List of usage examples for org.apache.hadoop.fs FileSystem exists
public boolean exists(Path f) throws IOException
From source file:com.linkedin.pinot.hadoop.job.SegmentCreationJob.java
License:Apache License
public void run() throws Exception { LOGGER.info("Starting {}", getClass().getSimpleName()); FileSystem fs = FileSystem.get(getConf()); Path inputPathPattern = new Path(_inputSegmentDir); if (fs.exists(new Path(_stagingDir))) { LOGGER.warn("Found the temp folder, deleting it"); fs.delete(new Path(_stagingDir), true); }//from w w w.j a v a2 s . co m fs.mkdirs(new Path(_stagingDir)); fs.mkdirs(new Path(_stagingDir + "/input/")); if (fs.exists(new Path(_outputDir))) { LOGGER.warn("Found the output folder, deleting it"); fs.delete(new Path(_outputDir), true); } fs.mkdirs(new Path(_outputDir)); List<FileStatus> inputDataFiles = new ArrayList<FileStatus>(); FileStatus[] fileStatusArr = fs.globStatus(inputPathPattern); for (FileStatus fileStatus : fileStatusArr) { inputDataFiles.addAll(getDataFilesFromPath(fs, fileStatus.getPath())); } for (int seqId = 0; seqId < inputDataFiles.size(); ++seqId) { FileStatus file = inputDataFiles.get(seqId); String completeFilePath = " " + file.getPath().toString() + " " + seqId; Path newOutPutFile = new Path((_stagingDir + "/input/" + file.getPath().toString().replace('.', '_').replace('/', '_').replace(':', '_') + ".txt")); FSDataOutputStream stream = fs.create(newOutPutFile); stream.writeUTF(completeFilePath); stream.flush(); stream.close(); } Job job = Job.getInstance(getConf()); job.setJarByClass(SegmentCreationJob.class); job.setJobName(_jobName); job.setMapperClass(HadoopSegmentCreationMapper.class); if (System.getenv("HADOOP_TOKEN_FILE_LOCATION") != null) { job.getConfiguration().set("mapreduce.job.credentials.binary", System.getenv("HADOOP_TOKEN_FILE_LOCATION")); } job.setInputFormatClass(TextInputFormat.class); job.setOutputFormatClass(TextOutputFormat.class); job.setMapOutputKeyClass(LongWritable.class); job.setMapOutputValueClass(Text.class); FileInputFormat.addInputPath(job, new Path(_stagingDir + "/input/")); FileOutputFormat.setOutputPath(job, new Path(_stagingDir + "/output/")); job.getConfiguration().setInt(JobContext.NUM_MAPS, inputDataFiles.size()); job.getConfiguration().set("data.schema", new ObjectMapper().writeValueAsString(_dataSchema)); job.setMaxReduceAttempts(1); job.setMaxMapAttempts(0); job.setNumReduceTasks(0); for (Object key : _properties.keySet()) { job.getConfiguration().set(key.toString(), _properties.getProperty(key.toString())); } if (_depsJarPath != null && _depsJarPath.length() > 0) { addDepsJarToDistributedCache(new Path(_depsJarPath), job); } // Submit the job for execution. job.waitForCompletion(true); if (!job.isSuccessful()) { throw new RuntimeException("Job failed : " + job); } LOGGER.info("Moving Segment Tar files from {} to: {}", _stagingDir + "/output/segmentTar", _outputDir); FileStatus[] segmentArr = fs.listStatus(new Path(_stagingDir + "/output/segmentTar")); for (FileStatus segment : segmentArr) { fs.rename(segment.getPath(), new Path(_outputDir, segment.getPath().getName())); } // Delete temporary directory. LOGGER.info("Cleanup the working directory."); LOGGER.info("Deleting the dir: {}", _stagingDir); fs.delete(new Path(_stagingDir), true); }
From source file:com.linkedin.thirdeye.bootstrap.segment.create.SegmentCreationPhaseJob.java
License:Apache License
public Job run() throws Exception { Job job = Job.getInstance(getConf()); job.setJarByClass(SegmentCreationPhaseJob.class); job.setJobName(name);/*from w ww . j a va2 s .c o m*/ FileSystem fs = FileSystem.get(getConf()); Configuration configuration = job.getConfiguration(); String schemaPath = getAndSetConfiguration(configuration, SEGMENT_CREATION_SCHEMA_PATH); LOGGER.info("Schema path : {}", schemaPath); String configPath = getAndSetConfiguration(configuration, SEGMENT_CREATION_CONFIG_PATH); LOGGER.info("Config path : {}", configPath); Schema dataSchema = createSchema(configPath); LOGGER.info("Data schema : {}", dataSchema); String inputSegmentDir = getAndSetConfiguration(configuration, SEGMENT_CREATION_INPUT_PATH); LOGGER.info("Input path : {}", inputSegmentDir); String outputDir = getAndSetConfiguration(configuration, SEGMENT_CREATION_OUTPUT_PATH); LOGGER.info("Output path : {}", outputDir); String stagingDir = new File(outputDir, TEMP).getAbsolutePath(); LOGGER.info("Staging dir : {}", stagingDir); String tableName = getAndSetConfiguration(configuration, SEGMENT_CREATION_SEGMENT_TABLE_NAME); LOGGER.info("Segment table name : {}", tableName); // Create temporary directory if (fs.exists(new Path(stagingDir))) { LOGGER.warn("Found the temp folder, deleting it"); fs.delete(new Path(stagingDir), true); } fs.mkdirs(new Path(stagingDir)); fs.mkdirs(new Path(stagingDir + "/input/")); if (fs.exists(new Path(outputDir))) { LOGGER.warn("Found the output folder deleting it"); fs.delete(new Path(outputDir), true); } fs.mkdirs(new Path(outputDir)); Path inputPathPattern = new Path(inputSegmentDir); List<FileStatus> inputDataFiles = Arrays.asList(fs.listStatus(inputPathPattern)); LOGGER.info("size {}", inputDataFiles.size()); try { for (int seqId = 0; seqId < inputDataFiles.size(); ++seqId) { FileStatus file = inputDataFiles.get(seqId); String completeFilePath = " " + file.getPath().toString() + " " + seqId; Path newOutPutFile = new Path((stagingDir + "/input/" + file.getPath().toString().replace('.', '_').replace('/', '_').replace(':', '_') + ".txt")); FSDataOutputStream stream = fs.create(newOutPutFile); LOGGER.info("wrote {}", completeFilePath); stream.writeUTF(completeFilePath); stream.flush(); stream.close(); } } catch (Exception e) { LOGGER.error("Exception while reading input files ", e); } job.setMapperClass(SegmentCreationPhaseMapReduceJob.SegmentCreationMapper.class); if (System.getenv("HADOOP_TOKEN_FILE_LOCATION") != null) { job.getConfiguration().set("mapreduce.job.credentials.binary", System.getenv("HADOOP_TOKEN_FILE_LOCATION")); } job.setInputFormatClass(TextInputFormat.class); job.setOutputFormatClass(TextOutputFormat.class); job.setMapOutputKeyClass(LongWritable.class); job.setMapOutputValueClass(Text.class); FileInputFormat.addInputPath(job, new Path(stagingDir + "/input/")); FileOutputFormat.setOutputPath(job, new Path(stagingDir + "/output/")); job.getConfiguration().setInt(JobContext.NUM_MAPS, inputDataFiles.size()); job.getConfiguration().set("data.schema", OBJECT_MAPPER.writeValueAsString(dataSchema)); if (!fs.exists(new Path(schemaPath))) { OBJECT_MAPPER.writerWithDefaultPrettyPrinter().writeValue(fs.create(new Path(schemaPath), false), dataSchema); } job.setMaxReduceAttempts(1); job.setMaxMapAttempts(0); job.setNumReduceTasks(0); for (Object key : props.keySet()) { job.getConfiguration().set(key.toString(), props.getProperty(key.toString())); } job.waitForCompletion(true); if (!job.isSuccessful()) { throw new RuntimeException("Job failed : " + job); } LOGGER.info("Moving Segment Tar files from {} to: {}", stagingDir + "/output/segmentTar", outputDir); FileStatus[] segmentArr = fs.listStatus(new Path(stagingDir + "/output/segmentTar")); for (FileStatus segment : segmentArr) { fs.rename(segment.getPath(), new Path(outputDir, segment.getPath().getName())); } // Delete temporary directory. LOGGER.info("Cleanup the working directory."); LOGGER.info("Deleting the dir: {}", stagingDir); fs.delete(new Path(stagingDir), true); return job; }
From source file:com.linkedin.thirdeye.hadoop.aggregation.AggregationPhaseJob.java
License:Apache License
public Job run() throws Exception { Job job = Job.getInstance(getConf()); job.setJobName(name);//from ww w .j ava 2 s.c om job.setJarByClass(AggregationPhaseJob.class); FileSystem fs = FileSystem.get(getConf()); Configuration configuration = job.getConfiguration(); // Properties LOGGER.info("Properties {}", props); // Input Path String inputPathDir = getAndSetConfiguration(configuration, AGG_PHASE_INPUT_PATH); LOGGER.info("Input path dir: " + inputPathDir); for (String inputPath : inputPathDir.split(ThirdEyeConstants.FIELD_SEPARATOR)) { LOGGER.info("Adding input:" + inputPath); Path input = new Path(inputPath); FileInputFormat.addInputPath(job, input); } // Output path Path outputPath = new Path(getAndSetConfiguration(configuration, AGG_PHASE_OUTPUT_PATH)); LOGGER.info("Output path dir: " + outputPath.toString()); if (fs.exists(outputPath)) { fs.delete(outputPath, true); } FileOutputFormat.setOutputPath(job, outputPath); // Schema Schema avroSchema = ThirdeyeAvroUtils.getSchema(inputPathDir); LOGGER.info("Schema : {}", avroSchema.toString(true)); job.getConfiguration().set(AGG_PHASE_AVRO_SCHEMA.toString(), avroSchema.toString()); // ThirdEyeConfig String metricTypesProperty = ThirdeyeAvroUtils.getMetricTypesProperty( props.getProperty(ThirdEyeConfigProperties.THIRDEYE_METRIC_NAMES.toString()), props.getProperty(ThirdEyeConfigProperties.THIRDEYE_METRIC_TYPES.toString()), avroSchema); props.setProperty(ThirdEyeConfigProperties.THIRDEYE_METRIC_TYPES.toString(), metricTypesProperty); ThirdEyeConfig thirdeyeConfig = ThirdEyeConfig.fromProperties(props); LOGGER.info("Thirdeye Config {}", thirdeyeConfig.encode()); job.getConfiguration().set(AGG_PHASE_THIRDEYE_CONFIG.toString(), OBJECT_MAPPER.writeValueAsString(thirdeyeConfig)); // Map config job.setMapperClass(AggregationMapper.class); job.setInputFormatClass(AvroKeyInputFormat.class); job.setMapOutputKeyClass(BytesWritable.class); job.setMapOutputValueClass(BytesWritable.class); // Reduce config job.setReducerClass(AggregationReducer.class); job.setOutputKeyClass(AvroKey.class); job.setOutputValueClass(NullWritable.class); AvroJob.setOutputKeySchema(job, avroSchema); job.setOutputFormatClass(AvroKeyOutputFormat.class); String numReducers = props.getProperty(ThirdEyeJobProperties.THIRDEYE_NUM_REDUCERS.getName()); LOGGER.info("Num Reducers : {}", numReducers); if (StringUtils.isNotBlank(numReducers)) { job.setNumReduceTasks(Integer.valueOf(numReducers)); LOGGER.info("Setting num reducers {}", job.getNumReduceTasks()); } job.waitForCompletion(true); Counter counter = job.getCounters().findCounter(AggregationCounter.NUMBER_OF_RECORDS); LOGGER.info(counter.getDisplayName() + " : " + counter.getValue()); if (counter.getValue() == 0) { throw new IllegalStateException("No input records in " + inputPathDir); } counter = job.getCounters().findCounter(AggregationCounter.NUMBER_OF_RECORDS_FLATTENED); LOGGER.info(counter.getDisplayName() + " : " + counter.getValue()); for (String metric : thirdeyeConfig.getMetricNames()) { counter = job.getCounters().findCounter(thirdeyeConfig.getCollection(), metric); LOGGER.info(counter.getDisplayName() + " : " + counter.getValue()); } return job; }
From source file:com.linkedin.thirdeye.hadoop.backfill.BackfillControllerAPIs.java
License:Apache License
/** * Downloads a segment from the controller, given the table name and segment name * @param segmentName//from w w w .ja va 2 s . c o m * @param hdfsSegmentPath * @throws IOException * @throws ArchiveException */ public void downloadSegment(String segmentName, Path hdfsSegmentPath) throws IOException, ArchiveException { FileSystem fs = FileSystem.get(new Configuration()); HttpClient controllerClient = new DefaultHttpClient(); HttpGet req = new HttpGet(SEGMENTS_ENDPOINT + URLEncoder.encode(tableName, UTF_8) + "/" + URLEncoder.encode(segmentName, UTF_8)); HttpResponse res = controllerClient.execute(controllerHttpHost, req); try { if (res.getStatusLine().getStatusCode() != 200) { throw new IllegalStateException(res.getStatusLine().toString()); } LOGGER.info("Fetching segment {}", segmentName); InputStream content = res.getEntity().getContent(); File tempDir = new File(Files.createTempDir(), "thirdeye_temp"); tempDir.mkdir(); LOGGER.info("Creating temporary dir for staging segments {}", tempDir); File tempSegmentDir = new File(tempDir, segmentName); File tempSegmentTar = new File(tempDir, segmentName + ThirdEyeConstants.TAR_SUFFIX); LOGGER.info("Downloading {} to {}", segmentName, tempSegmentTar); OutputStream out = new FileOutputStream(tempSegmentTar); IOUtils.copy(content, out); if (!tempSegmentTar.exists()) { throw new IllegalStateException("Download of " + segmentName + " unsuccessful"); } LOGGER.info("Extracting segment {} to {}", tempSegmentTar, tempDir); TarGzCompressionUtils.unTar(tempSegmentTar, tempDir); File[] files = tempDir.listFiles(new FilenameFilter() { @Override public boolean accept(File dir, String name) { return !name.endsWith(ThirdEyeConstants.TAR_SUFFIX) && new File(dir, name).isDirectory(); } }); if (files.length == 0) { throw new IllegalStateException("Failed to extract " + tempSegmentTar + " to " + tempDir); } else if (!files[0].getName().equals(tempSegmentDir.getName())) { LOGGER.info("Moving extracted segment to the segment dir {}", tempSegmentDir); FileUtils.moveDirectory(files[0], tempSegmentDir); } if (!tempSegmentDir.exists()) { throw new IllegalStateException("Failed to move " + files[0] + " to " + tempSegmentDir); } LOGGER.info("Copying segment from {} to hdfs {}", tempSegmentDir, hdfsSegmentPath); fs.copyFromLocalFile(new Path(tempSegmentDir.toString()), hdfsSegmentPath); Path hdfsSegmentDir = new Path(hdfsSegmentPath, segmentName); if (!fs.exists(hdfsSegmentDir)) { throw new IllegalStateException("Failed to copy segment " + segmentName + " from local path " + tempSegmentDir + " to hdfs path " + hdfsSegmentPath); } } finally { if (res.getEntity() != null) { EntityUtils.consume(res.getEntity()); } } LOGGER.info("Successfully downloaded segment {} to {}", segmentName, hdfsSegmentPath); }
From source file:com.linkedin.thirdeye.hadoop.backfill.BackfillPhaseJob.java
License:Apache License
public Job run() throws Exception { Job job = Job.getInstance(getConf()); job.setJarByClass(BackfillPhaseJob.class); job.setJobName(name);//from w w w . ja v a2 s. co m FileSystem fs = FileSystem.get(getConf()); Configuration configuration = job.getConfiguration(); LOGGER.info("*******************************************************************************"); String controllerHost = getAndSetConfiguration(configuration, BACKFILL_PHASE_CONTROLLER_HOST); String controllerPort = getAndSetConfiguration(configuration, BACKFILL_PHASE_CONTROLLER_PORT); LOGGER.info("Controller Host : {} Controller Port : {}", controllerHost, controllerPort); String segmentStartTime = getAndSetConfiguration(configuration, BACKFILL_PHASE_START_TIME); String segmentEndTime = getAndSetConfiguration(configuration, BACKFILL_PHASE_END_TIME); long startTime = Long.valueOf(segmentStartTime); long endTime = Long.valueOf(segmentEndTime); if (Long.valueOf(segmentStartTime) > Long.valueOf(segmentEndTime)) { throw new IllegalStateException("Start time cannot be greater than end time"); } String tableName = getAndSetConfiguration(configuration, BACKFILL_PHASE_TABLE_NAME); LOGGER.info("Start time : {} End time : {} Table name : {}", segmentStartTime, segmentEndTime, tableName); String outputPath = getAndSetConfiguration(configuration, BACKFILL_PHASE_OUTPUT_PATH); LOGGER.info("Output path : {}", outputPath); Path backfillDir = new Path(outputPath); if (fs.exists(backfillDir)) { LOGGER.warn("Found the output folder deleting it"); fs.delete(backfillDir, true); } Path downloadDir = new Path(backfillDir, DOWNLOAD); LOGGER.info("Creating download dir : {}", downloadDir); fs.mkdirs(downloadDir); Path inputDir = new Path(backfillDir, INPUT); LOGGER.info("Creating input dir : {}", inputDir); fs.mkdirs(inputDir); Path outputDir = new Path(backfillDir, OUTPUT); LOGGER.info("Creating output dir : {}", outputDir); BackfillControllerAPIs backfillControllerAPIs = new BackfillControllerAPIs(controllerHost, Integer.valueOf(controllerPort), tableName); LOGGER.info("Downloading segments in range {} to {}", startTime, endTime); List<String> allSegments = backfillControllerAPIs.getAllSegments(tableName); List<String> segmentsToDownload = backfillControllerAPIs.findSegmentsInRange(tableName, allSegments, startTime, endTime); for (String segmentName : segmentsToDownload) { backfillControllerAPIs.downloadSegment(segmentName, downloadDir); } LOGGER.info("Reading downloaded segment input files"); List<FileStatus> inputDataFiles = new ArrayList<>(); inputDataFiles.addAll(Lists.newArrayList(fs.listStatus(downloadDir))); LOGGER.info("size {}", inputDataFiles.size()); try { LOGGER.info("Creating input files at {} for segment input files", inputDir); for (int seqId = 0; seqId < inputDataFiles.size(); ++seqId) { FileStatus file = inputDataFiles.get(seqId); String completeFilePath = " " + file.getPath().toString() + " " + seqId; Path newOutPutFile = new Path((inputDir + "/" + file.getPath().toString().replace('.', '_').replace('/', '_').replace(':', '_') + ".txt")); FSDataOutputStream stream = fs.create(newOutPutFile); LOGGER.info("wrote {}", completeFilePath); stream.writeUTF(completeFilePath); stream.flush(); stream.close(); } } catch (Exception e) { LOGGER.error("Exception while reading input files ", e); } job.setMapperClass(BackfillPhaseMapJob.BackfillMapper.class); if (System.getenv("HADOOP_TOKEN_FILE_LOCATION") != null) { job.getConfiguration().set("mapreduce.job.credentials.binary", System.getenv("HADOOP_TOKEN_FILE_LOCATION")); } job.setInputFormatClass(TextInputFormat.class); job.setOutputFormatClass(TextOutputFormat.class); job.setMapOutputKeyClass(LongWritable.class); job.setMapOutputValueClass(Text.class); FileInputFormat.addInputPath(job, inputDir); FileOutputFormat.setOutputPath(job, outputDir); job.getConfiguration().setInt(JobContext.NUM_MAPS, inputDataFiles.size()); job.setMaxReduceAttempts(1); job.setMaxMapAttempts(0); job.setNumReduceTasks(0); for (Object key : props.keySet()) { job.getConfiguration().set(key.toString(), props.getProperty(key.toString())); } job.waitForCompletion(true); if (!job.isSuccessful()) { throw new RuntimeException("Job failed : " + job); } LOGGER.info("Cleanup the working directory"); LOGGER.info("Deleting the dir: {}", downloadDir); fs.delete(downloadDir, true); LOGGER.info("Deleting the dir: {}", inputDir); fs.delete(inputDir, true); LOGGER.info("Deleting the dir: {}", outputDir); fs.delete(outputDir, true); return job; }
From source file:com.linkedin.thirdeye.hadoop.derivedcolumn.transformation.DerivedColumnTransformationPhaseJob.java
License:Apache License
public Job run() throws Exception { Job job = Job.getInstance(getConf()); job.setJobName(name);/*from w w w . ja va 2 s . co m*/ job.setJarByClass(DerivedColumnTransformationPhaseJob.class); Configuration configuration = job.getConfiguration(); FileSystem fs = FileSystem.get(configuration); // Input Path String inputPathDir = getAndSetConfiguration(configuration, DERIVED_COLUMN_TRANSFORMATION_PHASE_INPUT_PATH); LOGGER.info("Input path dir: " + inputPathDir); for (String inputPath : inputPathDir.split(",")) { LOGGER.info("Adding input:" + inputPath); Path input = new Path(inputPath); FileInputFormat.addInputPath(job, input); } // Topk path String topkPath = getAndSetConfiguration(configuration, DERIVED_COLUMN_TRANSFORMATION_PHASE_TOPK_PATH); LOGGER.info("Topk path : " + topkPath); // Output path Path outputPath = new Path( getAndSetConfiguration(configuration, DERIVED_COLUMN_TRANSFORMATION_PHASE_OUTPUT_PATH)); LOGGER.info("Output path dir: " + outputPath.toString()); if (fs.exists(outputPath)) { fs.delete(outputPath, true); } FileOutputFormat.setOutputPath(job, outputPath); // Schema Schema avroSchema = ThirdeyeAvroUtils.getSchema(inputPathDir); LOGGER.info("Schema : {}", avroSchema.toString(true)); // ThirdEyeConfig String metricTypesProperty = ThirdeyeAvroUtils.getMetricTypesProperty( props.getProperty(ThirdEyeConfigProperties.THIRDEYE_METRIC_NAMES.toString()), props.getProperty(ThirdEyeConfigProperties.THIRDEYE_METRIC_TYPES.toString()), avroSchema); props.setProperty(ThirdEyeConfigProperties.THIRDEYE_METRIC_TYPES.toString(), metricTypesProperty); ThirdEyeConfig thirdeyeConfig = ThirdEyeConfig.fromProperties(props); job.getConfiguration().set(DERIVED_COLUMN_TRANSFORMATION_PHASE_THIRDEYE_CONFIG.toString(), OBJECT_MAPPER.writeValueAsString(thirdeyeConfig)); LOGGER.info("ThirdEyeConfig {}", thirdeyeConfig.encode()); // New schema Schema outputSchema = newSchema(thirdeyeConfig); job.getConfiguration().set(DERIVED_COLUMN_TRANSFORMATION_PHASE_OUTPUT_SCHEMA.toString(), outputSchema.toString()); // Map config job.setMapperClass(DerivedColumnTransformationPhaseMapper.class); job.setInputFormatClass(AvroKeyInputFormat.class); job.setMapOutputKeyClass(AvroKey.class); job.setMapOutputValueClass(NullWritable.class); AvroJob.setOutputKeySchema(job, outputSchema); LazyOutputFormat.setOutputFormatClass(job, AvroKeyOutputFormat.class); AvroMultipleOutputs.addNamedOutput(job, "avro", AvroKeyOutputFormat.class, outputSchema); job.setNumReduceTasks(0); job.waitForCompletion(true); return job; }
From source file:com.linkedin.thirdeye.hadoop.segment.creation.SegmentCreationPhaseJob.java
License:Apache License
public Job run() throws Exception { Job job = Job.getInstance(getConf()); job.setJarByClass(SegmentCreationPhaseJob.class); job.setJobName(name);/*from w w w .ja v a2s . c om*/ FileSystem fs = FileSystem.get(getConf()); Configuration configuration = job.getConfiguration(); String inputSegmentDir = getAndSetConfiguration(configuration, SEGMENT_CREATION_INPUT_PATH); LOGGER.info("Input path : {}", inputSegmentDir); Schema avroSchema = ThirdeyeAvroUtils.getSchema(inputSegmentDir); LOGGER.info("Schema : {}", avroSchema); String metricTypesProperty = ThirdeyeAvroUtils.getMetricTypesProperty( props.getProperty(ThirdEyeConfigProperties.THIRDEYE_METRIC_NAMES.toString()), props.getProperty(ThirdEyeConfigProperties.THIRDEYE_METRIC_TYPES.toString()), avroSchema); props.setProperty(ThirdEyeConfigProperties.THIRDEYE_METRIC_TYPES.toString(), metricTypesProperty); ThirdEyeConfig thirdeyeConfig = ThirdEyeConfig.fromProperties(props); LOGGER.info("ThirdEyeConfig {}", thirdeyeConfig.encode()); String outputDir = getAndSetConfiguration(configuration, SEGMENT_CREATION_OUTPUT_PATH); LOGGER.info("Output path : {}", outputDir); Path stagingDir = new Path(outputDir, TEMP); LOGGER.info("Staging dir : {}", stagingDir); String segmentWallClockStart = getAndSetConfiguration(configuration, SEGMENT_CREATION_WALLCLOCK_START_TIME); LOGGER.info("Segment wallclock start time : {}", segmentWallClockStart); String segmentWallClockEnd = getAndSetConfiguration(configuration, SEGMENT_CREATION_WALLCLOCK_END_TIME); LOGGER.info("Segment wallclock end time : {}", segmentWallClockEnd); String schedule = getAndSetConfiguration(configuration, SEGMENT_CREATION_SCHEDULE); LOGGER.info("Segment schedule : {}", schedule); String isBackfill = props.getProperty(SEGMENT_CREATION_BACKFILL.toString(), DEFAULT_BACKFILL); configuration.set(SEGMENT_CREATION_BACKFILL.toString(), isBackfill); LOGGER.info("Is Backfill : {}", configuration.get(SEGMENT_CREATION_BACKFILL.toString())); // Create temporary directory if (fs.exists(stagingDir)) { LOGGER.warn("Found the temp folder, deleting it"); fs.delete(stagingDir, true); } fs.mkdirs(stagingDir); fs.mkdirs(new Path(stagingDir + "/input/")); // Create output directory if (fs.exists(new Path(outputDir))) { LOGGER.warn("Found the output folder deleting it"); fs.delete(new Path(outputDir), true); } fs.mkdirs(new Path(outputDir)); // Read input files List<FileStatus> inputDataFiles = new ArrayList<>(); for (String input : inputSegmentDir.split(",")) { Path inputPathPattern = new Path(input); inputDataFiles.addAll(Arrays.asList(fs.listStatus(inputPathPattern))); } LOGGER.info("size {}", inputDataFiles.size()); try { for (int seqId = 0; seqId < inputDataFiles.size(); ++seqId) { FileStatus file = inputDataFiles.get(seqId); String completeFilePath = " " + file.getPath().toString() + " " + seqId; Path newOutPutFile = new Path((stagingDir + "/input/" + file.getPath().toString().replace('.', '_').replace('/', '_').replace(':', '_') + ".txt")); FSDataOutputStream stream = fs.create(newOutPutFile); LOGGER.info("wrote {}", completeFilePath); stream.writeUTF(completeFilePath); stream.flush(); stream.close(); } } catch (Exception e) { LOGGER.error("Exception while reading input files ", e); } job.setMapperClass(SegmentCreationPhaseMapReduceJob.SegmentCreationMapper.class); if (System.getenv("HADOOP_TOKEN_FILE_LOCATION") != null) { job.getConfiguration().set("mapreduce.job.credentials.binary", System.getenv("HADOOP_TOKEN_FILE_LOCATION")); } job.setInputFormatClass(TextInputFormat.class); job.setOutputFormatClass(TextOutputFormat.class); job.setMapOutputKeyClass(LongWritable.class); job.setMapOutputValueClass(Text.class); FileInputFormat.addInputPath(job, new Path(stagingDir + "/input/")); FileOutputFormat.setOutputPath(job, new Path(stagingDir + "/output/")); job.getConfiguration().setInt(JobContext.NUM_MAPS, inputDataFiles.size()); job.getConfiguration().set(SEGMENT_CREATION_THIRDEYE_CONFIG.toString(), OBJECT_MAPPER.writeValueAsString(thirdeyeConfig)); job.setMaxReduceAttempts(1); job.setMaxMapAttempts(0); job.setNumReduceTasks(0); for (Object key : props.keySet()) { job.getConfiguration().set(key.toString(), props.getProperty(key.toString())); } job.waitForCompletion(true); if (!job.isSuccessful()) { throw new RuntimeException("Job failed : " + job); } LOGGER.info("Moving Segment Tar files from {} to: {}", stagingDir + "/output/segmentTar", outputDir); FileStatus[] segmentArr = fs.listStatus(new Path(stagingDir + "/output/segmentTar")); for (FileStatus segment : segmentArr) { fs.rename(segment.getPath(), new Path(outputDir, segment.getPath().getName())); } // Delete temporary directory. LOGGER.info("Cleanup the working directory."); LOGGER.info("Deleting the dir: {}", stagingDir); fs.delete(stagingDir, true); return job; }
From source file:com.linkedin.thirdeye.hadoop.topk.TopKPhaseJob.java
License:Apache License
public Job run() throws Exception { Job job = Job.getInstance(getConf()); job.setJobName(name);//from www.ja v a2s . c om job.setJarByClass(TopKPhaseJob.class); Configuration configuration = job.getConfiguration(); FileSystem fs = FileSystem.get(configuration); // Properties LOGGER.info("Properties {}", props); // Input Path String inputPathDir = getAndSetConfiguration(configuration, TOPK_PHASE_INPUT_PATH); LOGGER.info("Input path dir: " + inputPathDir); for (String inputPath : inputPathDir.split(ThirdEyeConstants.FIELD_SEPARATOR)) { LOGGER.info("Adding input:" + inputPath); Path input = new Path(inputPath); FileInputFormat.addInputPath(job, input); } // Output path Path outputPath = new Path(getAndSetConfiguration(configuration, TOPK_PHASE_OUTPUT_PATH)); LOGGER.info("Output path dir: " + outputPath.toString()); if (fs.exists(outputPath)) { fs.delete(outputPath, true); } FileOutputFormat.setOutputPath(job, outputPath); // Schema Schema avroSchema = ThirdeyeAvroUtils.getSchema(inputPathDir); LOGGER.info("Schema : {}", avroSchema.toString(true)); // ThirdEyeConfig String metricTypesProperty = ThirdeyeAvroUtils.getMetricTypesProperty( props.getProperty(ThirdEyeConfigProperties.THIRDEYE_METRIC_NAMES.toString()), props.getProperty(ThirdEyeConfigProperties.THIRDEYE_METRIC_TYPES.toString()), avroSchema); props.setProperty(ThirdEyeConfigProperties.THIRDEYE_METRIC_TYPES.toString(), metricTypesProperty); ThirdEyeConfig thirdeyeConfig = ThirdEyeConfig.fromProperties(props); LOGGER.info("Thirdeye Config {}", thirdeyeConfig.encode()); job.getConfiguration().set(TOPK_PHASE_THIRDEYE_CONFIG.toString(), OBJECT_MAPPER.writeValueAsString(thirdeyeConfig)); // Map config job.setMapperClass(TopKPhaseMapper.class); job.setInputFormatClass(AvroKeyInputFormat.class); job.setMapOutputKeyClass(BytesWritable.class); job.setMapOutputValueClass(BytesWritable.class); // Combiner job.setCombinerClass(TopKPhaseCombiner.class); // Reduce config job.setReducerClass(TopKPhaseReducer.class); job.setOutputKeyClass(NullWritable.class); job.setOutputValueClass(NullWritable.class); job.setNumReduceTasks(1); job.waitForCompletion(true); return job; }
From source file:com.linkedin.whiteelephant.util.JobStatsProcessing.java
License:Apache License
public static List<ProcessingTask> getTasks(FileSystem fs, String logsRoot, String clusterName, String outputPathRoot, boolean incremental, int numDays, int numDaysForced) throws IOException { Calendar cal = Calendar.getInstance(timeZone); SimpleDateFormat yearFormat = new SimpleDateFormat("yyyy"); SimpleDateFormat dayFormat = new SimpleDateFormat("MMdd"); SimpleDateFormat idFormat = new SimpleDateFormat("yyyy-MM-dd"); yearFormat.setTimeZone(timeZone);//w w w. j a v a2s. com dayFormat.setTimeZone(timeZone); idFormat.setTimeZone(timeZone); List<ProcessingTask> processingTasks = new ArrayList<ProcessingTask>(); numDays = Math.max(numDays, numDaysForced); // Start processing previous day of data since current day isn't yet finished. Unless we are aggregating hourly data there is no point. cal.add(Calendar.DAY_OF_MONTH, -1); int numPaths = 0; long totalLength = 0; for (int i = 0; i < numDays; i++, cal.add(Calendar.DAY_OF_MONTH, -1)) { Date date = cal.getTime(); String pathFormat = String.format("%s/%s/daily/*/%s/%s/*.log", logsRoot, clusterName, yearFormat.format(date), dayFormat.format(date)); FileStatus[] stats = fs.globStatus(new Path(pathFormat)); StringBuilder msg = new StringBuilder(pathFormat + " => " + stats.length + " files"); String outputPathForDay = String.format("%s/%s/%s/%s", outputPathRoot, clusterName, yearFormat.format(date), dayFormat.format(date)); if (stats.length > 0) { if (!incremental || !fs.exists(new Path(outputPathForDay)) || i < numDaysForced) { for (FileStatus stat : stats) { totalLength += stat.getLen(); numPaths++; } String id = clusterName + "-" + idFormat.format(date); System.out.println(msg); processingTasks.add(new ProcessingTask(id, pathFormat, outputPathForDay, totalLength)); } else if (incremental && fs.exists(new Path(outputPathForDay))) { msg.append(" (skipping)"); System.out.println(msg); } } } System.out.println("Found " + numPaths + " paths to process, totalling " + totalLength + " bytes (" + (totalLength / (1024 * 1024 * 1024)) + " gigabytes)"); return processingTasks; }
From source file:com.littlehotspot.hadoop.mr.box.BoxLog.java
License:Open Source License
@Override public int run(String[] arg) throws Exception { try {//ww w . j a va 2 s. com // ??? if (arg.length > 2) { BOX_LOG_FORMAT_REGEX = Pattern.compile(arg[2]); } Job job = Job.getInstance(this.getConf(), BoxLog.class.getSimpleName()); job.setJarByClass(BoxLog.class); /***/ Path inputPath = new Path(arg[0]); FileInputFormat.addInputPath(job, inputPath); job.setMapperClass(BoxMapper.class); job.setMapOutputKeyClass(Text.class); job.setMapOutputValueClass(Text.class); /***/ Path outputPath = new Path(arg[1]); FileSystem fileSystem = FileSystem.get(new URI(outputPath.toString()), new Configuration()); if (fileSystem.exists(outputPath)) { fileSystem.delete(outputPath, true); } FileOutputFormat.setOutputPath(job, outputPath); job.setReducerClass(BoxReduce.class); job.setOutputKeyClass(Text.class); job.setOutputValueClass(Text.class); boolean status = job.waitForCompletion(true); if (!status) { throw new Exception("MapReduce task execute failed........."); } return 0; } catch (Exception e) { e.printStackTrace(); return 1; } }