use of org.apache.druid.indexing.common.IngestionStatsAndErrorsTaskReportData in project druid by druid-io.
the class AbstractITBatchIndexTest method submitTaskAndWait.
protected void submitTaskAndWait(String taskSpec, String dataSourceName, boolean waitForNewVersion, boolean waitForSegmentsToLoad, Pair<Boolean, Boolean> segmentAvailabilityConfirmationPair) {
final List<DataSegment> oldVersions = waitForNewVersion ? coordinator.getAvailableSegments(dataSourceName) : null;
long startSubTaskCount = -1;
final boolean assertRunsSubTasks = taskSpec.contains("index_parallel");
if (assertRunsSubTasks) {
startSubTaskCount = countCompleteSubTasks(dataSourceName, !taskSpec.contains("dynamic"));
}
final String taskID = indexer.submitTask(taskSpec);
LOG.info("TaskID for loading index task %s", taskID);
indexer.waitUntilTaskCompletes(taskID);
if (assertRunsSubTasks) {
final boolean perfectRollup = !taskSpec.contains("dynamic");
final long newSubTasks = countCompleteSubTasks(dataSourceName, perfectRollup) - startSubTaskCount;
Assert.assertTrue(newSubTasks > 0, StringUtils.format("The supervisor task[%s] didn't create any sub tasks. Was it executed in the parallel mode?", taskID));
}
if (segmentAvailabilityConfirmationPair.lhs != null && segmentAvailabilityConfirmationPair.lhs) {
TaskReport reportRaw = indexer.getTaskReport(taskID).get("ingestionStatsAndErrors");
IngestionStatsAndErrorsTaskReport report = (IngestionStatsAndErrorsTaskReport) reportRaw;
IngestionStatsAndErrorsTaskReportData reportData = (IngestionStatsAndErrorsTaskReportData) report.getPayload();
// Confirm that the task waited longer than 0ms for the task to complete.
Assert.assertTrue(reportData.getSegmentAvailabilityWaitTimeMs() > 0);
// Make sure that the result of waiting for segments to load matches the expected result
if (segmentAvailabilityConfirmationPair.rhs != null) {
Assert.assertEquals(Boolean.valueOf(reportData.isSegmentAvailabilityConfirmed()), segmentAvailabilityConfirmationPair.rhs);
}
}
// original segments have loaded.
if (waitForNewVersion) {
ITRetryUtil.retryUntilTrue(() -> {
final VersionedIntervalTimeline<String, DataSegment> timeline = VersionedIntervalTimeline.forSegments(coordinator.getAvailableSegments(dataSourceName));
final List<TimelineObjectHolder<String, DataSegment>> holders = timeline.lookup(Intervals.ETERNITY);
return FluentIterable.from(holders).transformAndConcat(TimelineObjectHolder::getObject).anyMatch(chunk -> FluentIterable.from(oldVersions).anyMatch(oldSegment -> chunk.getObject().overshadows(oldSegment)));
}, "See a new version");
}
if (waitForSegmentsToLoad) {
ITRetryUtil.retryUntilTrue(() -> coordinator.areSegmentsLoaded(dataSourceName), "Segment Load");
}
}
use of org.apache.druid.indexing.common.IngestionStatsAndErrorsTaskReportData in project druid by druid-io.
the class IndexTaskTest method testCsvWithHeaderOfEmptyTimestamp.
@Test
public void testCsvWithHeaderOfEmptyTimestamp() throws Exception {
final File tmpDir = temporaryFolder.newFolder();
final File tmpFile = File.createTempFile("druid", "index", tmpDir);
try (BufferedWriter writer = Files.newWriter(tmpFile, StandardCharsets.UTF_8)) {
writer.write(",,\n");
writer.write("2014-01-01T00:00:10Z,a,1\n");
}
final List<String> columns = Arrays.asList("ts", "", "");
// report parse exception
final IndexTuningConfig tuningConfig = createTuningConfig(2, null, null, null, null, false, true);
final IndexIngestionSpec ingestionSpec;
List<String> expectedMessages;
if (useInputFormatApi) {
ingestionSpec = createIngestionSpec(jsonMapper, tmpDir, DEFAULT_TIMESTAMP_SPEC, DimensionsSpec.EMPTY, new CsvInputFormat(columns, null, null, true, 0), null, null, tuningConfig, false, false);
expectedMessages = ImmutableList.of(StringUtils.format("Timestamp[null] is unparseable! Event: {column_1=2014-01-01T00:00:10Z, column_2=a, column_3=1} (Path: %s, Record: 1, Line: 2)", tmpFile.toURI()));
} else {
ingestionSpec = createIngestionSpec(jsonMapper, tmpDir, new CSVParseSpec(DEFAULT_TIMESTAMP_SPEC, DimensionsSpec.EMPTY, null, columns, true, 0), null, null, tuningConfig, false, false);
expectedMessages = ImmutableList.of("Timestamp[null] is unparseable! Event: {column_1=2014-01-01T00:00:10Z, column_2=a, column_3=1}");
}
IndexTask indexTask = new IndexTask(null, null, ingestionSpec, null);
TaskStatus status = runTask(indexTask).lhs;
Assert.assertEquals(TaskState.FAILED, status.getStatusCode());
checkTaskStatusErrorMsgForParseExceptionsExceeded(status);
IngestionStatsAndErrorsTaskReportData reportData = getTaskReportData();
List<LinkedHashMap> parseExceptionReports = (List<LinkedHashMap>) reportData.getUnparseableEvents().get(RowIngestionMeters.BUILD_SEGMENTS);
List<String> actualMessages = parseExceptionReports.stream().map((r) -> {
return ((List<String>) r.get("details")).get(0);
}).collect(Collectors.toList());
Assert.assertEquals(expectedMessages, actualMessages);
List<String> expectedInputs = ImmutableList.of("{column_1=2014-01-01T00:00:10Z, column_2=a, column_3=1}");
List<String> actualInputs = parseExceptionReports.stream().map((r) -> {
return (String) r.get("input");
}).collect(Collectors.toList());
Assert.assertEquals(expectedInputs, actualInputs);
}
use of org.apache.druid.indexing.common.IngestionStatsAndErrorsTaskReportData in project druid by druid-io.
the class IndexTaskTest method testMultipleParseExceptionsFailure.
@Test
public void testMultipleParseExceptionsFailure() throws Exception {
final File tmpDir = temporaryFolder.newFolder();
final File tmpFile = File.createTempFile("druid", "index", tmpDir);
try (BufferedWriter writer = Files.newWriter(tmpFile, StandardCharsets.UTF_8)) {
writer.write("time,dim,dimLong,dimFloat,val\n");
// unparseable
writer.write("unparseable,a,2,3.0,1\n");
// valid row
writer.write("2014-01-01T00:00:10Z,a,2,3.0,1\n");
// unparseable
writer.write("9.0,a,2,3.0,1\n");
// thrown away
writer.write("3014-03-01T00:00:10Z,outsideofinterval,2,3.0,1\n");
// unparseable
writer.write("99999999999-01-01T00:00:10Z,b,2,3.0,1\n");
}
// Allow up to 3 parse exceptions, and save up to 2 parse exceptions
final IndexTuningConfig tuningConfig = new IndexTuningConfig(null, null, null, null, null, null, null, null, null, null, new DynamicPartitionsSpec(2, null), INDEX_SPEC, null, null, false, false, null, null, null, true, 2, 5, null, null);
final TimestampSpec timestampSpec = new TimestampSpec("time", "auto", null);
final DimensionsSpec dimensionsSpec = new DimensionsSpec(Arrays.asList(new StringDimensionSchema("dim"), new LongDimensionSchema("dimLong"), new FloatDimensionSchema("dimFloat")));
final List<String> columns = Arrays.asList("time", "dim", "dimLong", "dimFloat", "val");
final IndexIngestionSpec ingestionSpec;
List<String> expectedMessages;
if (useInputFormatApi) {
ingestionSpec = createIngestionSpec(jsonMapper, tmpDir, timestampSpec, dimensionsSpec, new CsvInputFormat(columns, null, null, true, 0), null, null, tuningConfig, false, false);
expectedMessages = Arrays.asList(StringUtils.format("Timestamp[99999999999-01-01T00:00:10Z] is unparseable! Event: {time=99999999999-01-01T00:00:10Z, dim=b, dimLong=2, dimFloat=3.0, val=1} (Path: %s, Record: 3, Line: 6)", tmpFile.toURI()), StringUtils.format("Timestamp[9.0] is unparseable! Event: {time=9.0, dim=a, dimLong=2, dimFloat=3.0, val=1} (Path: %s, Record: 2, Line: 4)", tmpFile.toURI()), StringUtils.format("Timestamp[unparseable] is unparseable! Event: {time=unparseable, dim=a, dimLong=2, dimFloat=3.0, val=1} (Path: %s, Record: 1, Line: 2)", tmpFile.toURI()));
} else {
ingestionSpec = createIngestionSpec(jsonMapper, tmpDir, new CSVParseSpec(timestampSpec, dimensionsSpec, null, columns, true, 0), null, null, tuningConfig, false, false);
expectedMessages = Arrays.asList("Timestamp[99999999999-01-01T00:00:10Z] is unparseable! Event: {time=99999999999-01-01T00:00:10Z, dim=b, dimLong=2, dimFloat=3.0, val=1}", "Timestamp[9.0] is unparseable! Event: {time=9.0, dim=a, dimLong=2, dimFloat=3.0, val=1}", "Timestamp[unparseable] is unparseable! Event: {time=unparseable, dim=a, dimLong=2, dimFloat=3.0, val=1}");
}
IndexTask indexTask = new IndexTask(null, null, ingestionSpec, null);
TaskStatus status = runTask(indexTask).lhs;
Assert.assertEquals(TaskState.FAILED, status.getStatusCode());
checkTaskStatusErrorMsgForParseExceptionsExceeded(status);
IngestionStatsAndErrorsTaskReportData reportData = getTaskReportData();
Map<String, Object> expectedMetrics = ImmutableMap.of(RowIngestionMeters.DETERMINE_PARTITIONS, ImmutableMap.of(RowIngestionMeters.PROCESSED_WITH_ERROR, 0, RowIngestionMeters.PROCESSED, 0, RowIngestionMeters.UNPARSEABLE, 0, RowIngestionMeters.THROWN_AWAY, 0), RowIngestionMeters.BUILD_SEGMENTS, ImmutableMap.of(RowIngestionMeters.PROCESSED_WITH_ERROR, 0, RowIngestionMeters.PROCESSED, 1, RowIngestionMeters.UNPARSEABLE, 3, RowIngestionMeters.THROWN_AWAY, useInputFormatApi ? 1 : 2));
Assert.assertEquals(expectedMetrics, reportData.getRowStats());
List<LinkedHashMap> parseExceptionReports = (List<LinkedHashMap>) reportData.getUnparseableEvents().get(RowIngestionMeters.BUILD_SEGMENTS);
List<String> actualMessages = parseExceptionReports.stream().map((r) -> {
return ((List<String>) r.get("details")).get(0);
}).collect(Collectors.toList());
Assert.assertEquals(expectedMessages, actualMessages);
List<String> expectedInputs = Arrays.asList("{time=99999999999-01-01T00:00:10Z, dim=b, dimLong=2, dimFloat=3.0, val=1}", "{time=9.0, dim=a, dimLong=2, dimFloat=3.0, val=1}", "{time=unparseable, dim=a, dimLong=2, dimFloat=3.0, val=1}");
List<String> actualInputs = parseExceptionReports.stream().map((r) -> {
return (String) r.get("input");
}).collect(Collectors.toList());
Assert.assertEquals(expectedInputs, actualInputs);
}
use of org.apache.druid.indexing.common.IngestionStatsAndErrorsTaskReportData in project druid by druid-io.
the class IndexTaskTest method testReportParseException.
@Test
public void testReportParseException() throws Exception {
final File tmpDir = temporaryFolder.newFolder();
final File tmpFile = File.createTempFile("druid", "index", tmpDir);
try (BufferedWriter writer = Files.newWriter(tmpFile, StandardCharsets.UTF_8)) {
writer.write("time,d,val\n");
writer.write("unparseable,a,1\n");
writer.write("2014-01-01T00:00:10Z,a,1\n");
}
final TimestampSpec timestampSpec = new TimestampSpec("time", "auto", null);
final List<String> columns = Arrays.asList("time", "dim", "val");
// report parse exception
final IndexTuningConfig tuningConfig = createTuningConfig(2, null, null, null, null, false, true);
final IndexIngestionSpec indexIngestionSpec;
List<String> expectedMessages;
if (useInputFormatApi) {
indexIngestionSpec = createIngestionSpec(jsonMapper, tmpDir, timestampSpec, DimensionsSpec.EMPTY, new CsvInputFormat(columns, null, null, true, 0), null, null, tuningConfig, false, false);
expectedMessages = ImmutableList.of(StringUtils.format("Timestamp[unparseable] is unparseable! Event: {time=unparseable, d=a, val=1} (Path: %s, Record: 1, Line: 2)", tmpFile.toURI()));
} else {
indexIngestionSpec = createIngestionSpec(jsonMapper, tmpDir, new CSVParseSpec(timestampSpec, DimensionsSpec.EMPTY, null, columns, true, 0), null, null, tuningConfig, false, false);
expectedMessages = ImmutableList.of("Timestamp[unparseable] is unparseable! Event: {time=unparseable, d=a, val=1}");
}
IndexTask indexTask = new IndexTask(null, null, indexIngestionSpec, null);
TaskStatus status = runTask(indexTask).lhs;
Assert.assertEquals(TaskState.FAILED, status.getStatusCode());
checkTaskStatusErrorMsgForParseExceptionsExceeded(status);
IngestionStatsAndErrorsTaskReportData reportData = getTaskReportData();
List<LinkedHashMap> parseExceptionReports = (List<LinkedHashMap>) reportData.getUnparseableEvents().get(RowIngestionMeters.BUILD_SEGMENTS);
List<String> actualMessages = parseExceptionReports.stream().map((r) -> {
return ((List<String>) r.get("details")).get(0);
}).collect(Collectors.toList());
Assert.assertEquals(expectedMessages, actualMessages);
List<String> expectedInputs = ImmutableList.of("{time=unparseable, d=a, val=1}");
List<String> actualInputs = parseExceptionReports.stream().map((r) -> {
return (String) r.get("input");
}).collect(Collectors.toList());
Assert.assertEquals(expectedInputs, actualInputs);
}
use of org.apache.druid.indexing.common.IngestionStatsAndErrorsTaskReportData in project druid by druid-io.
the class IndexTaskTest method testMultipleParseExceptionsSuccess.
@Test
public void testMultipleParseExceptionsSuccess() throws Exception {
final File tmpDir = temporaryFolder.newFolder();
final File tmpFile = File.createTempFile("druid", "index", tmpDir);
try (BufferedWriter writer = Files.newWriter(tmpFile, StandardCharsets.UTF_8)) {
// unparseable time
writer.write("{\"time\":\"unparseable\",\"dim\":\"a\",\"dimLong\":2,\"dimFloat\":3.0,\"val\":1}\n");
// valid row
writer.write("{\"time\":\"2014-01-01T00:00:10Z\",\"dim\":\"a\",\"dimLong\":2,\"dimFloat\":3.0,\"val\":1}\n");
writer.write(// row with invalid long dimension
"{\"time\":\"2014-01-01T00:00:10Z\",\"dim\":\"b\",\"dimLong\":\"notnumber\",\"dimFloat\":3.0,\"val\":1}\n");
writer.write(// row with invalid float dimension
"{\"time\":\"2014-01-01T00:00:10Z\",\"dim\":\"b\",\"dimLong\":2,\"dimFloat\":\"notnumber\",\"val\":1}\n");
writer.write(// row with invalid metric
"{\"time\":\"2014-01-01T00:00:10Z\",\"dim\":\"b\",\"dimLong\":2,\"dimFloat\":4.0,\"val\":\"notnumber\"}\n");
// invalid JSON
writer.write("{\"time\":9.0x,\"dim\":\"a\",\"dimLong\":2,\"dimFloat\":3.0,\"val\":1}\n");
writer.write(// thrown away
"{\"time\":\"3014-03-01T00:00:10Z\",\"dim\":\"outsideofinterval\",\"dimLong\":2,\"dimFloat\":3.0,\"val\":1}\n");
// unparseable time
writer.write("{\"time\":\"99999999999-01-01T00:00:10Z\",\"dim\":\"b\",\"dimLong\":2,\"dimFloat\":3.0,\"val\":1}\n");
// invalid JSON
writer.write("this is not JSON\n");
}
final IndexTuningConfig tuningConfig = new IndexTuningConfig(null, null, null, null, null, null, null, null, null, null, new HashedPartitionsSpec(2, null, null), INDEX_SPEC, null, null, true, false, null, null, null, true, 7, 7, null, null);
final TimestampSpec timestampSpec = new TimestampSpec("time", "auto", null);
final DimensionsSpec dimensionsSpec = new DimensionsSpec(Arrays.asList(new StringDimensionSchema("dim"), new LongDimensionSchema("dimLong"), new FloatDimensionSchema("dimFloat")));
final IndexIngestionSpec ingestionSpec;
if (useInputFormatApi) {
ingestionSpec = createIngestionSpec(jsonMapper, tmpDir, timestampSpec, dimensionsSpec, new JsonInputFormat(null, null, null), null, null, tuningConfig, false, false);
} else {
ingestionSpec = createIngestionSpec(jsonMapper, tmpDir, new JSONParseSpec(timestampSpec, dimensionsSpec, null, null, null), null, null, tuningConfig, false, false);
}
IndexTask indexTask = new IndexTask(null, null, ingestionSpec, null);
TaskStatus status = runTask(indexTask).lhs;
Assert.assertEquals(TaskState.SUCCESS, status.getStatusCode());
Assert.assertEquals(null, status.getErrorMsg());
IngestionStatsAndErrorsTaskReportData reportData = getTaskReportData();
Map<String, Object> expectedMetrics = ImmutableMap.of(RowIngestionMeters.DETERMINE_PARTITIONS, ImmutableMap.of(RowIngestionMeters.PROCESSED_WITH_ERROR, 0, RowIngestionMeters.PROCESSED, 4, RowIngestionMeters.UNPARSEABLE, 4, RowIngestionMeters.THROWN_AWAY, 1), RowIngestionMeters.BUILD_SEGMENTS, ImmutableMap.of(RowIngestionMeters.PROCESSED_WITH_ERROR, 3, RowIngestionMeters.PROCESSED, 1, RowIngestionMeters.UNPARSEABLE, 4, RowIngestionMeters.THROWN_AWAY, 1));
Assert.assertEquals(expectedMetrics, reportData.getRowStats());
List<LinkedHashMap> parseExceptionReports = (List<LinkedHashMap>) reportData.getUnparseableEvents().get(RowIngestionMeters.BUILD_SEGMENTS);
List<String> expectedMessages;
if (useInputFormatApi) {
expectedMessages = Arrays.asList(StringUtils.format("Unable to parse row [this is not JSON] (Path: %s, Record: 6, Line: 9)", tmpFile.toURI()), StringUtils.format("Timestamp[99999999999-01-01T00:00:10Z] is unparseable! Event: {time=99999999999-01-01T00:00:10Z, dim=b, dimLong=2, dimFloat=3.0, val=1} (Path: %s, Record: 6, Line: 8)", tmpFile.toURI()), StringUtils.format("Unable to parse row [{\"time\":9.0x,\"dim\":\"a\",\"dimLong\":2,\"dimFloat\":3.0,\"val\":1}] (Path: %s, Record: 5, Line: 6)", tmpFile.toURI()), "Unable to parse value[notnumber] for field[val]", "could not convert value [notnumber] to float", "could not convert value [notnumber] to long", StringUtils.format("Timestamp[unparseable] is unparseable! Event: {time=unparseable, dim=a, dimLong=2, dimFloat=3.0, val=1} (Path: %s, Record: 1, Line: 1)", tmpFile.toURI()));
} else {
expectedMessages = Arrays.asList("Unable to parse row [this is not JSON]", "Timestamp[99999999999-01-01T00:00:10Z] is unparseable! Event: {time=99999999999-01-01T00:00:10Z, dim=b, dimLong=2, dimFloat=3.0, val=1}", "Unable to parse row [{\"time\":9.0x,\"dim\":\"a\",\"dimLong\":2,\"dimFloat\":3.0,\"val\":1}]", "Unable to parse value[notnumber] for field[val]", "could not convert value [notnumber] to float", "could not convert value [notnumber] to long", "Timestamp[unparseable] is unparseable! Event: {time=unparseable, dim=a, dimLong=2, dimFloat=3.0, val=1}");
}
List<String> actualMessages = parseExceptionReports.stream().map((r) -> {
return ((List<String>) r.get("details")).get(0);
}).collect(Collectors.toList());
Assert.assertEquals(expectedMessages, actualMessages);
List<String> expectedInputs = Arrays.asList("this is not JSON", "{time=99999999999-01-01T00:00:10Z, dim=b, dimLong=2, dimFloat=3.0, val=1}", "{\"time\":9.0x,\"dim\":\"a\",\"dimLong\":2,\"dimFloat\":3.0,\"val\":1}", "{time=2014-01-01T00:00:10Z, dim=b, dimLong=2, dimFloat=4.0, val=notnumber}", "{time=2014-01-01T00:00:10Z, dim=b, dimLong=2, dimFloat=notnumber, val=1}", "{time=2014-01-01T00:00:10Z, dim=b, dimLong=notnumber, dimFloat=3.0, val=1}", "{time=unparseable, dim=a, dimLong=2, dimFloat=3.0, val=1}");
List<String> actualInputs = parseExceptionReports.stream().map((r) -> {
return (String) r.get("input");
}).collect(Collectors.toList());
Assert.assertEquals(expectedInputs, actualInputs);
parseExceptionReports = (List<LinkedHashMap>) reportData.getUnparseableEvents().get(RowIngestionMeters.DETERMINE_PARTITIONS);
if (useInputFormatApi) {
expectedMessages = Arrays.asList(StringUtils.format("Unable to parse row [this is not JSON] (Path: %s, Record: 6, Line: 9)", tmpFile.toURI()), StringUtils.format("Timestamp[99999999999-01-01T00:00:10Z] is unparseable! Event: {time=99999999999-01-01T00:00:10Z, dim=b, dimLong=2, dimFloat=3.0, val=1} (Path: %s, Record: 6, Line: 8)", tmpFile.toURI()), StringUtils.format("Unable to parse row [{\"time\":9.0x,\"dim\":\"a\",\"dimLong\":2,\"dimFloat\":3.0,\"val\":1}] (Path: %s, Record: 5, Line: 6)", tmpFile.toURI()), StringUtils.format("Timestamp[unparseable] is unparseable! Event: {time=unparseable, dim=a, dimLong=2, dimFloat=3.0, val=1} (Path: %s, Record: 1, Line: 1)", tmpFile.toURI()));
} else {
expectedMessages = Arrays.asList("Unable to parse row [this is not JSON]", "Timestamp[99999999999-01-01T00:00:10Z] is unparseable! Event: {time=99999999999-01-01T00:00:10Z, dim=b, dimLong=2, dimFloat=3.0, val=1}", "Unable to parse row [{\"time\":9.0x,\"dim\":\"a\",\"dimLong\":2,\"dimFloat\":3.0,\"val\":1}]", "Timestamp[unparseable] is unparseable! Event: {time=unparseable, dim=a, dimLong=2, dimFloat=3.0, val=1}");
}
actualMessages = parseExceptionReports.stream().map((r) -> {
return ((List<String>) r.get("details")).get(0);
}).collect(Collectors.toList());
Assert.assertEquals(expectedMessages, actualMessages);
expectedInputs = Arrays.asList("this is not JSON", "{time=99999999999-01-01T00:00:10Z, dim=b, dimLong=2, dimFloat=3.0, val=1}", "{\"time\":9.0x,\"dim\":\"a\",\"dimLong\":2,\"dimFloat\":3.0,\"val\":1}", "{time=unparseable, dim=a, dimLong=2, dimFloat=3.0, val=1}");
actualInputs = parseExceptionReports.stream().map((r) -> {
return (String) r.get("input");
}).collect(Collectors.toList());
Assert.assertEquals(expectedInputs, actualInputs);
}
Aggregations