org.apache.flink.runtime.jobgraph.tasks.CheckpointCoordinatorConfiguration Java Examples
The following examples show how to use
org.apache.flink.runtime.jobgraph.tasks.CheckpointCoordinatorConfiguration.
You can vote up the ones you like or vote down the ones you don't like,
and go to the original project or source file by following the links above each example. You may check out the related API usage on the sidebar.
Example #1
Source Project: Flink-CEPplus Author: ljygz File: JobMasterTest.java License: Apache License 2.0 | 6 votes |
@Nonnull private JobGraph createJobGraphFromJobVerticesWithCheckpointing(SavepointRestoreSettings savepointRestoreSettings, JobVertex... jobVertices) { final JobGraph jobGraph = new JobGraph(jobVertices); // enable checkpointing which is required to resume from a savepoint final CheckpointCoordinatorConfiguration checkpoinCoordinatorConfiguration = new CheckpointCoordinatorConfiguration( 1000L, 1000L, 1000L, 1, CheckpointRetentionPolicy.NEVER_RETAIN_AFTER_TERMINATION, true); final JobCheckpointingSettings checkpointingSettings = new JobCheckpointingSettings( Collections.emptyList(), Collections.emptyList(), Collections.emptyList(), checkpoinCoordinatorConfiguration, null); jobGraph.setSnapshotSettings(checkpointingSettings); jobGraph.setSavepointRestoreSettings(savepointRestoreSettings); return jobGraph; }
Example #2
Source Project: flink Author: flink-tpc-ds File: CheckpointConfigHandler.java License: Apache License 2.0 | 6 votes |
private static CheckpointConfigInfo createCheckpointConfigInfo(AccessExecutionGraph executionGraph) throws RestHandlerException { final CheckpointCoordinatorConfiguration checkpointCoordinatorConfiguration = executionGraph.getCheckpointCoordinatorConfiguration(); if (checkpointCoordinatorConfiguration == null) { throw new RestHandlerException( "Checkpointing is not enabled for this job (" + executionGraph.getJobID() + ").", HttpResponseStatus.NOT_FOUND); } else { CheckpointRetentionPolicy retentionPolicy = checkpointCoordinatorConfiguration.getCheckpointRetentionPolicy(); CheckpointConfigInfo.ExternalizedCheckpointInfo externalizedCheckpointInfo = new CheckpointConfigInfo.ExternalizedCheckpointInfo( retentionPolicy != CheckpointRetentionPolicy.NEVER_RETAIN_AFTER_TERMINATION, retentionPolicy != CheckpointRetentionPolicy.RETAIN_ON_CANCELLATION); return new CheckpointConfigInfo( checkpointCoordinatorConfiguration.isExactlyOnce() ? CheckpointConfigInfo.ProcessingMode.EXACTLY_ONCE : CheckpointConfigInfo.ProcessingMode.AT_LEAST_ONCE, checkpointCoordinatorConfiguration.getCheckpointInterval(), checkpointCoordinatorConfiguration.getCheckpointTimeout(), checkpointCoordinatorConfiguration.getMinPauseBetweenCheckpoints(), checkpointCoordinatorConfiguration.getMaxConcurrentCheckpoints(), externalizedCheckpointInfo); } }
Example #3
Source Project: flink Author: apache File: JobMasterTest.java License: Apache License 2.0 | 6 votes |
@Nonnull private JobGraph createJobGraphFromJobVerticesWithCheckpointing(SavepointRestoreSettings savepointRestoreSettings, JobVertex... jobVertices) { final JobGraph jobGraph = new JobGraph(jobVertices); // enable checkpointing which is required to resume from a savepoint final CheckpointCoordinatorConfiguration checkpoinCoordinatorConfiguration = new CheckpointCoordinatorConfiguration( 1000L, 1000L, 1000L, 1, CheckpointRetentionPolicy.NEVER_RETAIN_AFTER_TERMINATION, true, false, false, 0); final JobCheckpointingSettings checkpointingSettings = new JobCheckpointingSettings( Collections.emptyList(), Collections.emptyList(), Collections.emptyList(), checkpoinCoordinatorConfiguration, null); jobGraph.setSnapshotSettings(checkpointingSettings); jobGraph.setSavepointRestoreSettings(savepointRestoreSettings); return jobGraph; }
Example #4
Source Project: flink Author: flink-tpc-ds File: CheckpointCoordinatorMasterHooksTest.java License: Apache License 2.0 | 6 votes |
private static CheckpointCoordinator instantiateCheckpointCoordinator(JobID jid, ExecutionVertex... ackVertices) { CheckpointCoordinatorConfiguration chkConfig = new CheckpointCoordinatorConfiguration( 10000000L, 600000L, 0L, 1, CheckpointRetentionPolicy.NEVER_RETAIN_AFTER_TERMINATION, true, false, 0); return new CheckpointCoordinator( jid, chkConfig, new ExecutionVertex[0], ackVertices, new ExecutionVertex[0], new StandaloneCheckpointIDCounter(), new StandaloneCompletedCheckpointStore(10), new MemoryStateBackend(), Executors.directExecutor(), SharedStateRegistry.DEFAULT_FACTORY, new CheckpointFailureManager( 0, NoOpFailJobCall.INSTANCE)); }
Example #5
Source Project: flink Author: apache File: JobGraphTest.java License: Apache License 2.0 | 6 votes |
private static JobCheckpointingSettings createCheckpointSettingsWithInterval(final long checkpointInterval) { final CheckpointCoordinatorConfiguration checkpointCoordinatorConfiguration = new CheckpointCoordinatorConfiguration( checkpointInterval, Long.MAX_VALUE, Long.MAX_VALUE, Integer.MAX_VALUE, CheckpointRetentionPolicy.NEVER_RETAIN_AFTER_TERMINATION, true, false, false, 0); return new JobCheckpointingSettings( Collections.emptyList(), Collections.emptyList(), Collections.emptyList(), checkpointCoordinatorConfiguration, null); }
Example #6
Source Project: flink Author: flink-tpc-ds File: FailoverRegionTest.java License: Apache License 2.0 | 6 votes |
private static void enableCheckpointing(ExecutionGraph eg) { ArrayList<ExecutionJobVertex> jobVertices = new ArrayList<>(eg.getAllVertices().values()); CheckpointCoordinatorConfiguration chkConfig = new CheckpointCoordinatorConfiguration( 1000, 100, 0, 1, CheckpointRetentionPolicy.RETAIN_ON_CANCELLATION, true, false, 0); eg.enableCheckpointing( chkConfig, jobVertices, jobVertices, jobVertices, Collections.emptyList(), new StandaloneCheckpointIDCounter(), new StandaloneCompletedCheckpointStore(1), new MemoryStateBackend(), new CheckpointStatsTracker( 0, jobVertices, mock(CheckpointCoordinatorConfiguration.class), new UnregisteredMetricsGroup())); }
Example #7
Source Project: flink Author: apache File: CheckpointConfigHandler.java License: Apache License 2.0 | 6 votes |
private static CheckpointConfigInfo createCheckpointConfigInfo(AccessExecutionGraph executionGraph) throws RestHandlerException { final CheckpointCoordinatorConfiguration checkpointCoordinatorConfiguration = executionGraph.getCheckpointCoordinatorConfiguration(); if (checkpointCoordinatorConfiguration == null) { throw new RestHandlerException( "Checkpointing is not enabled for this job (" + executionGraph.getJobID() + ").", HttpResponseStatus.NOT_FOUND); } else { CheckpointRetentionPolicy retentionPolicy = checkpointCoordinatorConfiguration.getCheckpointRetentionPolicy(); CheckpointConfigInfo.ExternalizedCheckpointInfo externalizedCheckpointInfo = new CheckpointConfigInfo.ExternalizedCheckpointInfo( retentionPolicy != CheckpointRetentionPolicy.NEVER_RETAIN_AFTER_TERMINATION, retentionPolicy != CheckpointRetentionPolicy.RETAIN_ON_CANCELLATION); String stateBackendName = executionGraph.getStateBackendName().orElse(null); return new CheckpointConfigInfo( checkpointCoordinatorConfiguration.isExactlyOnce() ? CheckpointConfigInfo.ProcessingMode.EXACTLY_ONCE : CheckpointConfigInfo.ProcessingMode.AT_LEAST_ONCE, checkpointCoordinatorConfiguration.getCheckpointInterval(), checkpointCoordinatorConfiguration.getCheckpointTimeout(), checkpointCoordinatorConfiguration.getMinPauseBetweenCheckpoints(), checkpointCoordinatorConfiguration.getMaxConcurrentCheckpoints(), externalizedCheckpointInfo, stateBackendName); } }
Example #8
Source Project: flink Author: apache File: CheckpointCoordinatorTest.java License: Apache License 2.0 | 6 votes |
/** * Tests that no minimum delay between savepoints is enforced. */ @Test public void testMinDelayBetweenSavepoints() throws Exception { CheckpointCoordinatorConfiguration chkConfig = new CheckpointCoordinatorConfigurationBuilder() .setMinPauseBetweenCheckpoints(100000000L) // very long min delay => should not affect savepoints .setMaxConcurrentCheckpoints(1) .build(); CheckpointCoordinator coord = new CheckpointCoordinatorBuilder() .setCheckpointCoordinatorConfiguration(chkConfig) .setCompletedCheckpointStore(new StandaloneCompletedCheckpointStore(2)) .setTimer(manuallyTriggeredScheduledExecutor) .build(); String savepointDir = tmpFolder.newFolder().getAbsolutePath(); CompletableFuture<CompletedCheckpoint> savepoint0 = coord.triggerSavepoint(savepointDir); assertFalse("Did not trigger savepoint", savepoint0.isDone()); CompletableFuture<CompletedCheckpoint> savepoint1 = coord.triggerSavepoint(savepointDir); assertFalse("Did not trigger savepoint", savepoint1.isDone()); }
Example #9
Source Project: Flink-CEPplus Author: ljygz File: CheckpointStatsTracker.java License: Apache License 2.0 | 5 votes |
/** * Creates a new checkpoint stats tracker. * * @param numRememberedCheckpoints Maximum number of checkpoints to remember, including in progress ones. * @param jobVertices Job vertices involved in the checkpoints. * @param jobCheckpointingConfiguration Checkpointing configuration. * @param metricGroup Metric group for exposed metrics */ public CheckpointStatsTracker( int numRememberedCheckpoints, List<ExecutionJobVertex> jobVertices, CheckpointCoordinatorConfiguration jobCheckpointingConfiguration, MetricGroup metricGroup) { checkArgument(numRememberedCheckpoints >= 0, "Negative number of remembered checkpoints"); this.history = new CheckpointStatsHistory(numRememberedCheckpoints); this.jobVertices = checkNotNull(jobVertices, "JobVertices"); this.jobCheckpointingConfiguration = checkNotNull(jobCheckpointingConfiguration); // Compute the total subtask count. We do this here in order to only // do it once. int count = 0; for (ExecutionJobVertex vertex : jobVertices) { count += vertex.getParallelism(); } this.totalSubtaskCount = count; // Latest snapshot is empty latestSnapshot = new CheckpointStatsSnapshot( counts.createSnapshot(), summary.createSnapshot(), history.createSnapshot(), null); // Register the metrics registerMetrics(metricGroup); }
Example #10
Source Project: flink Author: apache File: CheckpointCoordinatorMasterHooksTest.java License: Apache License 2.0 | 5 votes |
private CheckpointCoordinator instantiateCheckpointCoordinator( JobID jid, ScheduledExecutor testingScheduledExecutor, ExecutionVertex... ackVertices) { CheckpointCoordinatorConfiguration chkConfig = new CheckpointCoordinatorConfiguration( 10000000L, 600000L, 0L, 1, CheckpointRetentionPolicy.NEVER_RETAIN_AFTER_TERMINATION, true, false, false, 0); return new CheckpointCoordinator( jid, chkConfig, new ExecutionVertex[0], ackVertices, new ExecutionVertex[0], Collections.emptyList(), new StandaloneCheckpointIDCounter(), new StandaloneCompletedCheckpointStore(10), new MemoryStateBackend(), Executors.directExecutor(), testingScheduledExecutor, SharedStateRegistry.DEFAULT_FACTORY, new CheckpointFailureManager( 0, NoOpFailJobCall.INSTANCE)); }
Example #11
Source Project: Flink-CEPplus Author: ljygz File: ExecutionGraph.java License: Apache License 2.0 | 5 votes |
@Override public CheckpointCoordinatorConfiguration getCheckpointCoordinatorConfiguration() { if (checkpointStatsTracker != null) { return checkpointStatsTracker.getJobCheckpointingConfiguration(); } else { return null; } }
Example #12
Source Project: Flink-CEPplus Author: ljygz File: ArchivedExecutionGraph.java License: Apache License 2.0 | 5 votes |
public ArchivedExecutionGraph( JobID jobID, String jobName, Map<JobVertexID, ArchivedExecutionJobVertex> tasks, List<ArchivedExecutionJobVertex> verticesInCreationOrder, long[] stateTimestamps, JobStatus state, @Nullable ErrorInfo failureCause, String jsonPlan, StringifiedAccumulatorResult[] archivedUserAccumulators, Map<String, SerializedValue<OptionalFailure<Object>>> serializedUserAccumulators, ArchivedExecutionConfig executionConfig, boolean isStoppable, @Nullable CheckpointCoordinatorConfiguration jobCheckpointingConfiguration, @Nullable CheckpointStatsSnapshot checkpointStatsSnapshot) { this.jobID = Preconditions.checkNotNull(jobID); this.jobName = Preconditions.checkNotNull(jobName); this.tasks = Preconditions.checkNotNull(tasks); this.verticesInCreationOrder = Preconditions.checkNotNull(verticesInCreationOrder); this.stateTimestamps = Preconditions.checkNotNull(stateTimestamps); this.state = Preconditions.checkNotNull(state); this.failureCause = failureCause; this.jsonPlan = Preconditions.checkNotNull(jsonPlan); this.archivedUserAccumulators = Preconditions.checkNotNull(archivedUserAccumulators); this.serializedUserAccumulators = Preconditions.checkNotNull(serializedUserAccumulators); this.archivedExecutionConfig = Preconditions.checkNotNull(executionConfig); this.isStoppable = isStoppable; this.jobCheckpointingConfiguration = jobCheckpointingConfiguration; this.checkpointStatsSnapshot = checkpointStatsSnapshot; }
Example #13
Source Project: Flink-CEPplus Author: ljygz File: CheckpointStatsTrackerTest.java License: Apache License 2.0 | 5 votes |
/** * Tests access to the snapshotting settings. */ @Test public void testGetSnapshottingSettings() throws Exception { ExecutionJobVertex jobVertex = mock(ExecutionJobVertex.class); when(jobVertex.getJobVertexId()).thenReturn(new JobVertexID()); when(jobVertex.getParallelism()).thenReturn(1); JobCheckpointingSettings snapshottingSettings = new JobCheckpointingSettings( Collections.singletonList(new JobVertexID()), Collections.singletonList(new JobVertexID()), Collections.singletonList(new JobVertexID()), new CheckpointCoordinatorConfiguration( 181238123L, 19191992L, 191929L, 123, CheckpointRetentionPolicy.NEVER_RETAIN_AFTER_TERMINATION, false ), null); CheckpointStatsTracker tracker = new CheckpointStatsTracker( 0, Collections.singletonList(jobVertex), snapshottingSettings.getCheckpointCoordinatorConfiguration(), new UnregisteredMetricsGroup()); assertEquals(snapshottingSettings.getCheckpointCoordinatorConfiguration(), tracker.getJobCheckpointingConfiguration()); }
Example #14
Source Project: Flink-CEPplus Author: ljygz File: CheckpointStatsTrackerTest.java License: Apache License 2.0 | 5 votes |
/** * Creates a "disabled" checkpoint tracker for tests. */ static CheckpointStatsTracker createTestTracker() { ExecutionJobVertex jobVertex = mock(ExecutionJobVertex.class); when(jobVertex.getJobVertexId()).thenReturn(new JobVertexID()); when(jobVertex.getParallelism()).thenReturn(1); return new CheckpointStatsTracker( 0, Collections.singletonList(jobVertex), mock(CheckpointCoordinatorConfiguration.class), new UnregisteredMetricsGroup()); }
Example #15
Source Project: Flink-CEPplus Author: ljygz File: ExecutionGraphDeploymentTest.java License: Apache License 2.0 | 5 votes |
private ExecutionGraph createExecutionGraph(Configuration configuration) throws Exception { final ScheduledExecutorService executor = TestingUtils.defaultExecutor(); final JobID jobId = new JobID(); final JobGraph jobGraph = new JobGraph(jobId, "test"); jobGraph.setSnapshotSettings( new JobCheckpointingSettings( Collections.<JobVertexID>emptyList(), Collections.<JobVertexID>emptyList(), Collections.<JobVertexID>emptyList(), new CheckpointCoordinatorConfiguration( 100, 10 * 60 * 1000, 0, 1, CheckpointRetentionPolicy.NEVER_RETAIN_AFTER_TERMINATION, false), null)); final Time timeout = Time.seconds(10L); return ExecutionGraphBuilder.buildGraph( null, jobGraph, configuration, executor, executor, new ProgrammedSlotProvider(1), getClass().getClassLoader(), new StandaloneCheckpointRecoveryFactory(), timeout, new NoRestartStrategy(), new UnregisteredMetricsGroup(), 1, blobWriter, timeout, LoggerFactory.getLogger(getClass())); }
Example #16
Source Project: flink Author: flink-tpc-ds File: CheckpointStatsTracker.java License: Apache License 2.0 | 5 votes |
/** * Creates a new checkpoint stats tracker. * * @param numRememberedCheckpoints Maximum number of checkpoints to remember, including in progress ones. * @param jobVertices Job vertices involved in the checkpoints. * @param jobCheckpointingConfiguration Checkpointing configuration. * @param metricGroup Metric group for exposed metrics */ public CheckpointStatsTracker( int numRememberedCheckpoints, List<ExecutionJobVertex> jobVertices, CheckpointCoordinatorConfiguration jobCheckpointingConfiguration, MetricGroup metricGroup) { checkArgument(numRememberedCheckpoints >= 0, "Negative number of remembered checkpoints"); this.history = new CheckpointStatsHistory(numRememberedCheckpoints); this.jobVertices = checkNotNull(jobVertices, "JobVertices"); this.jobCheckpointingConfiguration = checkNotNull(jobCheckpointingConfiguration); // Compute the total subtask count. We do this here in order to only // do it once. int count = 0; for (ExecutionJobVertex vertex : jobVertices) { count += vertex.getParallelism(); } this.totalSubtaskCount = count; // Latest snapshot is empty latestSnapshot = new CheckpointStatsSnapshot( counts.createSnapshot(), summary.createSnapshot(), history.createSnapshot(), null); // Register the metrics registerMetrics(metricGroup); }
Example #17
Source Project: flink Author: flink-tpc-ds File: ExecutionGraph.java License: Apache License 2.0 | 5 votes |
@Override public CheckpointCoordinatorConfiguration getCheckpointCoordinatorConfiguration() { if (checkpointStatsTracker != null) { return checkpointStatsTracker.getJobCheckpointingConfiguration(); } else { return null; } }
Example #18
Source Project: flink Author: flink-tpc-ds File: ArchivedExecutionGraph.java License: Apache License 2.0 | 5 votes |
public ArchivedExecutionGraph( JobID jobID, String jobName, Map<JobVertexID, ArchivedExecutionJobVertex> tasks, List<ArchivedExecutionJobVertex> verticesInCreationOrder, long[] stateTimestamps, JobStatus state, @Nullable ErrorInfo failureCause, String jsonPlan, StringifiedAccumulatorResult[] archivedUserAccumulators, Map<String, SerializedValue<OptionalFailure<Object>>> serializedUserAccumulators, ArchivedExecutionConfig executionConfig, boolean isStoppable, @Nullable CheckpointCoordinatorConfiguration jobCheckpointingConfiguration, @Nullable CheckpointStatsSnapshot checkpointStatsSnapshot) { this.jobID = Preconditions.checkNotNull(jobID); this.jobName = Preconditions.checkNotNull(jobName); this.tasks = Preconditions.checkNotNull(tasks); this.verticesInCreationOrder = Preconditions.checkNotNull(verticesInCreationOrder); this.stateTimestamps = Preconditions.checkNotNull(stateTimestamps); this.state = Preconditions.checkNotNull(state); this.failureCause = failureCause; this.jsonPlan = Preconditions.checkNotNull(jsonPlan); this.archivedUserAccumulators = Preconditions.checkNotNull(archivedUserAccumulators); this.serializedUserAccumulators = Preconditions.checkNotNull(serializedUserAccumulators); this.archivedExecutionConfig = Preconditions.checkNotNull(executionConfig); this.isStoppable = isStoppable; this.jobCheckpointingConfiguration = jobCheckpointingConfiguration; this.checkpointStatsSnapshot = checkpointStatsSnapshot; }
Example #19
Source Project: flink Author: flink-tpc-ds File: CheckpointCoordinatorTest.java License: Apache License 2.0 | 5 votes |
/** * Tests that no minimum delay between savepoints is enforced. */ @Test public void testMinDelayBetweenSavepoints() throws Exception { JobID jobId = new JobID(); final ExecutionAttemptID attemptID1 = new ExecutionAttemptID(); ExecutionVertex vertex1 = mockExecutionVertex(attemptID1); CheckpointCoordinatorConfiguration chkConfig = new CheckpointCoordinatorConfiguration( 100000, 200000, 100000000L, // very long min delay => should not affect savepoints 1, CheckpointRetentionPolicy.NEVER_RETAIN_AFTER_TERMINATION, true, false, 0); CheckpointCoordinator coord = new CheckpointCoordinator( jobId, chkConfig, new ExecutionVertex[] { vertex1 }, new ExecutionVertex[] { vertex1 }, new ExecutionVertex[] { vertex1 }, new StandaloneCheckpointIDCounter(), new StandaloneCompletedCheckpointStore(2), new MemoryStateBackend(), Executors.directExecutor(), SharedStateRegistry.DEFAULT_FACTORY, failureManager); String savepointDir = tmpFolder.newFolder().getAbsolutePath(); CompletableFuture<CompletedCheckpoint> savepoint0 = coord.triggerSavepoint(0, savepointDir); assertFalse("Did not trigger savepoint", savepoint0.isDone()); CompletableFuture<CompletedCheckpoint> savepoint1 = coord.triggerSavepoint(1, savepointDir); assertFalse("Did not trigger savepoint", savepoint1.isDone()); }
Example #20
Source Project: flink Author: flink-tpc-ds File: CheckpointCoordinatorTest.java License: Apache License 2.0 | 5 votes |
private CheckpointCoordinator getCheckpointCoordinator( final JobID jobId, final ExecutionVertex vertex1, final ExecutionVertex vertex2, final CheckpointFailureManager failureManager) { final CheckpointCoordinatorConfiguration chkConfig = new CheckpointCoordinatorConfiguration( 600000, 600000, 0, Integer.MAX_VALUE, CheckpointRetentionPolicy.NEVER_RETAIN_AFTER_TERMINATION, true, false, 0); return new CheckpointCoordinator( jobId, chkConfig, new ExecutionVertex[]{vertex1, vertex2}, new ExecutionVertex[]{vertex1, vertex2}, new ExecutionVertex[]{vertex1, vertex2}, new StandaloneCheckpointIDCounter(), new StandaloneCompletedCheckpointStore(1), new MemoryStateBackend(), Executors.directExecutor(), SharedStateRegistry.DEFAULT_FACTORY, failureManager); }
Example #21
Source Project: flink Author: flink-tpc-ds File: CheckpointStatsTrackerTest.java License: Apache License 2.0 | 5 votes |
/** * Tests access to the snapshotting settings. */ @Test public void testGetSnapshottingSettings() throws Exception { ExecutionJobVertex jobVertex = mock(ExecutionJobVertex.class); when(jobVertex.getJobVertexId()).thenReturn(new JobVertexID()); when(jobVertex.getParallelism()).thenReturn(1); JobCheckpointingSettings snapshottingSettings = new JobCheckpointingSettings( Collections.singletonList(new JobVertexID()), Collections.singletonList(new JobVertexID()), Collections.singletonList(new JobVertexID()), new CheckpointCoordinatorConfiguration( 181238123L, 19191992L, 191929L, 123, CheckpointRetentionPolicy.NEVER_RETAIN_AFTER_TERMINATION, false, false, 0 ), null); CheckpointStatsTracker tracker = new CheckpointStatsTracker( 0, Collections.singletonList(jobVertex), snapshottingSettings.getCheckpointCoordinatorConfiguration(), new UnregisteredMetricsGroup()); assertEquals(snapshottingSettings.getCheckpointCoordinatorConfiguration(), tracker.getJobCheckpointingConfiguration()); }
Example #22
Source Project: flink Author: flink-tpc-ds File: CheckpointStatsTrackerTest.java License: Apache License 2.0 | 5 votes |
/** * Creates a "disabled" checkpoint tracker for tests. */ static CheckpointStatsTracker createTestTracker() { ExecutionJobVertex jobVertex = mock(ExecutionJobVertex.class); when(jobVertex.getJobVertexId()).thenReturn(new JobVertexID()); when(jobVertex.getParallelism()).thenReturn(1); return new CheckpointStatsTracker( 0, Collections.singletonList(jobVertex), mock(CheckpointCoordinatorConfiguration.class), new UnregisteredMetricsGroup()); }
Example #23
Source Project: flink Author: flink-tpc-ds File: ExecutionGraphCheckpointCoordinatorTest.java License: Apache License 2.0 | 5 votes |
private ExecutionGraph createExecutionGraphAndEnableCheckpointing( CheckpointIDCounter counter, CompletedCheckpointStore store) throws Exception { final Time timeout = Time.days(1L); JobVertex jobVertex = new JobVertex("MockVertex"); jobVertex.setInvokableClass(AbstractInvokable.class); final ExecutionGraph executionGraph = new ExecutionGraphTestUtils.TestingExecutionGraphBuilder(jobVertex) .setRpcTimeout(timeout) .setAllocationTimeout(timeout) .allowQueuedScheduling() .build(); executionGraph.start(ComponentMainThreadExecutorServiceAdapter.forMainThread()); CheckpointCoordinatorConfiguration chkConfig = new CheckpointCoordinatorConfiguration( 100, 100, 100, 1, CheckpointRetentionPolicy.NEVER_RETAIN_AFTER_TERMINATION, true, false, 0); executionGraph.enableCheckpointing( chkConfig, Collections.emptyList(), Collections.emptyList(), Collections.emptyList(), Collections.emptyList(), counter, store, new MemoryStateBackend(), CheckpointStatsTrackerTest.createTestTracker()); return executionGraph; }
Example #24
Source Project: flink Author: flink-tpc-ds File: ExecutionGraphDeploymentTest.java License: Apache License 2.0 | 5 votes |
private ExecutionGraph createExecutionGraph(Configuration configuration) throws Exception { final ScheduledExecutorService executor = TestingUtils.defaultExecutor(); final JobID jobId = new JobID(); final JobGraph jobGraph = new JobGraph(jobId, "test"); jobGraph.setSnapshotSettings( new JobCheckpointingSettings( Collections.<JobVertexID>emptyList(), Collections.<JobVertexID>emptyList(), Collections.<JobVertexID>emptyList(), new CheckpointCoordinatorConfiguration( 100, 10 * 60 * 1000, 0, 1, CheckpointRetentionPolicy.NEVER_RETAIN_AFTER_TERMINATION, false, false, 0), null)); final Time timeout = Time.seconds(10L); return ExecutionGraphBuilder.buildGraph( null, jobGraph, configuration, executor, executor, new ProgrammedSlotProvider(1), getClass().getClassLoader(), new StandaloneCheckpointRecoveryFactory(), timeout, new NoRestartStrategy(), new UnregisteredMetricsGroup(), blobWriter, timeout, LoggerFactory.getLogger(getClass()), NettyShuffleMaster.INSTANCE, NoOpPartitionTracker.INSTANCE); }
Example #25
Source Project: flink Author: flink-tpc-ds File: AdaptedRestartPipelinedRegionStrategyNGAbortPendingCheckpointsTest.java License: Apache License 2.0 | 5 votes |
private static void enableCheckpointing(final ExecutionGraph executionGraph) { final List<ExecutionJobVertex> jobVertices = new ArrayList<>(executionGraph.getAllVertices().values()); final CheckpointCoordinatorConfiguration checkpointCoordinatorConfiguration = new CheckpointCoordinatorConfiguration( Long.MAX_VALUE, Long.MAX_VALUE, 0, 1, CheckpointRetentionPolicy.RETAIN_ON_CANCELLATION, true, false, 0); executionGraph.enableCheckpointing( checkpointCoordinatorConfiguration, jobVertices, jobVertices, jobVertices, Collections.emptyList(), new StandaloneCheckpointIDCounter(), new StandaloneCompletedCheckpointStore(1), new MemoryStateBackend(), new CheckpointStatsTracker( 0, jobVertices, checkpointCoordinatorConfiguration, new UnregisteredMetricsGroup())); }
Example #26
Source Project: flink Author: apache File: CheckpointCoordinator.java License: Apache License 2.0 | 5 votes |
public CheckpointCoordinator( JobID job, CheckpointCoordinatorConfiguration chkConfig, ExecutionVertex[] tasksToTrigger, ExecutionVertex[] tasksToWaitFor, ExecutionVertex[] tasksToCommitTo, Collection<OperatorCoordinatorCheckpointContext> coordinatorsToCheckpoint, CheckpointIDCounter checkpointIDCounter, CompletedCheckpointStore completedCheckpointStore, StateBackend checkpointStateBackend, Executor executor, ScheduledExecutor timer, SharedStateRegistryFactory sharedStateRegistryFactory, CheckpointFailureManager failureManager) { this( job, chkConfig, tasksToTrigger, tasksToWaitFor, tasksToCommitTo, coordinatorsToCheckpoint, checkpointIDCounter, completedCheckpointStore, checkpointStateBackend, executor, timer, sharedStateRegistryFactory, failureManager, SystemClock.getInstance()); }
Example #27
Source Project: flink Author: apache File: CheckpointStatsTracker.java License: Apache License 2.0 | 5 votes |
/** * Creates a new checkpoint stats tracker. * * @param numRememberedCheckpoints Maximum number of checkpoints to remember, including in progress ones. * @param jobVertices Job vertices involved in the checkpoints. * @param jobCheckpointingConfiguration Checkpointing configuration. * @param metricGroup Metric group for exposed metrics */ public CheckpointStatsTracker( int numRememberedCheckpoints, List<ExecutionJobVertex> jobVertices, CheckpointCoordinatorConfiguration jobCheckpointingConfiguration, MetricGroup metricGroup) { checkArgument(numRememberedCheckpoints >= 0, "Negative number of remembered checkpoints"); this.history = new CheckpointStatsHistory(numRememberedCheckpoints); this.jobVertices = checkNotNull(jobVertices, "JobVertices"); this.jobCheckpointingConfiguration = checkNotNull(jobCheckpointingConfiguration); // Compute the total subtask count. We do this here in order to only // do it once. int count = 0; for (ExecutionJobVertex vertex : jobVertices) { count += vertex.getParallelism(); } this.totalSubtaskCount = count; // Latest snapshot is empty latestSnapshot = new CheckpointStatsSnapshot( counts.createSnapshot(), summary.createSnapshot(), history.createSnapshot(), null); // Register the metrics registerMetrics(metricGroup); }
Example #28
Source Project: flink Author: apache File: ExecutionGraph.java License: Apache License 2.0 | 5 votes |
@Override public CheckpointCoordinatorConfiguration getCheckpointCoordinatorConfiguration() { if (checkpointStatsTracker != null) { return checkpointStatsTracker.getJobCheckpointingConfiguration(); } else { return null; } }
Example #29
Source Project: flink Author: apache File: SchedulerTestingUtils.java License: Apache License 2.0 | 5 votes |
public static void enableCheckpointing(final JobGraph jobGraph, @Nullable StateBackend stateBackend) { final List<JobVertexID> triggerVertices = new ArrayList<>(); final List<JobVertexID> allVertices = new ArrayList<>(); for (JobVertex vertex : jobGraph.getVertices()) { if (vertex.isInputVertex()) { triggerVertices.add(vertex.getID()); } allVertices.add(vertex.getID()); } final CheckpointCoordinatorConfiguration config = new CheckpointCoordinatorConfiguration( Long.MAX_VALUE, // disable periodical checkpointing DEFAULT_CHECKPOINT_TIMEOUT_MS, 0, 1, CheckpointRetentionPolicy.NEVER_RETAIN_AFTER_TERMINATION, false, false, false, 0); SerializedValue<StateBackend> serializedStateBackend = null; if (stateBackend != null) { try { serializedStateBackend = new SerializedValue<>(stateBackend); } catch (IOException e) { throw new RuntimeException("could not serialize state backend", e); } } jobGraph.setSnapshotSettings(new JobCheckpointingSettings( triggerVertices, allVertices, allVertices, config, serializedStateBackend)); }
Example #30
Source Project: flink Author: apache File: CheckpointCoordinatorTest.java License: Apache License 2.0 | 5 votes |
/** * Tests that the externalized checkpoint configuration is respected. */ @Test public void testExternalizedCheckpoints() throws Exception { try { // set up the coordinator and validate the initial state CheckpointCoordinatorConfiguration chkConfig = new CheckpointCoordinatorConfigurationBuilder() .setCheckpointRetentionPolicy(CheckpointRetentionPolicy.RETAIN_ON_FAILURE) .build(); CheckpointCoordinator coord = new CheckpointCoordinatorBuilder() .setCheckpointCoordinatorConfiguration(chkConfig) .setTimer(manuallyTriggeredScheduledExecutor) .build(); CompletableFuture<CompletedCheckpoint> checkpointFuture = coord.triggerCheckpoint(false); manuallyTriggeredScheduledExecutor.triggerAll(); assertFalse(checkpointFuture.isCompletedExceptionally()); for (PendingCheckpoint checkpoint : coord.getPendingCheckpoints().values()) { CheckpointProperties props = checkpoint.getProps(); CheckpointProperties expected = CheckpointProperties.forCheckpoint(CheckpointRetentionPolicy.RETAIN_ON_FAILURE); assertEquals(expected, props); } // the now we should have a completed checkpoint coord.shutdown(JobStatus.FINISHED); } catch (Exception e) { e.printStackTrace(); fail(e.getMessage()); } }