Package org.apache.spark.sql.pipelines.graph

Class GraphExecution

Object

org.apache.spark.sql.pipelines.graph.GraphExecution

All Implemented Interfaces:

org.apache.spark.internal.Logging

Direct Known Subclasses:

TriggeredGraphExecution

public abstract class GraphExecution
extends Object
implements org.apache.spark.internal.Logging

Nested Class Summary

Nested Classes

Modifier and Type	Class	Description
static interface	GraphExecution.FlowExecutionAction
static interface	GraphExecution.FlowExecutionStopReason	Represents the reason why a flow execution should be stopped.
static class	GraphExecution.RetryFlowExecution$	Indicates that the flow execution should be retried.
static class	GraphExecution.StopFlowExecution	Indicates that the flow execution should be stopped with a specific reason.
static class	GraphExecution.StopFlowExecution$

Nested classes/interfaces inherited from interface org.apache.spark.internal.Logging

org.apache.spark.internal.Logging.LogStringContext, org.apache.spark.internal.Logging.SparkShellLoggingFilter

Constructor Summary

Constructors

Constructor	Description
GraphExecution(DataflowGraph graphForExecution, PipelineUpdateContext env)

Method Summary

Modifier and Type	Method	Description
abstract void	awaitCompletion()	Blocks the current thread while any flows are queued or running.
static GraphExecution.FlowExecutionAction	determineFlowExecutionActionFromError(scala.Function0<Throwable> ex, scala.Function0<String> flowDisplayName, scala.Function0<Object> currentNumTries, scala.Function0<Object> maxAllowedRetries)	Analyze the exception thrown by flow execution and figure out if we should retry the execution, or we need to reanalyze the flow entirely to resolve issues like schema changes.
scala.collection.concurrent.TrieMap<org.apache.spark.sql.catalyst.TableIdentifier,FlowExecution>	flowExecutions()	FlowExecutions currently being executed and tracked by the graph execution.
abstract RunTerminationReason	getRunTerminationReason()	Returns the reason why this flow execution has terminated.
DataflowGraph	graphForExecution()
static org.apache.spark.internal.Logging.LogStringContext	LogStringContext(scala.StringContext sc)
int	maxRetryAttemptsForFlow(org.apache.spark.sql.catalyst.TableIdentifier flowName)
static org.slf4j.Logger	org$apache$spark$internal$Logging$$log_()
static void	org$apache$spark$internal$Logging$$log__$eq(org.slf4j.Logger x$1)
scala.Option<FlowExecution>	planAndStartFlow(ResolvedFlow flow)	Plans the logical ResolvedFlow into a FlowExecution and then starts executing it.
void	start()	Starts the execution of flows in graphForExecution.
void	stop()	Stops this execution by stopping all streams and terminating any other resources.
void	stopFlow(FlowExecution pf)	Stops execution of a `FlowExecution`.
void	stopThread(Thread thread)	Stop a thread timeout.
abstract Trigger	streamTrigger(Flow flow)	The `Trigger` configuration for a streaming flow.

Methods inherited from class java.lang.Object

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

Methods inherited from interface org.apache.spark.internal.Logging

initializeForcefully, initializeLogIfNecessary, initializeLogIfNecessary, initializeLogIfNecessary$default$2, isTraceEnabled, log, logBasedOnLevel, logDebug, logDebug, logDebug, logDebug, logError, logError, logError, logError, logInfo, logInfo, logInfo, logInfo, logName, LogStringContext, logTrace, logTrace, logTrace, logTrace, logWarning, logWarning, logWarning, logWarning, org$apache$spark$internal$Logging$$log_, org$apache$spark$internal$Logging$$log__$eq, withLogContext

Constructor Details

GraphExecution

public GraphExecution(DataflowGraph graphForExecution,
 PipelineUpdateContext env)

Method Details

determineFlowExecutionActionFromError

public static GraphExecution.FlowExecutionAction determineFlowExecutionActionFromError(scala.Function0<Throwable> ex,
 scala.Function0<String> flowDisplayName,
 scala.Function0<Object> currentNumTries,
 scala.Function0<Object> maxAllowedRetries)

Analyze the exception thrown by flow execution and figure out if we should retry the execution, or we need to reanalyze the flow entirely to resolve issues like schema changes. This should be the narrow waist for all exception analysis in flow execution. TODO: currently it only handles schema change and max retries, we should aim to extend this to include other non-retryable exception as well so we can have a single SoT for all these error matching logic.

Parameters:

ex - Exception to analyze.

flowDisplayName - The user facing flow name with the error.

currentNumTries - Number of times the flow has been tried.

maxAllowedRetries - Maximum number of retries allowed for the flow.

Returns:

(undocumented)

org$apache$spark$internal$Logging$$log_

public static org.slf4j.Logger org$apache$spark$internal$Logging$$log_()

org$apache$spark$internal$Logging$$log__$eq

public static void org$apache$spark$internal$Logging$$log__$eq(org.slf4j.Logger x$1)

LogStringContext

public static org.apache.spark.internal.Logging.LogStringContext LogStringContext(scala.StringContext sc)

graphForExecution

public DataflowGraph graphForExecution()

streamTrigger

public abstract Trigger streamTrigger(Flow flow)

The `Trigger` configuration for a streaming flow.

flowExecutions

public scala.collection.concurrent.TrieMap<org.apache.spark.sql.catalyst.TableIdentifier,FlowExecution> flowExecutions()

FlowExecutions currently being executed and tracked by the graph execution.

Returns:

(undocumented)

planAndStartFlow

public scala.Option<FlowExecution> planAndStartFlow(ResolvedFlow flow)

Plans the logical ResolvedFlow into a FlowExecution and then starts executing it. Implementation note: Thread safe

Parameters:

flow - (undocumented)

Returns:

None if the flow planner decided that there is no actual update required here. Otherwise returns the corresponding physical flow.

start

public void start()

Starts the execution of flows in graphForExecution. Does not block.

stop

public void stop()

Stops this execution by stopping all streams and terminating any other resources.

This method may be called multiple times due to race conditions and must be idempotent.

stopFlow

public void stopFlow(FlowExecution pf)

Stops execution of a `FlowExecution`.

awaitCompletion

public abstract void awaitCompletion()

Blocks the current thread while any flows are queued or running. Returns when all flows that could be run have completed. When this returns, all flows are either SUCCESSFUL, TERMINATED_WITH_ERROR, SKIPPED, CANCELED, or EXCLUDED.

getRunTerminationReason

public abstract RunTerminationReason getRunTerminationReason()

Returns the reason why this flow execution has terminated. If the function is called before the flow has not terminated yet, the behavior is undefined, and may return UnexpectedRunFailure.

Returns:

(undocumented)

maxRetryAttemptsForFlow

public int maxRetryAttemptsForFlow(org.apache.spark.sql.catalyst.TableIdentifier flowName)

stopThread

public void stopThread(Thread thread)

Stop a thread timeout.

Parameters:

thread - (undocumented)