org.apache.flink.runtime.checkpoint.CheckpointRetentionPolicy Java Examples
The following examples show how to use
org.apache.flink.runtime.checkpoint.CheckpointRetentionPolicy.
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 File: CheckpointCoordinatorConfiguration.java From Flink-CEPplus with Apache License 2.0 | 6 votes |
public CheckpointCoordinatorConfiguration( long checkpointInterval, long checkpointTimeout, long minPauseBetweenCheckpoints, int maxConcurrentCheckpoints, CheckpointRetentionPolicy checkpointRetentionPolicy, boolean isExactlyOnce) { // sanity checks if (checkpointInterval < 1 || checkpointTimeout < 1 || minPauseBetweenCheckpoints < 0 || maxConcurrentCheckpoints < 1) { throw new IllegalArgumentException(); } this.checkpointInterval = checkpointInterval; this.checkpointTimeout = checkpointTimeout; this.minPauseBetweenCheckpoints = minPauseBetweenCheckpoints; this.maxConcurrentCheckpoints = maxConcurrentCheckpoints; this.checkpointRetentionPolicy = Preconditions.checkNotNull(checkpointRetentionPolicy); this.isExactlyOnce = isExactlyOnce; }
Example #2
Source File: JobGraphTest.java From flink with 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 #3
Source File: JobMasterTest.java From flink with 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 File: CheckpointCoordinatorConfiguration.java From flink with Apache License 2.0 | 6 votes |
/** * @deprecated use {@link #builder()}. */ @Deprecated @VisibleForTesting public CheckpointCoordinatorConfiguration( long checkpointInterval, long checkpointTimeout, long minPauseBetweenCheckpoints, int maxConcurrentCheckpoints, CheckpointRetentionPolicy checkpointRetentionPolicy, boolean isExactlyOnce, boolean isUnalignedCheckpoint, boolean isPreferCheckpointForRecovery, int tolerableCpFailureNumber) { this( checkpointInterval, checkpointTimeout, minPauseBetweenCheckpoints, maxConcurrentCheckpoints, checkpointRetentionPolicy, isExactlyOnce, isPreferCheckpointForRecovery, tolerableCpFailureNumber, isUnalignedCheckpoint); }
Example #5
Source File: CheckpointConfigHandler.java From flink with 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 #6
Source File: FailoverRegionTest.java From flink with 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 File: JobGraphTest.java From flink with 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, 0); return new JobCheckpointingSettings( Collections.emptyList(), Collections.emptyList(), Collections.emptyList(), checkpointCoordinatorConfiguration, null); }
Example #8
Source File: JobCheckpointingSettingsTest.java From flink with Apache License 2.0 | 6 votes |
/** * Tests that the settings are actually serializable. */ @Test public void testIsJavaSerializable() throws Exception { JobCheckpointingSettings settings = new JobCheckpointingSettings( Arrays.asList(new JobVertexID(), new JobVertexID()), Arrays.asList(new JobVertexID(), new JobVertexID()), Arrays.asList(new JobVertexID(), new JobVertexID()), new CheckpointCoordinatorConfiguration( 1231231, 1231, 112, 12, CheckpointRetentionPolicy.RETAIN_ON_FAILURE, false, false, 0), new SerializedValue<>(new MemoryStateBackend())); JobCheckpointingSettings copy = CommonTestUtils.createCopySerializable(settings); assertEquals(settings.getVerticesToAcknowledge(), copy.getVerticesToAcknowledge()); assertEquals(settings.getVerticesToConfirm(), copy.getVerticesToConfirm()); assertEquals(settings.getVerticesToTrigger(), copy.getVerticesToTrigger()); assertEquals(settings.getCheckpointCoordinatorConfiguration(), copy.getCheckpointCoordinatorConfiguration()); assertNotNull(copy.getDefaultStateBackend()); assertTrue(copy.getDefaultStateBackend().deserializeValue(this.getClass().getClassLoader()).getClass() == MemoryStateBackend.class); }
Example #9
Source File: JobMasterTest.java From flink with 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, 0); final JobCheckpointingSettings checkpointingSettings = new JobCheckpointingSettings( Collections.emptyList(), Collections.emptyList(), Collections.emptyList(), checkpoinCoordinatorConfiguration, null); jobGraph.setSnapshotSettings(checkpointingSettings); jobGraph.setSavepointRestoreSettings(savepointRestoreSettings); return jobGraph; }
Example #10
Source File: CheckpointCoordinatorConfiguration.java From flink with Apache License 2.0 | 6 votes |
public CheckpointCoordinatorConfiguration( long checkpointInterval, long checkpointTimeout, long minPauseBetweenCheckpoints, int maxConcurrentCheckpoints, CheckpointRetentionPolicy checkpointRetentionPolicy, boolean isExactlyOnce, boolean isPreferCheckpointForRecovery, int tolerableCpFailureNumber) { // sanity checks if (checkpointInterval < MINIMAL_CHECKPOINT_TIME || checkpointTimeout < MINIMAL_CHECKPOINT_TIME || minPauseBetweenCheckpoints < 0 || maxConcurrentCheckpoints < 1 || tolerableCpFailureNumber < 0) { throw new IllegalArgumentException(); } this.checkpointInterval = checkpointInterval; this.checkpointTimeout = checkpointTimeout; this.minPauseBetweenCheckpoints = minPauseBetweenCheckpoints; this.maxConcurrentCheckpoints = maxConcurrentCheckpoints; this.checkpointRetentionPolicy = Preconditions.checkNotNull(checkpointRetentionPolicy); this.isExactlyOnce = isExactlyOnce; this.isPreferCheckpointForRecovery = isPreferCheckpointForRecovery; this.tolerableCheckpointFailureNumber = tolerableCpFailureNumber; }
Example #11
Source File: CheckpointConfigHandler.java From flink with 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 #12
Source File: JobCheckpointingSettingsTest.java From Flink-CEPplus with Apache License 2.0 | 6 votes |
/** * Tests that the settings are actually serializable. */ @Test public void testIsJavaSerializable() throws Exception { JobCheckpointingSettings settings = new JobCheckpointingSettings( Arrays.asList(new JobVertexID(), new JobVertexID()), Arrays.asList(new JobVertexID(), new JobVertexID()), Arrays.asList(new JobVertexID(), new JobVertexID()), new CheckpointCoordinatorConfiguration( 1231231, 1231, 112, 12, CheckpointRetentionPolicy.RETAIN_ON_FAILURE, false), new SerializedValue<>(new MemoryStateBackend())); JobCheckpointingSettings copy = CommonTestUtils.createCopySerializable(settings); assertEquals(settings.getVerticesToAcknowledge(), copy.getVerticesToAcknowledge()); assertEquals(settings.getVerticesToConfirm(), copy.getVerticesToConfirm()); assertEquals(settings.getVerticesToTrigger(), copy.getVerticesToTrigger()); assertEquals(settings.getCheckpointCoordinatorConfiguration(), copy.getCheckpointCoordinatorConfiguration()); assertNotNull(copy.getDefaultStateBackend()); assertTrue(copy.getDefaultStateBackend().deserializeValue(this.getClass().getClassLoader()).getClass() == MemoryStateBackend.class); }
Example #13
Source File: CheckpointConfigHandler.java From Flink-CEPplus with 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 #14
Source File: JobMasterTest.java From Flink-CEPplus with 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 #15
Source File: SchedulerTestingUtils.java From flink with 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 #16
Source File: ExecutionGraphDeploymentTest.java From Flink-CEPplus with 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 #17
Source File: ExecutionGraphDeploymentTest.java From flink with 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 #18
Source File: FailoverRegionTest.java From flink with Apache License 2.0 | 5 votes |
/** * Attach pending checkpoints of chk-42 and chk-43 to the execution graph. * If {@link #acknowledgeAllCheckpoints(CheckpointCoordinator, Iterator)} called then, * chk-42 would become the completed checkpoint. */ private void attachPendingCheckpoints(ExecutionGraph eg) throws IOException { final Map<Long, PendingCheckpoint> pendingCheckpoints = new HashMap<>(); final Map<ExecutionAttemptID, ExecutionVertex> verticesToConfirm = new HashMap<>(); eg.getAllExecutionVertices().forEach(e -> { Execution ee = e.getCurrentExecutionAttempt(); if (ee != null) { verticesToConfirm.put(ee.getAttemptId(), e); } }); CheckpointCoordinator checkpointCoordinator = eg.getCheckpointCoordinator(); assertNotNull(checkpointCoordinator); CheckpointStorageCoordinatorView checkpointStorage = checkpointCoordinator.getCheckpointStorage(); pendingCheckpoints.put(checkpointId, new PendingCheckpoint( eg.getJobID(), checkpointId, 0L, verticesToConfirm, CheckpointProperties.forCheckpoint(CheckpointRetentionPolicy.RETAIN_ON_FAILURE), checkpointStorage.initializeLocationForCheckpoint(checkpointId), eg.getFutureExecutor())); long newCheckpointId = checkpointId + 1; pendingCheckpoints.put(newCheckpointId, new PendingCheckpoint( eg.getJobID(), newCheckpointId, 0L, verticesToConfirm, CheckpointProperties.forCheckpoint(CheckpointRetentionPolicy.RETAIN_ON_FAILURE), checkpointStorage.initializeLocationForCheckpoint(newCheckpointId), eg.getFutureExecutor())); Whitebox.setInternalState(checkpointCoordinator, "pendingCheckpoints", pendingCheckpoints); }
Example #19
Source File: AdaptedRestartPipelinedRegionStrategyNGAbortPendingCheckpointsTest.java From flink with 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 #20
Source File: JobCheckpointingSettingsTest.java From flink with Apache License 2.0 | 5 votes |
/** * Tests that the settings are actually serializable. */ @Test public void testIsJavaSerializable() throws Exception { JobCheckpointingSettings settings = new JobCheckpointingSettings( Arrays.asList(new JobVertexID(), new JobVertexID()), Arrays.asList(new JobVertexID(), new JobVertexID()), Arrays.asList(new JobVertexID(), new JobVertexID()), new CheckpointCoordinatorConfiguration( 1231231, 1231, 112, 12, CheckpointRetentionPolicy.RETAIN_ON_FAILURE, false, false, false, 0), new SerializedValue<>(new MemoryStateBackend())); JobCheckpointingSettings copy = CommonTestUtils.createCopySerializable(settings); assertEquals(settings.getVerticesToAcknowledge(), copy.getVerticesToAcknowledge()); assertEquals(settings.getVerticesToConfirm(), copy.getVerticesToConfirm()); assertEquals(settings.getVerticesToTrigger(), copy.getVerticesToTrigger()); assertEquals(settings.getCheckpointCoordinatorConfiguration(), copy.getCheckpointCoordinatorConfiguration()); assertNotNull(copy.getDefaultStateBackend()); assertTrue(copy.getDefaultStateBackend().deserializeValue(this.getClass().getClassLoader()).getClass() == MemoryStateBackend.class); }
Example #21
Source File: CheckpointCoordinatorConfiguration.java From flink with Apache License 2.0 | 5 votes |
private CheckpointCoordinatorConfiguration( long checkpointInterval, long checkpointTimeout, long minPauseBetweenCheckpoints, int maxConcurrentCheckpoints, CheckpointRetentionPolicy checkpointRetentionPolicy, boolean isExactlyOnce, boolean isPreferCheckpointForRecovery, int tolerableCpFailureNumber, boolean isUnalignedCheckpointsEnabled) { // sanity checks if (checkpointInterval < MINIMAL_CHECKPOINT_TIME || checkpointTimeout < MINIMAL_CHECKPOINT_TIME || minPauseBetweenCheckpoints < 0 || maxConcurrentCheckpoints < 1 || tolerableCpFailureNumber < 0) { throw new IllegalArgumentException(); } Preconditions.checkArgument(!isUnalignedCheckpointsEnabled || maxConcurrentCheckpoints <= 1, "maxConcurrentCheckpoints can't be > 1 if UnalignedCheckpoints enabled"); this.checkpointInterval = checkpointInterval; this.checkpointTimeout = checkpointTimeout; this.minPauseBetweenCheckpoints = minPauseBetweenCheckpoints; this.maxConcurrentCheckpoints = maxConcurrentCheckpoints; this.checkpointRetentionPolicy = Preconditions.checkNotNull(checkpointRetentionPolicy); this.isExactlyOnce = isExactlyOnce; this.isPreferCheckpointForRecovery = isPreferCheckpointForRecovery; this.tolerableCheckpointFailureNumber = tolerableCpFailureNumber; this.isUnalignedCheckpointsEnabled = isUnalignedCheckpointsEnabled; }
Example #22
Source File: JMXJobManagerMetricTest.java From flink with Apache License 2.0 | 4 votes |
/** * Tests that metrics registered on the JobManager are actually accessible via JMX. */ @Test public void testJobManagerJMXMetricAccess() throws Exception { Deadline deadline = Deadline.now().plus(Duration.ofMinutes(2)); try { JobVertex sourceJobVertex = new JobVertex("Source"); sourceJobVertex.setInvokableClass(BlockingInvokable.class); JobGraph jobGraph = new JobGraph("TestingJob", sourceJobVertex); jobGraph.setSnapshotSettings(new JobCheckpointingSettings( Collections.<JobVertexID>emptyList(), Collections.<JobVertexID>emptyList(), Collections.<JobVertexID>emptyList(), new CheckpointCoordinatorConfiguration( 500, 500, 50, 5, CheckpointRetentionPolicy.NEVER_RETAIN_AFTER_TERMINATION, true, false, false, 0), null)); ClusterClient<?> client = MINI_CLUSTER_RESOURCE.getClusterClient(); ClientUtils.submitJob(client, jobGraph); FutureUtils.retrySuccessfulWithDelay( () -> client.getJobStatus(jobGraph.getJobID()), Time.milliseconds(10), deadline, status -> status == JobStatus.RUNNING, TestingUtils.defaultScheduledExecutor() ).get(deadline.timeLeft().toMillis(), TimeUnit.MILLISECONDS); MBeanServer mBeanServer = ManagementFactory.getPlatformMBeanServer(); Set<ObjectName> nameSet = mBeanServer.queryNames(new ObjectName("org.apache.flink.jobmanager.job.lastCheckpointSize:job_name=TestingJob,*"), null); Assert.assertEquals(1, nameSet.size()); assertEquals(-1L, mBeanServer.getAttribute(nameSet.iterator().next(), "Value")); BlockingInvokable.unblock(); } finally { BlockingInvokable.unblock(); } }
Example #23
Source File: ArchivedExecutionGraphTest.java From flink with Apache License 2.0 | 4 votes |
@BeforeClass public static void setupExecutionGraph() throws Exception { // ------------------------------------------------------------------------------------------------------------- // Setup // ------------------------------------------------------------------------------------------------------------- JobVertexID v1ID = new JobVertexID(); JobVertexID v2ID = new JobVertexID(); JobVertex v1 = new JobVertex("v1", v1ID); JobVertex v2 = new JobVertex("v2", v2ID); v1.setParallelism(1); v2.setParallelism(2); v1.setInvokableClass(AbstractInvokable.class); v2.setInvokableClass(AbstractInvokable.class); JobGraph jobGraph = new JobGraph(v1, v2); ExecutionConfig config = new ExecutionConfig(); config.setExecutionMode(ExecutionMode.BATCH_FORCED); config.setRestartStrategy(new RestartStrategies.NoRestartStrategyConfiguration()); config.setParallelism(4); config.enableObjectReuse(); config.setGlobalJobParameters(new TestJobParameters()); jobGraph.setExecutionConfig(config); runtimeGraph = TestingExecutionGraphBuilder .newBuilder() .setJobGraph(jobGraph) .build(); runtimeGraph.start(ComponentMainThreadExecutorServiceAdapter.forMainThread()); List<ExecutionJobVertex> jobVertices = new ArrayList<>(); jobVertices.add(runtimeGraph.getJobVertex(v1ID)); jobVertices.add(runtimeGraph.getJobVertex(v2ID)); CheckpointStatsTracker statsTracker = new CheckpointStatsTracker( 0, jobVertices, mock(CheckpointCoordinatorConfiguration.class), new UnregisteredMetricsGroup()); CheckpointCoordinatorConfiguration chkConfig = new CheckpointCoordinatorConfiguration( 100, 100, 100, 1, CheckpointRetentionPolicy.NEVER_RETAIN_AFTER_TERMINATION, true, false, false, 0); runtimeGraph.enableCheckpointing( chkConfig, Collections.<ExecutionJobVertex>emptyList(), Collections.<ExecutionJobVertex>emptyList(), Collections.<ExecutionJobVertex>emptyList(), Collections.<MasterTriggerRestoreHook<?>>emptyList(), new StandaloneCheckpointIDCounter(), new StandaloneCompletedCheckpointStore(1), new MemoryStateBackend(), statsTracker); runtimeGraph.setJsonPlan("{}"); runtimeGraph.getJobVertex(v2ID).getTaskVertices()[0].getCurrentExecutionAttempt().fail(new RuntimeException("This exception was thrown on purpose.")); }
Example #24
Source File: ExecutionGraphDeploymentTest.java From flink with Apache License 2.0 | 4 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, 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, NoOpJobMasterPartitionTracker.INSTANCE); }
Example #25
Source File: CheckpointCoordinatorConfiguration.java From Flink-CEPplus with Apache License 2.0 | 4 votes |
public CheckpointRetentionPolicy getCheckpointRetentionPolicy() { return checkpointRetentionPolicy; }
Example #26
Source File: JobMasterTest.java From flink with Apache License 2.0 | 4 votes |
/** * Tests that an existing checkpoint will have precedence over an savepoint. */ @Test public void testCheckpointPrecedesSavepointRecovery() throws Exception { // create savepoint data final long savepointId = 42L; final File savepointFile = createSavepoint(savepointId); // set savepoint settings final SavepointRestoreSettings savepointRestoreSettings = SavepointRestoreSettings.forPath("" + savepointFile.getAbsolutePath(), true); final JobGraph jobGraph = createJobGraphWithCheckpointing(savepointRestoreSettings); final long checkpointId = 1L; final CompletedCheckpoint completedCheckpoint = new CompletedCheckpoint( jobGraph.getJobID(), checkpointId, 1L, 1L, Collections.emptyMap(), null, CheckpointProperties.forCheckpoint(CheckpointRetentionPolicy.NEVER_RETAIN_AFTER_TERMINATION), new DummyCheckpointStorageLocation()); final StandaloneCompletedCheckpointStore completedCheckpointStore = new StandaloneCompletedCheckpointStore(1); completedCheckpointStore.addCheckpoint(completedCheckpoint); final TestingCheckpointRecoveryFactory testingCheckpointRecoveryFactory = new TestingCheckpointRecoveryFactory(completedCheckpointStore, new StandaloneCheckpointIDCounter()); haServices.setCheckpointRecoveryFactory(testingCheckpointRecoveryFactory); final JobMaster jobMaster = createJobMaster( configuration, jobGraph, haServices, new TestingJobManagerSharedServicesBuilder().build()); try { // starting the JobMaster should have read the savepoint final CompletedCheckpoint savepointCheckpoint = completedCheckpointStore.getLatestCheckpoint(false); assertThat(savepointCheckpoint, Matchers.notNullValue()); assertThat(savepointCheckpoint.getCheckpointID(), is(checkpointId)); } finally { RpcUtils.terminateRpcEndpoint(jobMaster, testingTimeout); } }
Example #27
Source File: CheckpointCoordinatorConfiguration.java From flink with Apache License 2.0 | 4 votes |
public CheckpointCoordinatorConfigurationBuilder setCheckpointRetentionPolicy(CheckpointRetentionPolicy checkpointRetentionPolicy) { this.checkpointRetentionPolicy = checkpointRetentionPolicy; return this; }
Example #28
Source File: CheckpointCoordinatorConfiguration.java From flink with Apache License 2.0 | 4 votes |
public CheckpointRetentionPolicy getCheckpointRetentionPolicy() { return checkpointRetentionPolicy; }
Example #29
Source File: JobMasterTest.java From Flink-CEPplus with Apache License 2.0 | 4 votes |
/** * Tests that an existing checkpoint will have precedence over an savepoint. */ @Test public void testCheckpointPrecedesSavepointRecovery() throws Exception { // create savepoint data final long savepointId = 42L; final File savepointFile = createSavepoint(savepointId); // set savepoint settings final SavepointRestoreSettings savepointRestoreSettings = SavepointRestoreSettings.forPath("" + savepointFile.getAbsolutePath(), true); final JobGraph jobGraph = createJobGraphWithCheckpointing(savepointRestoreSettings); final long checkpointId = 1L; final CompletedCheckpoint completedCheckpoint = new CompletedCheckpoint( jobGraph.getJobID(), checkpointId, 1L, 1L, Collections.emptyMap(), null, CheckpointProperties.forCheckpoint(CheckpointRetentionPolicy.NEVER_RETAIN_AFTER_TERMINATION), new DummyCheckpointStorageLocation()); final StandaloneCompletedCheckpointStore completedCheckpointStore = new StandaloneCompletedCheckpointStore(1); completedCheckpointStore.addCheckpoint(completedCheckpoint); final TestingCheckpointRecoveryFactory testingCheckpointRecoveryFactory = new TestingCheckpointRecoveryFactory(completedCheckpointStore, new StandaloneCheckpointIDCounter()); haServices.setCheckpointRecoveryFactory(testingCheckpointRecoveryFactory); final JobMaster jobMaster = createJobMaster( configuration, jobGraph, haServices, new TestingJobManagerSharedServicesBuilder().build()); try { // starting the JobMaster should have read the savepoint final CompletedCheckpoint savepointCheckpoint = completedCheckpointStore.getLatestCheckpoint(); assertThat(savepointCheckpoint, Matchers.notNullValue()); assertThat(savepointCheckpoint.getCheckpointID(), is(checkpointId)); } finally { RpcUtils.terminateRpcEndpoint(jobMaster, testingTimeout); } }
Example #30
Source File: JMXJobManagerMetricTest.java From flink with Apache License 2.0 | 4 votes |
/** * Tests that metrics registered on the JobManager are actually accessible via JMX. */ @Test public void testJobManagerJMXMetricAccess() throws Exception { Deadline deadline = Deadline.now().plus(Duration.ofMinutes(2)); try { JobVertex sourceJobVertex = new JobVertex("Source"); sourceJobVertex.setInvokableClass(BlockingInvokable.class); JobGraph jobGraph = new JobGraph("TestingJob", sourceJobVertex); jobGraph.setSnapshotSettings(new JobCheckpointingSettings( Collections.<JobVertexID>emptyList(), Collections.<JobVertexID>emptyList(), Collections.<JobVertexID>emptyList(), new CheckpointCoordinatorConfiguration( 500, 500, 50, 5, CheckpointRetentionPolicy.NEVER_RETAIN_AFTER_TERMINATION, true, false, 0), null)); ClusterClient<?> client = MINI_CLUSTER_RESOURCE.getClusterClient(); client.setDetached(true); client.submitJob(jobGraph, JMXJobManagerMetricTest.class.getClassLoader()); FutureUtils.retrySuccessfulWithDelay( () -> client.getJobStatus(jobGraph.getJobID()), Time.milliseconds(10), deadline, status -> status == JobStatus.RUNNING, TestingUtils.defaultScheduledExecutor() ).get(deadline.timeLeft().toMillis(), TimeUnit.MILLISECONDS); MBeanServer mBeanServer = ManagementFactory.getPlatformMBeanServer(); Set<ObjectName> nameSet = mBeanServer.queryNames(new ObjectName("org.apache.flink.jobmanager.job.lastCheckpointSize:job_name=TestingJob,*"), null); Assert.assertEquals(1, nameSet.size()); assertEquals(-1L, mBeanServer.getAttribute(nameSet.iterator().next(), "Value")); BlockingInvokable.unblock(); } finally { BlockingInvokable.unblock(); } }