Examples with TaskToolbox org.apache.druid.indexing.common.TaskToolbox used on opensource projects

Example 6 with TaskToolbox

use of org.apache.druid.indexing.common.TaskToolbox in project druid by druid-io.

the class IndexTaskTest method testWaitForSegmentAvailabilityMultipleSegmentsTimeout.

@Test
public void testWaitForSegmentAvailabilityMultipleSegmentsTimeout() throws IOException {
    final File tmpDir = temporaryFolder.newFolder();
    TaskToolbox mockToolbox = EasyMock.createMock(TaskToolbox.class);
    SegmentHandoffNotifierFactory mockFactory = EasyMock.createMock(SegmentHandoffNotifierFactory.class);
    SegmentHandoffNotifier mockNotifier = EasyMock.createMock(SegmentHandoffNotifier.class);
    DataSegment mockDataSegment1 = EasyMock.createMock(DataSegment.class);
    DataSegment mockDataSegment2 = EasyMock.createMock(DataSegment.class);
    List<DataSegment> segmentsToWaitFor = new ArrayList<>();
    segmentsToWaitFor.add(mockDataSegment1);
    segmentsToWaitFor.add(mockDataSegment2);
    IndexTask indexTask = new IndexTask(null, null, createDefaultIngestionSpec(jsonMapper, tmpDir, new UniformGranularitySpec(Granularities.HOUR, Granularities.MINUTE, null), null, createTuningConfigWithMaxRowsPerSegment(2, true), false, false), null);
    EasyMock.expect(mockDataSegment1.getInterval()).andReturn(Intervals.of("1970-01-01/2100-01-01")).once();
    EasyMock.expect(mockDataSegment1.getVersion()).andReturn("dummyString").once();
    EasyMock.expect(mockDataSegment1.getShardSpec()).andReturn(EasyMock.createMock(ShardSpec.class)).once();
    EasyMock.expect(mockDataSegment2.getInterval()).andReturn(Intervals.of("1970-01-01/2100-01-01")).once();
    EasyMock.expect(mockDataSegment2.getVersion()).andReturn("dummyString").once();
    EasyMock.expect(mockDataSegment2.getShardSpec()).andReturn(EasyMock.createMock(ShardSpec.class)).once();
    EasyMock.expect(mockToolbox.getSegmentHandoffNotifierFactory()).andReturn(mockFactory).once();
    EasyMock.expect(mockToolbox.getEmitter()).andReturn(new NoopServiceEmitter()).anyTimes();
    EasyMock.expect(mockDataSegment1.getDataSource()).andReturn("MockDataSource").once();
    EasyMock.expect(mockFactory.createSegmentHandoffNotifier("MockDataSource")).andReturn(mockNotifier).once();
    mockNotifier.start();
    EasyMock.expectLastCall().once();
    mockNotifier.registerSegmentHandoffCallback(EasyMock.anyObject(), EasyMock.anyObject(), EasyMock.anyObject());
    EasyMock.expectLastCall().andReturn(true).times(2);
    mockNotifier.close();
    EasyMock.expectLastCall().once();
    EasyMock.replay(mockToolbox);
    EasyMock.replay(mockDataSegment1, mockDataSegment2);
    EasyMock.replay(mockFactory, mockNotifier);
    Assert.assertFalse(indexTask.waitForSegmentAvailability(mockToolbox, segmentsToWaitFor, 1000));
    EasyMock.verify(mockToolbox);
    EasyMock.verify(mockDataSegment1, mockDataSegment2);
    EasyMock.verify(mockFactory, mockNotifier);
}

Example 7 with TaskToolbox

use of org.apache.druid.indexing.common.TaskToolbox in project druid by druid-io.

the class RealtimeIndexTaskTest method testTransformSpec.

@Test(timeout = 60_000L)
public void testTransformSpec() throws Exception {
    final TestIndexerMetadataStorageCoordinator mdc = new TestIndexerMetadataStorageCoordinator();
    final TransformSpec transformSpec = new TransformSpec(new SelectorDimFilter("dim1", "foo", null), ImmutableList.of(new ExpressionTransform("dim1t", "concat(dim1,dim1)", ExprMacroTable.nil())));
    final RealtimeIndexTask task = makeRealtimeTask(null, transformSpec, true, 0);
    final TaskToolbox taskToolbox = makeToolbox(task, mdc, tempFolder.newFolder());
    final ListenableFuture<TaskStatus> statusFuture = runTask(task, taskToolbox);
    final DataSegment publishedSegment;
    // Wait for firehose to show up, it starts off null.
    while (task.getFirehose() == null) {
        Thread.sleep(50);
    }
    final TestFirehose firehose = (TestFirehose) task.getFirehose();
    firehose.addRows(ImmutableList.of(ImmutableMap.of("t", now.getMillis(), "dim1", "foo", "met1", "1"), ImmutableMap.of("t", now.minus(new Period("P1D")).getMillis(), "dim1", "foo", "met1", 2.0), ImmutableMap.of("t", now.getMillis(), "dim2", "bar", "met1", 2.0)));
    // Stop the firehose, this will drain out existing events.
    firehose.close();
    // Wait for publish.
    while (mdc.getPublished().isEmpty()) {
        Thread.sleep(50);
    }
    publishedSegment = Iterables.getOnlyElement(mdc.getPublished());
    // Check metrics.
    Assert.assertEquals(1, task.getMetrics().processed());
    Assert.assertEquals(2, task.getMetrics().thrownAway());
    Assert.assertEquals(0, task.getMetrics().unparseable());
    // Do some queries.
    Assert.assertEquals(1, sumMetric(task, null, "rows").longValue());
    Assert.assertEquals(1, sumMetric(task, new SelectorDimFilter("dim1t", "foofoo", null), "rows").longValue());
    if (NullHandling.replaceWithDefault()) {
        Assert.assertEquals(0, sumMetric(task, new SelectorDimFilter("dim1t", "barbar", null), "rows").longValue());
    } else {
        Assert.assertNull(sumMetric(task, new SelectorDimFilter("dim1t", "barbar", null), "rows"));
    }
    Assert.assertEquals(1, sumMetric(task, null, "met1").longValue());
    // Simulate handoff.
    for (Map.Entry<SegmentDescriptor, Pair<Executor, Runnable>> entry : handOffCallbacks.entrySet()) {
        final Pair<Executor, Runnable> executorRunnablePair = entry.getValue();
        Assert.assertEquals(new SegmentDescriptor(publishedSegment.getInterval(), publishedSegment.getVersion(), publishedSegment.getShardSpec().getPartitionNum()), entry.getKey());
        executorRunnablePair.lhs.execute(executorRunnablePair.rhs);
    }
    handOffCallbacks.clear();
    // Wait for the task to finish.
    final TaskStatus taskStatus = statusFuture.get();
    Assert.assertEquals(TaskState.SUCCESS, taskStatus.getStatusCode());
}

Example 8 with TaskToolbox

use of org.apache.druid.indexing.common.TaskToolbox in project druid by druid-io.

the class RealtimeIndexTaskTest method testRestoreCorruptData.

@Test(timeout = 60_000L)
public void testRestoreCorruptData() throws Exception {
    final File directory = tempFolder.newFolder();
    final RealtimeIndexTask task1 = makeRealtimeTask(null);
    // First run:
    {
        final TestIndexerMetadataStorageCoordinator mdc = new TestIndexerMetadataStorageCoordinator();
        final TaskToolbox taskToolbox = makeToolbox(task1, mdc, directory);
        final ListenableFuture<TaskStatus> statusFuture = runTask(task1, taskToolbox);
        // Wait for firehose to show up, it starts off null.
        while (task1.getFirehose() == null) {
            Thread.sleep(50);
        }
        final TestFirehose firehose = (TestFirehose) task1.getFirehose();
        firehose.addRows(ImmutableList.of(ImmutableMap.of("t", now.getMillis(), "dim1", "foo")));
        // Trigger graceful shutdown.
        task1.stopGracefully(taskToolbox.getConfig());
        // Wait for the task to finish. The status doesn't really matter, but we'll check it anyway.
        final TaskStatus taskStatus = statusFuture.get();
        Assert.assertEquals(TaskState.SUCCESS, taskStatus.getStatusCode());
        // Nothing should be published.
        Assert.assertEquals(new HashSet<>(), mdc.getPublished());
    }
    // Corrupt the data:
    final File smooshFile = new File(StringUtils.format("%s/persistent/task/%s/work/persist/%s/%s_%s/0/00000.smoosh", directory, task1.getId(), task1.getDataSource(), Granularities.DAY.bucketStart(now), Granularities.DAY.bucketEnd(now)));
    Files.write(smooshFile.toPath(), StringUtils.toUtf8("oops!"));
    // Second run:
    {
        final TestIndexerMetadataStorageCoordinator mdc = new TestIndexerMetadataStorageCoordinator();
        final RealtimeIndexTask task2 = makeRealtimeTask(task1.getId());
        final TaskToolbox taskToolbox = makeToolbox(task2, mdc, directory);
        final ListenableFuture<TaskStatus> statusFuture = runTask(task2, taskToolbox);
        // Wait for the task to finish.
        boolean caught = false;
        try {
            statusFuture.get();
        } catch (Exception e) {
            caught = true;
        }
        Assert.assertTrue("expected exception", caught);
    }
}

Example 9 with TaskToolbox

use of org.apache.druid.indexing.common.TaskToolbox in project druid by druid-io.

the class RealtimeIndexTaskTest method testRestoreAfterHandoffAttemptDuringShutdown.

@Test(timeout = 60_000L)
public void testRestoreAfterHandoffAttemptDuringShutdown() throws Exception {
    final TaskStorage taskStorage = new HeapMemoryTaskStorage(new TaskStorageConfig(null));
    final TestIndexerMetadataStorageCoordinator mdc = new TestIndexerMetadataStorageCoordinator();
    final File directory = tempFolder.newFolder();
    final RealtimeIndexTask task1 = makeRealtimeTask(null);
    final DataSegment publishedSegment;
    // First run:
    {
        final TaskToolbox taskToolbox = makeToolbox(task1, taskStorage, mdc, directory);
        final ListenableFuture<TaskStatus> statusFuture = runTask(task1, taskToolbox);
        // Wait for firehose to show up, it starts off null.
        while (task1.getFirehose() == null) {
            Thread.sleep(50);
        }
        final TestFirehose firehose = (TestFirehose) task1.getFirehose();
        firehose.addRows(ImmutableList.of(ImmutableMap.of("t", now.getMillis(), "dim1", "foo")));
        // Stop the firehose, this will trigger a finishJob.
        firehose.close();
        // Wait for publish.
        while (mdc.getPublished().isEmpty()) {
            Thread.sleep(50);
        }
        publishedSegment = Iterables.getOnlyElement(mdc.getPublished());
        // Do a query.
        Assert.assertEquals(1, sumMetric(task1, null, "rows").longValue());
        // Trigger graceful shutdown.
        task1.stopGracefully(taskToolbox.getConfig());
        // Wait for the task to finish. The status doesn't really matter.
        while (!statusFuture.isDone()) {
            Thread.sleep(50);
        }
    }
    // Second run:
    {
        final RealtimeIndexTask task2 = makeRealtimeTask(task1.getId());
        final TaskToolbox taskToolbox = makeToolbox(task2, taskStorage, mdc, directory);
        final ListenableFuture<TaskStatus> statusFuture = runTask(task2, taskToolbox);
        // Wait for firehose to show up, it starts off null.
        while (task2.getFirehose() == null) {
            Thread.sleep(50);
        }
        // Stop the firehose again, this will start another handoff.
        final TestFirehose firehose = (TestFirehose) task2.getFirehose();
        // Stop the firehose, this will trigger a finishJob.
        firehose.close();
        // publishedSegment is still published. No reason it shouldn't be.
        Assert.assertEquals(ImmutableSet.of(publishedSegment), mdc.getPublished());
        // Wait for a handoffCallback to show up.
        while (handOffCallbacks.isEmpty()) {
            Thread.sleep(50);
        }
        // Simulate handoff.
        for (Map.Entry<SegmentDescriptor, Pair<Executor, Runnable>> entry : handOffCallbacks.entrySet()) {
            final Pair<Executor, Runnable> executorRunnablePair = entry.getValue();
            Assert.assertEquals(new SegmentDescriptor(publishedSegment.getInterval(), publishedSegment.getVersion(), publishedSegment.getShardSpec().getPartitionNum()), entry.getKey());
            executorRunnablePair.lhs.execute(executorRunnablePair.rhs);
        }
        handOffCallbacks.clear();
        // Wait for the task to finish.
        final TaskStatus taskStatus = statusFuture.get();
        Assert.assertEquals(TaskState.SUCCESS, taskStatus.getStatusCode());
    }
}

Example 10 with TaskToolbox

use of org.apache.druid.indexing.common.TaskToolbox in project druid by druid-io.

the class RangePartitionTaskKillTest method failsFirstPhase.

@Test(timeout = 5000L)
public void failsFirstPhase() throws Exception {
    int targetRowsPerSegment = NUM_ROW * 2 / DIM_FILE_CARDINALITY / NUM_PARTITION;
    final ParallelIndexSupervisorTask task = newTask(TIMESTAMP_SPEC, DIMENSIONS_SPEC, INPUT_FORMAT, null, INTERVAL_TO_INDEX, inputDir, TEST_FILE_NAME_PREFIX + "*", new SingleDimensionPartitionsSpec(targetRowsPerSegment, null, DIM1, false), 2, false, 0);
    final TaskActionClient actionClient = createActionClient(task);
    final TaskToolbox toolbox = createTaskToolbox(task, actionClient);
    prepareTaskForLocking(task);
    Assert.assertTrue(task.isReady(actionClient));
    task.stopGracefully(null);
    TaskStatus taskStatus = task.runRangePartitionMultiPhaseParallel(toolbox);
    Assert.assertTrue(taskStatus.isFailure());
    Assert.assertEquals("Failed in phase[PHASE-1]. See task logs for details.", taskStatus.getErrorMsg());
}