint directory, or CheckpointingOptions.SAVEPOINT_DIRECTORY if it is null.

Specified by:

triggerSavepoint in interface JobClient

Parameters:

savepointDirectory - directory the savepoint should be written to

formatType - binary format of the savepoint

Returns:

a CompletableFuture containing the path where the savepoint is located

getAccumulators

public CompletableFuture<Map<String,Object>> getAccumulators()

Description copied from interface: JobClient

Requests the accumulators of the associated job. Accumulators can be requested while it is running or after it has finished. The class loader is used to deserialize the incoming accumulator results.

Specified by:

getAccumulators in interface JobClient

getJobExecutionResult

public CompletableFuture<JobExecutionResult> getJobExecutionResult()

Description copied from interface: JobClient

Returns the result of the job execution of the submitted job.

Specified by:

getJobExecutionResult in interface JobClient

sendCoordinationRequest

public CompletableFuture<CoordinationResponse> sendCoordinationRequest(String operatorUid,
                                                                       CoordinationRequest request)

Description copied from interface: CoordinationRequestGateway

Send out a request to a specified coordinator and return the response.

On the client side, a unique operatorUid must be defined to identify an operator. Otherwise, the query cannot be executed correctly. Note that we use operatorUid instead of operatorID because the latter is an internal runtime concept that cannot be recognized by the client.

Specified by:

sendCoordinationRequest in interface CoordinationRequestGateway

Parameters:

operatorUid - specifies which coordinator to receive the request

request - the request to send

Returns:

the response from the coordinator