Search in sources :

Example 6 with RowLevelPolicyChecker

use of org.apache.gobblin.qualitychecker.row.RowLevelPolicyChecker in project incubator-gobblin by apache.

the class TaskTest method testRetryTask.

/**
 * Check if a {@link WorkUnitState.WorkingState} of a {@link Task} is set properly after a {@link Task} fails once,
 * but then is successful the next time.
 */
@Test(dataProvider = "stateOverrides")
public void testRetryTask(State overrides) throws Exception {
    // Create a TaskState
    TaskState taskState = getEmptyTestTaskState("testRetryTaskId");
    taskState.addAll(overrides);
    // Create a mock TaskContext
    TaskContext mockTaskContext = mock(TaskContext.class);
    when(mockTaskContext.getTaskMetrics()).thenReturn(TaskMetrics.get(taskState));
    when(mockTaskContext.getExtractor()).thenReturn(new FailOnceExtractor());
    when(mockTaskContext.getForkOperator()).thenReturn(new IdentityForkOperator());
    when(mockTaskContext.getTaskState()).thenReturn(taskState);
    when(mockTaskContext.getTaskLevelPolicyChecker(any(TaskState.class), anyInt())).thenReturn(mock(TaskLevelPolicyChecker.class));
    when(mockTaskContext.getRowLevelPolicyChecker()).thenReturn(new RowLevelPolicyChecker(Lists.newArrayList(), "ss", FileSystem.getLocal(new Configuration())));
    when(mockTaskContext.getRowLevelPolicyChecker(anyInt())).thenReturn(new RowLevelPolicyChecker(Lists.newArrayList(), "ss", FileSystem.getLocal(new Configuration())));
    // Create a mock TaskPublisher
    TaskPublisher mockTaskPublisher = mock(TaskPublisher.class);
    when(mockTaskPublisher.canPublish()).thenReturn(TaskPublisher.PublisherState.SUCCESS);
    when(mockTaskContext.getTaskPublisher(any(TaskState.class), any(TaskLevelPolicyCheckResults.class))).thenReturn(mockTaskPublisher);
    // Create a mock TaskStateTracker
    TaskStateTracker mockTaskStateTracker = mock(TaskStateTracker.class);
    // Create a TaskExecutor - a real TaskExecutor must be created so a Fork is run in a separate thread
    TaskExecutor taskExecutor = new TaskExecutor(new Properties());
    // Create the Task
    Task realTask = new Task(mockTaskContext, mockTaskStateTracker, taskExecutor, Optional.<CountDownLatch>absent());
    Task task = spy(realTask);
    doNothing().when(task).submitTaskCommittedEvent();
    // The first run of the Task should fail
    task.run();
    task.commit();
    Assert.assertEquals(task.getTaskState().getWorkingState(), WorkUnitState.WorkingState.FAILED);
    // The second run of the Task should succeed
    task.run();
    task.commit();
    Assert.assertEquals(task.getTaskState().getWorkingState(), WorkUnitState.WorkingState.SUCCESSFUL);
}
Also used : TaskPublisher(org.apache.gobblin.publisher.TaskPublisher) Configuration(org.apache.hadoop.conf.Configuration) TaskLevelPolicyChecker(org.apache.gobblin.qualitychecker.task.TaskLevelPolicyChecker) TaskLevelPolicyCheckResults(org.apache.gobblin.qualitychecker.task.TaskLevelPolicyCheckResults) Properties(java.util.Properties) IdentityForkOperator(org.apache.gobblin.fork.IdentityForkOperator) RowLevelPolicyChecker(org.apache.gobblin.qualitychecker.row.RowLevelPolicyChecker) Test(org.testng.annotations.Test)

Example 7 with RowLevelPolicyChecker

use of org.apache.gobblin.qualitychecker.row.RowLevelPolicyChecker in project incubator-gobblin by apache.

the class TestRecordStream method testAcks.

@Test
public void testAcks() throws Exception {
    StreamEntity[] entities = new StreamEntity[] { new RecordEnvelope<>("a"), new BasicTestControlMessage("1"), new RecordEnvelope<>("b"), new BasicTestControlMessage("2") };
    BasicAckableForTesting ackable = new BasicAckableForTesting();
    for (int i = 0; i < entities.length; i++) {
        entities[i].addCallBack(ackable);
    }
    MyExtractor extractor = new MyExtractor(entities);
    MyConverter converter = new MyConverter();
    MyDataWriter writer = new MyDataWriter();
    // Create a TaskState
    TaskState taskState = getEmptyTestTaskState("testRetryTaskId");
    taskState.setProp(ConfigurationKeys.TASK_SYNCHRONOUS_EXECUTION_MODEL_KEY, false);
    // Create a mock TaskContext
    TaskContext mockTaskContext = mock(TaskContext.class);
    when(mockTaskContext.getExtractor()).thenReturn(extractor);
    when(mockTaskContext.getForkOperator()).thenReturn(new IdentityForkOperator());
    when(mockTaskContext.getTaskState()).thenReturn(taskState);
    when(mockTaskContext.getConverters()).thenReturn(Lists.newArrayList(converter));
    when(mockTaskContext.getTaskLevelPolicyChecker(any(TaskState.class), anyInt())).thenReturn(mock(TaskLevelPolicyChecker.class));
    when(mockTaskContext.getRowLevelPolicyChecker()).thenReturn(new RowLevelPolicyChecker(Lists.newArrayList(), "ss", FileSystem.getLocal(new Configuration())));
    when(mockTaskContext.getRowLevelPolicyChecker(anyInt())).thenReturn(new RowLevelPolicyChecker(Lists.newArrayList(), "ss", FileSystem.getLocal(new Configuration())));
    when(mockTaskContext.getDataWriterBuilder(anyInt(), anyInt())).thenReturn(writer);
    // Create a mock TaskPublisher
    TaskPublisher mockTaskPublisher = mock(TaskPublisher.class);
    when(mockTaskPublisher.canPublish()).thenReturn(TaskPublisher.PublisherState.SUCCESS);
    when(mockTaskContext.getTaskPublisher(any(TaskState.class), any(TaskLevelPolicyCheckResults.class))).thenReturn(mockTaskPublisher);
    // Create a mock TaskStateTracker
    TaskStateTracker mockTaskStateTracker = mock(TaskStateTracker.class);
    // Create a TaskExecutor - a real TaskExecutor must be created so a Fork is run in a separate thread
    TaskExecutor taskExecutor = new TaskExecutor(new Properties());
    // Create the Task
    Task realTask = new Task(mockTaskContext, mockTaskStateTracker, taskExecutor, Optional.<CountDownLatch>absent());
    Task task = spy(realTask);
    doNothing().when(task).submitTaskCommittedEvent();
    task.run();
    task.commit();
    Assert.assertEquals(task.getTaskState().getWorkingState(), WorkUnitState.WorkingState.SUCCESSFUL);
    Assert.assertEquals(ackable.acked, 4);
}
Also used : TaskPublisher(org.apache.gobblin.publisher.TaskPublisher) Configuration(org.apache.hadoop.conf.Configuration) RecordEnvelope(org.apache.gobblin.stream.RecordEnvelope) TaskLevelPolicyChecker(org.apache.gobblin.qualitychecker.task.TaskLevelPolicyChecker) StreamEntity(org.apache.gobblin.stream.StreamEntity) TaskLevelPolicyCheckResults(org.apache.gobblin.qualitychecker.task.TaskLevelPolicyCheckResults) Properties(java.util.Properties) IdentityForkOperator(org.apache.gobblin.fork.IdentityForkOperator) RowLevelPolicyChecker(org.apache.gobblin.qualitychecker.row.RowLevelPolicyChecker) BasicAckableForTesting(org.apache.gobblin.ack.BasicAckableForTesting) Test(org.testng.annotations.Test)

Aggregations

RowLevelPolicyChecker (org.apache.gobblin.qualitychecker.row.RowLevelPolicyChecker)7 Configuration (org.apache.hadoop.conf.Configuration)6 TaskPublisher (org.apache.gobblin.publisher.TaskPublisher)5 TaskLevelPolicyCheckResults (org.apache.gobblin.qualitychecker.task.TaskLevelPolicyCheckResults)5 TaskLevelPolicyChecker (org.apache.gobblin.qualitychecker.task.TaskLevelPolicyChecker)5 IdentityForkOperator (org.apache.gobblin.fork.IdentityForkOperator)4 Test (org.testng.annotations.Test)4 Properties (java.util.Properties)3 State (org.apache.gobblin.configuration.State)3 RowLevelPolicyCheckResults (org.apache.gobblin.qualitychecker.row.RowLevelPolicyCheckResults)3 GenericRecord (org.apache.avro.generic.GenericRecord)2 RowLevelPolicyCheckerBuilderFactory (org.apache.gobblin.qualitychecker.row.RowLevelPolicyCheckerBuilderFactory)2 URI (java.net.URI)1 BasicAckableForTesting (org.apache.gobblin.ack.BasicAckableForTesting)1 WorkUnitState (org.apache.gobblin.configuration.WorkUnitState)1 RecordEnvelope (org.apache.gobblin.stream.RecordEnvelope)1 StreamEntity (org.apache.gobblin.stream.StreamEntity)1 WatermarkStorage (org.apache.gobblin.writer.WatermarkStorage)1 FileSystem (org.apache.hadoop.fs.FileSystem)1 Path (org.apache.hadoop.fs.Path)1