Examples with KafkaIndexTaskClientFactory org.apache.druid.indexing.kafka.KafkaIndexTaskClientFactory used on opensource projects

Example 1 with KafkaIndexTaskClientFactory

use of org.apache.druid.indexing.kafka.KafkaIndexTaskClientFactory in project druid by druid-io.

the class KafkaSupervisorTest method getSupervisor.

/**
 * Use when you don't want generateSequenceNumber overridden
 */
private KafkaSupervisor getSupervisor(int replicas, int taskCount, boolean useEarliestOffset, String duration, Period lateMessageRejectionPeriod, Period earlyMessageRejectionPeriod, boolean suspended, String kafkaHost, DataSchema dataSchema, KafkaSupervisorTuningConfig tuningConfig) {
    Map<String, Object> consumerProperties = new HashMap<>();
    consumerProperties.put("myCustomKey", "myCustomValue");
    consumerProperties.put("bootstrap.servers", kafkaHost);
    consumerProperties.put("isolation.level", "read_committed");
    KafkaSupervisorIOConfig kafkaSupervisorIOConfig = new KafkaSupervisorIOConfig(topic, INPUT_FORMAT, replicas, taskCount, new Period(duration), consumerProperties, null, KafkaSupervisorIOConfig.DEFAULT_POLL_TIMEOUT_MILLIS, new Period("P1D"), new Period("PT30S"), useEarliestOffset, new Period("PT30M"), lateMessageRejectionPeriod, earlyMessageRejectionPeriod, null);
    KafkaIndexTaskClientFactory taskClientFactory = new KafkaIndexTaskClientFactory(null, null) {

        @Override
        public KafkaIndexTaskClient build(TaskInfoProvider taskInfoProvider, String dataSource, int numThreads, Duration httpTimeout, long numRetries) {
            Assert.assertEquals(TEST_CHAT_THREADS, numThreads);
            Assert.assertEquals(TEST_HTTP_TIMEOUT.toStandardDuration(), httpTimeout);
            Assert.assertEquals(TEST_CHAT_RETRIES, numRetries);
            return taskClient;
        }
    };
    return new KafkaSupervisor(taskStorage, taskMaster, indexerMetadataStorageCoordinator, taskClientFactory, OBJECT_MAPPER, new KafkaSupervisorSpec(null, dataSchema, tuningConfig, kafkaSupervisorIOConfig, null, suspended, taskStorage, taskMaster, indexerMetadataStorageCoordinator, taskClientFactory, OBJECT_MAPPER, new NoopServiceEmitter(), new DruidMonitorSchedulerConfig(), rowIngestionMetersFactory, supervisorConfig), rowIngestionMetersFactory);
}

Example 2 with KafkaIndexTaskClientFactory

use of org.apache.druid.indexing.kafka.KafkaIndexTaskClientFactory in project druid by druid-io.

the class KafkaSupervisorTest method testNoInitialStateWithAutoscaler.

@Test
public void testNoInitialStateWithAutoscaler() throws Exception {
    KafkaIndexTaskClientFactory taskClientFactory = new KafkaIndexTaskClientFactory(null, null) {

        @Override
        public KafkaIndexTaskClient build(TaskInfoProvider taskInfoProvider, String dataSource, int numThreads, Duration httpTimeout, long numRetries) {
            Assert.assertEquals(TEST_CHAT_THREADS, numThreads);
            Assert.assertEquals(TEST_HTTP_TIMEOUT.toStandardDuration(), httpTimeout);
            Assert.assertEquals(TEST_CHAT_RETRIES, numRetries);
            return taskClient;
        }
    };
    HashMap<String, Object> autoScalerConfig = new HashMap<>();
    autoScalerConfig.put("enableTaskAutoScaler", true);
    autoScalerConfig.put("lagCollectionIntervalMillis", 500);
    autoScalerConfig.put("lagCollectionRangeMillis", 500);
    autoScalerConfig.put("scaleOutThreshold", 0);
    autoScalerConfig.put("triggerScaleOutFractionThreshold", 0.0);
    autoScalerConfig.put("scaleInThreshold", 1000000);
    autoScalerConfig.put("triggerScaleInFractionThreshold", 0.8);
    autoScalerConfig.put("scaleActionStartDelayMillis", 0);
    autoScalerConfig.put("scaleActionPeriodMillis", 100);
    autoScalerConfig.put("taskCountMax", 2);
    autoScalerConfig.put("taskCountMin", 1);
    autoScalerConfig.put("scaleInStep", 1);
    autoScalerConfig.put("scaleOutStep", 2);
    autoScalerConfig.put("minTriggerScaleActionFrequencyMillis", 1200000);
    final Map<String, Object> consumerProperties = KafkaConsumerConfigs.getConsumerProperties();
    consumerProperties.put("myCustomKey", "myCustomValue");
    consumerProperties.put("bootstrap.servers", kafkaHost);
    KafkaSupervisorIOConfig kafkaSupervisorIOConfig = new KafkaSupervisorIOConfig(topic, INPUT_FORMAT, 1, 1, new Period("PT1H"), consumerProperties, OBJECT_MAPPER.convertValue(autoScalerConfig, LagBasedAutoScalerConfig.class), KafkaSupervisorIOConfig.DEFAULT_POLL_TIMEOUT_MILLIS, new Period("P1D"), new Period("PT30S"), true, new Period("PT30M"), null, null, null);
    final KafkaSupervisorTuningConfig tuningConfigOri = new KafkaSupervisorTuningConfig(null, 1000, null, null, 50000, null, new Period("P1Y"), new File("/test"), null, null, null, false, null, false, null, numThreads, TEST_CHAT_THREADS, TEST_CHAT_RETRIES, TEST_HTTP_TIMEOUT, TEST_SHUTDOWN_TIMEOUT, null, null, null, null, null);
    EasyMock.expect(ingestionSchema.getIOConfig()).andReturn(kafkaSupervisorIOConfig).anyTimes();
    EasyMock.expect(ingestionSchema.getDataSchema()).andReturn(dataSchema).anyTimes();
    EasyMock.expect(ingestionSchema.getTuningConfig()).andReturn(tuningConfigOri).anyTimes();
    EasyMock.replay(ingestionSchema);
    SeekableStreamSupervisorSpec testableSupervisorSpec = new KafkaSupervisorSpec(ingestionSchema, dataSchema, tuningConfigOri, kafkaSupervisorIOConfig, null, false, taskStorage, taskMaster, indexerMetadataStorageCoordinator, taskClientFactory, OBJECT_MAPPER, new NoopServiceEmitter(), new DruidMonitorSchedulerConfig(), rowIngestionMetersFactory, new SupervisorStateManagerConfig());
    supervisor = new TestableKafkaSupervisor(taskStorage, taskMaster, indexerMetadataStorageCoordinator, taskClientFactory, OBJECT_MAPPER, (KafkaSupervisorSpec) testableSupervisorSpec, rowIngestionMetersFactory);
    SupervisorTaskAutoScaler autoscaler = testableSupervisorSpec.createAutoscaler(supervisor);
    final KafkaSupervisorTuningConfig tuningConfig = supervisor.getTuningConfig();
    addSomeEvents(1);
    Capture<KafkaIndexTask> captured = Capture.newInstance();
    EasyMock.expect(taskMaster.getTaskQueue()).andReturn(Optional.of(taskQueue)).anyTimes();
    EasyMock.expect(taskMaster.getTaskRunner()).andReturn(Optional.of(taskRunner)).anyTimes();
    EasyMock.expect(taskMaster.getSupervisorManager()).andReturn(Optional.absent()).anyTimes();
    EasyMock.expect(taskStorage.getActiveTasksByDatasource(DATASOURCE)).andReturn(ImmutableList.of()).anyTimes();
    EasyMock.expect(indexerMetadataStorageCoordinator.retrieveDataSourceMetadata(DATASOURCE)).andReturn(new KafkaDataSourceMetadata(null)).anyTimes();
    EasyMock.expect(taskQueue.add(EasyMock.capture(captured))).andReturn(true);
    taskRunner.registerListener(EasyMock.anyObject(TaskRunnerListener.class), EasyMock.anyObject(Executor.class));
    replayAll();
    supervisor.start();
    int taskCountBeforeScale = supervisor.getIoConfig().getTaskCount();
    Assert.assertEquals(1, taskCountBeforeScale);
    autoscaler.start();
    supervisor.runInternal();
    Thread.sleep(1 * 1000);
    verifyAll();
    int taskCountAfterScale = supervisor.getIoConfig().getTaskCount();
    Assert.assertEquals(2, taskCountAfterScale);
    KafkaIndexTask task = captured.getValue();
    Assert.assertEquals(KafkaSupervisorTest.dataSchema, task.getDataSchema());
    Assert.assertEquals(tuningConfig.convertToTaskTuningConfig(), task.getTuningConfig());
    KafkaIndexTaskIOConfig taskConfig = task.getIOConfig();
    Assert.assertEquals(kafkaHost, taskConfig.getConsumerProperties().get("bootstrap.servers"));
    Assert.assertEquals("myCustomValue", taskConfig.getConsumerProperties().get("myCustomKey"));
    Assert.assertEquals("sequenceName-0", taskConfig.getBaseSequenceName());
    Assert.assertTrue("isUseTransaction", taskConfig.isUseTransaction());
    Assert.assertFalse("minimumMessageTime", taskConfig.getMinimumMessageTime().isPresent());
    Assert.assertFalse("maximumMessageTime", taskConfig.getMaximumMessageTime().isPresent());
    Assert.assertEquals(topic, taskConfig.getStartSequenceNumbers().getStream());
    Assert.assertEquals(0L, (long) taskConfig.getStartSequenceNumbers().getPartitionSequenceNumberMap().get(0));
    Assert.assertEquals(0L, (long) taskConfig.getStartSequenceNumbers().getPartitionSequenceNumberMap().get(1));
    Assert.assertEquals(0L, (long) taskConfig.getStartSequenceNumbers().getPartitionSequenceNumberMap().get(2));
    Assert.assertEquals(topic, taskConfig.getEndSequenceNumbers().getStream());
    Assert.assertEquals(Long.MAX_VALUE, (long) taskConfig.getEndSequenceNumbers().getPartitionSequenceNumberMap().get(0));
    Assert.assertEquals(Long.MAX_VALUE, (long) taskConfig.getEndSequenceNumbers().getPartitionSequenceNumberMap().get(1));
    Assert.assertEquals(Long.MAX_VALUE, (long) taskConfig.getEndSequenceNumbers().getPartitionSequenceNumberMap().get(2));
    autoscaler.reset();
    autoscaler.stop();
}

Example 3 with KafkaIndexTaskClientFactory

use of org.apache.druid.indexing.kafka.KafkaIndexTaskClientFactory in project druid by druid-io.

the class KafkaSupervisorTest method getTestableSupervisor.

private TestableKafkaSupervisor getTestableSupervisor(int replicas, int taskCount, boolean useEarliestOffset, boolean resetOffsetAutomatically, String duration, Period lateMessageRejectionPeriod, Period earlyMessageRejectionPeriod, boolean suspended, String kafkaHost) {
    final Map<String, Object> consumerProperties = KafkaConsumerConfigs.getConsumerProperties();
    consumerProperties.put("myCustomKey", "myCustomValue");
    consumerProperties.put("bootstrap.servers", kafkaHost);
    KafkaSupervisorIOConfig kafkaSupervisorIOConfig = new KafkaSupervisorIOConfig(topic, INPUT_FORMAT, replicas, taskCount, new Period(duration), consumerProperties, null, KafkaSupervisorIOConfig.DEFAULT_POLL_TIMEOUT_MILLIS, new Period("P1D"), new Period("PT30S"), useEarliestOffset, new Period("PT30M"), lateMessageRejectionPeriod, earlyMessageRejectionPeriod, null);
    KafkaIndexTaskClientFactory taskClientFactory = new KafkaIndexTaskClientFactory(null, null) {

        @Override
        public KafkaIndexTaskClient build(TaskInfoProvider taskInfoProvider, String dataSource, int numThreads, Duration httpTimeout, long numRetries) {
            Assert.assertEquals(TEST_CHAT_THREADS, numThreads);
            Assert.assertEquals(TEST_HTTP_TIMEOUT.toStandardDuration(), httpTimeout);
            Assert.assertEquals(TEST_CHAT_RETRIES, numRetries);
            return taskClient;
        }
    };
    final KafkaSupervisorTuningConfig tuningConfig = new KafkaSupervisorTuningConfig(null, 1000, null, null, 50000, null, new Period("P1Y"), new File("/test"), null, null, null, false, null, resetOffsetAutomatically, null, numThreads, TEST_CHAT_THREADS, TEST_CHAT_RETRIES, TEST_HTTP_TIMEOUT, TEST_SHUTDOWN_TIMEOUT, null, null, null, null, 10);
    return new TestableKafkaSupervisor(taskStorage, taskMaster, indexerMetadataStorageCoordinator, taskClientFactory, OBJECT_MAPPER, new KafkaSupervisorSpec(null, dataSchema, tuningConfig, kafkaSupervisorIOConfig, null, suspended, taskStorage, taskMaster, indexerMetadataStorageCoordinator, taskClientFactory, OBJECT_MAPPER, new NoopServiceEmitter(), new DruidMonitorSchedulerConfig(), rowIngestionMetersFactory, new SupervisorStateManagerConfig()), rowIngestionMetersFactory);
}

Example 4 with KafkaIndexTaskClientFactory

use of org.apache.druid.indexing.kafka.KafkaIndexTaskClientFactory in project druid by druid-io.

the class KafkaSupervisorTest method getTestableSupervisorCustomIsTaskCurrent.

/**
 * Use when you want to mock the return value of SeekableStreamSupervisor#isTaskCurrent()
 */
private TestableKafkaSupervisor getTestableSupervisorCustomIsTaskCurrent(int replicas, int taskCount, boolean useEarliestOffset, String duration, Period lateMessageRejectionPeriod, Period earlyMessageRejectionPeriod, boolean suspended, boolean isTaskCurrentReturn) {
    Map<String, Object> consumerProperties = new HashMap<>();
    consumerProperties.put("myCustomKey", "myCustomValue");
    consumerProperties.put("bootstrap.servers", kafkaHost);
    consumerProperties.put("isolation.level", "read_committed");
    KafkaSupervisorIOConfig kafkaSupervisorIOConfig = new KafkaSupervisorIOConfig(topic, INPUT_FORMAT, replicas, taskCount, new Period(duration), consumerProperties, null, KafkaSupervisorIOConfig.DEFAULT_POLL_TIMEOUT_MILLIS, new Period("P1D"), new Period("PT30S"), useEarliestOffset, new Period("PT30M"), lateMessageRejectionPeriod, earlyMessageRejectionPeriod, null);
    KafkaIndexTaskClientFactory taskClientFactory = new KafkaIndexTaskClientFactory(null, null) {

        @Override
        public KafkaIndexTaskClient build(TaskInfoProvider taskInfoProvider, String dataSource, int numThreads, Duration httpTimeout, long numRetries) {
            Assert.assertEquals(TEST_CHAT_THREADS, numThreads);
            Assert.assertEquals(TEST_HTTP_TIMEOUT.toStandardDuration(), httpTimeout);
            Assert.assertEquals(TEST_CHAT_RETRIES, numRetries);
            return taskClient;
        }
    };
    final KafkaSupervisorTuningConfig tuningConfig = new KafkaSupervisorTuningConfig(null, 1000, null, null, 50000, null, new Period("P1Y"), new File("/test"), null, null, null, false, null, false, null, numThreads, TEST_CHAT_THREADS, TEST_CHAT_RETRIES, TEST_HTTP_TIMEOUT, TEST_SHUTDOWN_TIMEOUT, null, null, null, null, null);
    return new TestableKafkaSupervisorWithCustomIsTaskCurrent(taskStorage, taskMaster, indexerMetadataStorageCoordinator, taskClientFactory, OBJECT_MAPPER, new KafkaSupervisorSpec(null, dataSchema, tuningConfig, kafkaSupervisorIOConfig, null, suspended, taskStorage, taskMaster, indexerMetadataStorageCoordinator, taskClientFactory, OBJECT_MAPPER, new NoopServiceEmitter(), new DruidMonitorSchedulerConfig(), rowIngestionMetersFactory, supervisorConfig), rowIngestionMetersFactory, isTaskCurrentReturn);
}