Search in sources :

Example 1 with LinearShardSpec

use of io.druid.timeline.partition.LinearShardSpec in project druid by druid-io.

the class RealtimeManagerTest method setUp.

@Before
public void setUp() throws Exception {
    final List<TestInputRowHolder> rows = Arrays.asList(makeRow(new DateTime("9000-01-01").getMillis()), makeRow(new ParseException("parse error")), null, makeRow(new DateTime().getMillis()));
    ObjectMapper jsonMapper = new DefaultObjectMapper();
    schema = new DataSchema("test", null, new AggregatorFactory[] { new CountAggregatorFactory("rows") }, new UniformGranularitySpec(Granularities.HOUR, Granularities.NONE, null), jsonMapper);
    schema2 = new DataSchema("testV2", null, new AggregatorFactory[] { new CountAggregatorFactory("rows") }, new UniformGranularitySpec(Granularities.HOUR, Granularities.NONE, null), jsonMapper);
    RealtimeIOConfig ioConfig = new RealtimeIOConfig(new FirehoseFactory() {

        @Override
        public Firehose connect(InputRowParser parser) throws IOException {
            return new TestFirehose(rows.iterator());
        }
    }, new PlumberSchool() {

        @Override
        public Plumber findPlumber(DataSchema schema, RealtimeTuningConfig config, FireDepartmentMetrics metrics) {
            return plumber;
        }
    }, null);
    RealtimeIOConfig ioConfig2 = new RealtimeIOConfig(null, new PlumberSchool() {

        @Override
        public Plumber findPlumber(DataSchema schema, RealtimeTuningConfig config, FireDepartmentMetrics metrics) {
            return plumber2;
        }
    }, new FirehoseFactoryV2() {

        @Override
        public FirehoseV2 connect(InputRowParser parser, Object arg1) throws IOException, ParseException {
            return new TestFirehoseV2(rows.iterator());
        }
    });
    RealtimeTuningConfig tuningConfig = new RealtimeTuningConfig(1, new Period("P1Y"), null, null, null, null, null, null, null, null, 0, 0, null, null);
    plumber = new TestPlumber(new Sink(new Interval("0/P5000Y"), schema, tuningConfig.getShardSpec(), new DateTime().toString(), tuningConfig.getMaxRowsInMemory(), tuningConfig.isReportParseExceptions()));
    realtimeManager = new RealtimeManager(Arrays.<FireDepartment>asList(new FireDepartment(schema, ioConfig, tuningConfig)), null);
    plumber2 = new TestPlumber(new Sink(new Interval("0/P5000Y"), schema2, tuningConfig.getShardSpec(), new DateTime().toString(), tuningConfig.getMaxRowsInMemory(), tuningConfig.isReportParseExceptions()));
    realtimeManager2 = new RealtimeManager(Arrays.<FireDepartment>asList(new FireDepartment(schema2, ioConfig2, tuningConfig)), null);
    tuningConfig_0 = new RealtimeTuningConfig(1, new Period("P1Y"), null, null, null, null, null, new LinearShardSpec(0), null, null, 0, 0, null, null);
    tuningConfig_1 = new RealtimeTuningConfig(1, new Period("P1Y"), null, null, null, null, null, new LinearShardSpec(1), null, null, 0, 0, null, null);
    schema3 = new DataSchema("testing", null, new AggregatorFactory[] { new CountAggregatorFactory("ignore") }, new UniformGranularitySpec(Granularities.HOUR, Granularities.NONE, null), jsonMapper);
    FireDepartment department_0 = new FireDepartment(schema3, ioConfig, tuningConfig_0);
    FireDepartment department_1 = new FireDepartment(schema3, ioConfig2, tuningConfig_1);
    QueryRunnerFactoryConglomerate conglomerate = new QueryRunnerFactoryConglomerate() {

        @Override
        public <T, QueryType extends Query<T>> QueryRunnerFactory<T, QueryType> findFactory(QueryType query) {
            return factory;
        }
    };
    chiefStartedLatch = new CountDownLatch(2);
    RealtimeManager.FireChief fireChief_0 = new RealtimeManager.FireChief(department_0, conglomerate) {

        @Override
        public void run() {
            super.initPlumber();
            chiefStartedLatch.countDown();
        }
    };
    RealtimeManager.FireChief fireChief_1 = new RealtimeManager.FireChief(department_1, conglomerate) {

        @Override
        public void run() {
            super.initPlumber();
            chiefStartedLatch.countDown();
        }
    };
    realtimeManager3 = new RealtimeManager(Arrays.asList(department_0, department_1), conglomerate, ImmutableMap.<String, Map<Integer, RealtimeManager.FireChief>>of("testing", ImmutableMap.of(0, fireChief_0, 1, fireChief_1)));
    startFireChiefWithPartitionNum(fireChief_0, 0);
    startFireChiefWithPartitionNum(fireChief_1, 1);
}
Also used : FirehoseV2(io.druid.data.input.FirehoseV2) RealtimeIOConfig(io.druid.segment.indexing.RealtimeIOConfig) BaseQuery(io.druid.query.BaseQuery) Query(io.druid.query.Query) GroupByQuery(io.druid.query.groupby.GroupByQuery) FirehoseFactory(io.druid.data.input.FirehoseFactory) LinearShardSpec(io.druid.timeline.partition.LinearShardSpec) DateTime(org.joda.time.DateTime) UniformGranularitySpec(io.druid.segment.indexing.granularity.UniformGranularitySpec) QueryRunnerFactoryConglomerate(io.druid.query.QueryRunnerFactoryConglomerate) Sink(io.druid.segment.realtime.plumber.Sink) DefaultObjectMapper(io.druid.jackson.DefaultObjectMapper) ObjectMapper(com.fasterxml.jackson.databind.ObjectMapper) Firehose(io.druid.data.input.Firehose) Period(org.joda.time.Period) IOException(java.io.IOException) PlumberSchool(io.druid.segment.realtime.plumber.PlumberSchool) CountAggregatorFactory(io.druid.query.aggregation.CountAggregatorFactory) AggregatorFactory(io.druid.query.aggregation.AggregatorFactory) LongSumAggregatorFactory(io.druid.query.aggregation.LongSumAggregatorFactory) RealtimeTuningConfig(io.druid.segment.indexing.RealtimeTuningConfig) CountDownLatch(java.util.concurrent.CountDownLatch) DataSchema(io.druid.segment.indexing.DataSchema) CountAggregatorFactory(io.druid.query.aggregation.CountAggregatorFactory) Plumber(io.druid.segment.realtime.plumber.Plumber) ParseException(io.druid.java.util.common.parsers.ParseException) DefaultObjectMapper(io.druid.jackson.DefaultObjectMapper) InputRowParser(io.druid.data.input.impl.InputRowParser) Map(java.util.Map) ImmutableMap(com.google.common.collect.ImmutableMap) FirehoseFactoryV2(io.druid.data.input.FirehoseFactoryV2) Interval(org.joda.time.Interval) Before(org.junit.Before)

Example 2 with LinearShardSpec

use of io.druid.timeline.partition.LinearShardSpec in project druid by druid-io.

the class DruidCoordinatorSegmentMergerTest method testMergeMixedShardSpecs.

@Test
public void testMergeMixedShardSpecs() {
    final List<DataSegment> segments = ImmutableList.of(DataSegment.builder().dataSource("foo").interval(new Interval("2012-01-01/P1D")).version("1").build(), DataSegment.builder().dataSource("foo").interval(new Interval("2012-01-02/P1D")).version("1").build(), DataSegment.builder().dataSource("foo").interval(new Interval("2012-01-03/P1D")).version("1").shardSpec(new LinearShardSpec(1500)).build(), DataSegment.builder().dataSource("foo").interval(new Interval("2012-01-04/P1D")).version("1").build(), DataSegment.builder().dataSource("foo").interval(new Interval("2012-01-05/P1D")).version("1").build());
    Assert.assertEquals(ImmutableList.of(ImmutableList.of(segments.get(0), segments.get(1)), ImmutableList.of(segments.get(3), segments.get(4))), merge(segments));
}
Also used : LinearShardSpec(io.druid.timeline.partition.LinearShardSpec) DataSegment(io.druid.timeline.DataSegment) Interval(org.joda.time.Interval) Test(org.junit.Test)

Example 3 with LinearShardSpec

use of io.druid.timeline.partition.LinearShardSpec in project druid by druid-io.

the class DruidCoordinatorSegmentMergerTest method testMergeLinearShardSpecs.

@Test
public void testMergeLinearShardSpecs() {
    final List<DataSegment> segments = ImmutableList.of(DataSegment.builder().dataSource("foo").interval(new Interval("2012-01-01/P1D")).version("1").shardSpec(new LinearShardSpec(1)).build(), DataSegment.builder().dataSource("foo").interval(new Interval("2012-01-02/P1D")).version("1").shardSpec(new LinearShardSpec(7)).build(), DataSegment.builder().dataSource("foo").interval(new Interval("2012-01-03/P1D")).version("1").shardSpec(new LinearShardSpec(1500)).build());
    Assert.assertEquals(ImmutableList.of(), merge(segments));
}
Also used : LinearShardSpec(io.druid.timeline.partition.LinearShardSpec) DataSegment(io.druid.timeline.DataSegment) Interval(org.joda.time.Interval) Test(org.junit.Test)

Example 4 with LinearShardSpec

use of io.druid.timeline.partition.LinearShardSpec in project druid by druid-io.

the class SameIntervalMergeTaskTest method runTask.

private List<DataSegment> runTask(final SameIntervalMergeTask mergeTask, final String version) throws Exception {
    boolean isReady = mergeTask.isReady(new TaskActionClient() {

        @Override
        public <RetType> RetType submit(TaskAction<RetType> taskAction) throws IOException {
            if (taskAction instanceof LockTryAcquireAction) {
                // the lock of this interval is required
                Assert.assertEquals(mergeTask.getInterval(), ((LockTryAcquireAction) taskAction).getInterval());
                isRedayCountDown.countDown();
                taskLock = new TaskLock(mergeTask.getGroupId(), mergeTask.getDataSource(), mergeTask.getInterval(), version);
                return (RetType) taskLock;
            }
            return null;
        }
    });
    // ensure LockTryAcquireAction is submitted
    Assert.assertTrue(isReady);
    final List<DataSegment> segments = Lists.newArrayList();
    mergeTask.run(new TaskToolbox(null, null, new TaskActionClient() {

        @Override
        public <RetType> RetType submit(TaskAction<RetType> taskAction) throws IOException {
            if (taskAction instanceof LockListAction) {
                Assert.assertNotNull("taskLock should be acquired before list", taskLock);
                return (RetType) Arrays.asList(taskLock);
            }
            if (taskAction instanceof SegmentListUsedAction) {
                List<DataSegment> segments = ImmutableList.of(DataSegment.builder().dataSource(mergeTask.getDataSource()).interval(new Interval("2010-01-01/PT1H")).version("oldVersion").shardSpec(new LinearShardSpec(0)).build(), DataSegment.builder().dataSource(mergeTask.getDataSource()).interval(new Interval("2010-01-01/PT1H")).version("oldVersion").shardSpec(new LinearShardSpec(0)).build(), DataSegment.builder().dataSource(mergeTask.getDataSource()).interval(new Interval("2010-01-01/PT2H")).version("oldVersion").shardSpec(new LinearShardSpec(0)).build());
                return (RetType) segments;
            }
            if (taskAction instanceof SegmentInsertAction) {
                publishCountDown.countDown();
                return null;
            }
            return null;
        }
    }, new NoopServiceEmitter(), new DataSegmentPusher() {

        @Deprecated
        @Override
        public String getPathForHadoop(String dataSource) {
            return getPathForHadoop();
        }

        @Override
        public String getPathForHadoop() {
            return null;
        }

        @Override
        public DataSegment push(File file, DataSegment segment) throws IOException {
            // the merged segment is pushed to storage
            segments.add(segment);
            return segment;
        }
    }, null, null, null, null, null, null, null, null, new SegmentLoader() {

        @Override
        public boolean isSegmentLoaded(DataSegment segment) throws SegmentLoadingException {
            return false;
        }

        @Override
        public Segment getSegment(DataSegment segment) throws SegmentLoadingException {
            return null;
        }

        @Override
        public File getSegmentFiles(DataSegment segment) throws SegmentLoadingException {
            // dummy file to represent the downloaded segment's dir
            return new File("" + segment.getShardSpec().getPartitionNum());
        }

        @Override
        public void cleanup(DataSegment segment) throws SegmentLoadingException {
        }
    }, jsonMapper, temporaryFolder.newFolder(), EasyMock.createMock(IndexMerger.class), indexIO, null, null, EasyMock.createMock(IndexMergerV9.class)));
    return segments;
}
Also used : LockListAction(io.druid.indexing.common.actions.LockListAction) DataSegmentPusher(io.druid.segment.loading.DataSegmentPusher) LockTryAcquireAction(io.druid.indexing.common.actions.LockTryAcquireAction) TaskAction(io.druid.indexing.common.actions.TaskAction) LinearShardSpec(io.druid.timeline.partition.LinearShardSpec) NoopServiceEmitter(io.druid.server.metrics.NoopServiceEmitter) IOException(java.io.IOException) DataSegment(io.druid.timeline.DataSegment) SegmentLoader(io.druid.segment.loading.SegmentLoader) TaskToolbox(io.druid.indexing.common.TaskToolbox) TaskActionClient(io.druid.indexing.common.actions.TaskActionClient) TaskLock(io.druid.indexing.common.TaskLock) SegmentInsertAction(io.druid.indexing.common.actions.SegmentInsertAction) SegmentListUsedAction(io.druid.indexing.common.actions.SegmentListUsedAction) File(java.io.File) Interval(org.joda.time.Interval)

Example 5 with LinearShardSpec

use of io.druid.timeline.partition.LinearShardSpec in project druid by druid-io.

the class SqlBenchmark method setup.

@Setup(Level.Trial)
public void setup() throws Exception {
    tmpDir = Files.createTempDir();
    log.info("Starting benchmark setup using tmpDir[%s], rows[%,d].", tmpDir, rowsPerSegment);
    if (ComplexMetrics.getSerdeForType("hyperUnique") == null) {
        ComplexMetrics.registerSerde("hyperUnique", new HyperUniquesSerde(HyperLogLogHash.getDefault()));
    }
    final BenchmarkSchemaInfo schemaInfo = BenchmarkSchemas.SCHEMA_MAP.get("basic");
    final BenchmarkDataGenerator dataGenerator = new BenchmarkDataGenerator(schemaInfo.getColumnSchemas(), RNG_SEED + 1, schemaInfo.getDataInterval(), rowsPerSegment);
    final List<InputRow> rows = Lists.newArrayList();
    for (int i = 0; i < rowsPerSegment; i++) {
        final InputRow row = dataGenerator.nextRow();
        if (i % 20000 == 0) {
            log.info("%,d/%,d rows generated.", i, rowsPerSegment);
        }
        rows.add(row);
    }
    log.info("%,d/%,d rows generated.", rows.size(), rowsPerSegment);
    final PlannerConfig plannerConfig = new PlannerConfig();
    final QueryRunnerFactoryConglomerate conglomerate = CalciteTests.queryRunnerFactoryConglomerate();
    final QueryableIndex index = IndexBuilder.create().tmpDir(new File(tmpDir, "1")).indexMerger(TestHelper.getTestIndexMergerV9()).rows(rows).buildMMappedIndex();
    this.walker = new SpecificSegmentsQuerySegmentWalker(conglomerate).add(DataSegment.builder().dataSource("foo").interval(index.getDataInterval()).version("1").shardSpec(new LinearShardSpec(0)).build(), index);
    final Map<String, Table> tableMap = ImmutableMap.<String, Table>of("foo", new DruidTable(new TableDataSource("foo"), RowSignature.builder().add("__time", ValueType.LONG).add("dimSequential", ValueType.STRING).add("dimZipf", ValueType.STRING).add("dimUniform", ValueType.STRING).build()));
    final Schema druidSchema = new AbstractSchema() {

        @Override
        protected Map<String, Table> getTableMap() {
            return tableMap;
        }
    };
    plannerFactory = new PlannerFactory(Calcites.createRootSchema(druidSchema), walker, CalciteTests.createOperatorTable(), plannerConfig);
    groupByQuery = GroupByQuery.builder().setDataSource("foo").setInterval(new Interval(JodaUtils.MIN_INSTANT, JodaUtils.MAX_INSTANT)).setDimensions(Arrays.<DimensionSpec>asList(new DefaultDimensionSpec("dimZipf", "d0"), new DefaultDimensionSpec("dimSequential", "d1"))).setAggregatorSpecs(Arrays.<AggregatorFactory>asList(new CountAggregatorFactory("c"))).setGranularity(Granularities.ALL).build();
    sqlQuery = "SELECT\n" + "  dimZipf AS d0," + "  dimSequential AS d1,\n" + "  COUNT(*) AS c\n" + "FROM druid.foo\n" + "GROUP BY dimZipf, dimSequential";
}
Also used : DruidTable(io.druid.sql.calcite.table.DruidTable) Table(org.apache.calcite.schema.Table) LinearShardSpec(io.druid.timeline.partition.LinearShardSpec) Schema(org.apache.calcite.schema.Schema) AbstractSchema(org.apache.calcite.schema.impl.AbstractSchema) BenchmarkDataGenerator(io.druid.benchmark.datagen.BenchmarkDataGenerator) HyperUniquesSerde(io.druid.query.aggregation.hyperloglog.HyperUniquesSerde) DruidTable(io.druid.sql.calcite.table.DruidTable) CountAggregatorFactory(io.druid.query.aggregation.CountAggregatorFactory) AggregatorFactory(io.druid.query.aggregation.AggregatorFactory) DefaultDimensionSpec(io.druid.query.dimension.DefaultDimensionSpec) QueryRunnerFactoryConglomerate(io.druid.query.QueryRunnerFactoryConglomerate) SpecificSegmentsQuerySegmentWalker(io.druid.sql.calcite.util.SpecificSegmentsQuerySegmentWalker) TableDataSource(io.druid.query.TableDataSource) CountAggregatorFactory(io.druid.query.aggregation.CountAggregatorFactory) AbstractSchema(org.apache.calcite.schema.impl.AbstractSchema) QueryableIndex(io.druid.segment.QueryableIndex) BenchmarkSchemaInfo(io.druid.benchmark.datagen.BenchmarkSchemaInfo) PlannerConfig(io.druid.sql.calcite.planner.PlannerConfig) InputRow(io.druid.data.input.InputRow) PlannerFactory(io.druid.sql.calcite.planner.PlannerFactory) File(java.io.File) Interval(org.joda.time.Interval) Setup(org.openjdk.jmh.annotations.Setup)

Aggregations

LinearShardSpec (io.druid.timeline.partition.LinearShardSpec)13 Interval (org.joda.time.Interval)9 CountAggregatorFactory (io.druid.query.aggregation.CountAggregatorFactory)5 DataSegment (io.druid.timeline.DataSegment)5 Test (org.junit.Test)5 AggregatorFactory (io.druid.query.aggregation.AggregatorFactory)4 QueryableIndex (io.druid.segment.QueryableIndex)4 SegmentIdentifier (io.druid.segment.realtime.appenderator.SegmentIdentifier)4 File (java.io.File)4 LongSumAggregatorFactory (io.druid.query.aggregation.LongSumAggregatorFactory)3 ObjectMapper (com.fasterxml.jackson.databind.ObjectMapper)2 SegmentListUsedAction (io.druid.indexing.common.actions.SegmentListUsedAction)2 TaskActionClient (io.druid.indexing.common.actions.TaskActionClient)2 QueryRunnerFactoryConglomerate (io.druid.query.QueryRunnerFactoryConglomerate)2 DoubleSumAggregatorFactory (io.druid.query.aggregation.DoubleSumAggregatorFactory)2 IndexBuilder (io.druid.segment.IndexBuilder)2 DataSchema (io.druid.segment.indexing.DataSchema)2 RealtimeTuningConfig (io.druid.segment.indexing.RealtimeTuningConfig)2 SpecificSegmentsQuerySegmentWalker (io.druid.sql.calcite.util.SpecificSegmentsQuerySegmentWalker)2 IOException (java.io.IOException)2