List of usage examples for org.apache.hadoop.mapreduce Job submit
public void submit() throws IOException, InterruptedException, ClassNotFoundException
From source file:com.ikanow.infinit.e.core.mapreduce.HadoopJobRunner.java
License:Open Source License
@SuppressWarnings({ "unchecked", "rawtypes" }) private String runHadoopJob(CustomMapReduceJobPojo job, String tempJarLocation) throws IOException, SAXException, ParserConfigurationException { StringWriter xml = new StringWriter(); createConfigXML(xml, job.jobtitle, job.inputCollection, getQueryOrProcessing(job.query, QuerySpec.INPUTFIELDS), job.isCustomTable, job.getOutputDatabase(), job._id.toString(), job.outputCollectionTemp, job.mapper, job.reducer, job.combiner, getQueryOrProcessing(job.query, QuerySpec.QUERY), job.communityIds, job.outputKey, job.outputValue, job.arguments);/*from w w w . j a v a2 s . c o m*/ ClassLoader savedClassLoader = Thread.currentThread().getContextClassLoader(); URLClassLoader child = new URLClassLoader(new URL[] { new File(tempJarLocation).toURI().toURL() }, savedClassLoader); Thread.currentThread().setContextClassLoader(child); // Now load the XML into a configuration object: Configuration config = new Configuration(); try { DocumentBuilderFactory dbFactory = DocumentBuilderFactory.newInstance(); DocumentBuilder dBuilder = dbFactory.newDocumentBuilder(); Document doc = dBuilder.parse(new ByteArrayInputStream(xml.toString().getBytes())); NodeList nList = doc.getElementsByTagName("property"); for (int temp = 0; temp < nList.getLength(); temp++) { Node nNode = nList.item(temp); if (nNode.getNodeType() == Node.ELEMENT_NODE) { Element eElement = (Element) nNode; String name = getTagValue("name", eElement); String value = getTagValue("value", eElement); if ((null != name) && (null != value)) { config.set(name, value); } } } } catch (Exception e) { throw new IOException(e.getMessage()); } // Now run the JAR file try { config.setBoolean("mapred.used.genericoptionsparser", true); // (just stops an annoying warning from appearing) if (bLocalMode) { config.set("mapred.job.tracker", "local"); config.set("fs.default.name", "local"); } else { String trackerUrl = HadoopUtils.getXMLProperty( prop_custom.getHadoopConfigPath() + "/hadoop/mapred-site.xml", "mapred.job.tracker"); String fsUrl = HadoopUtils.getXMLProperty( prop_custom.getHadoopConfigPath() + "/hadoop/core-site.xml", "fs.default.name"); config.set("mapred.job.tracker", trackerUrl); config.set("fs.default.name", fsUrl); } Job hj = new Job(config); Class<?> classToLoad = Class.forName(job.mapper, true, child); hj.setJarByClass(classToLoad); hj.setInputFormatClass((Class<? extends InputFormat>) Class .forName("com.ikanow.infinit.e.data_model.custom.InfiniteMongoInputFormat", true, child)); if ((null != job.exportToHdfs) && job.exportToHdfs) { hj.setOutputFormatClass((Class<? extends OutputFormat>) Class .forName("org.apache.hadoop.mapreduce.lib.output.SequenceFileOutputFormat", true, child)); Path outPath = this.ensureOutputDirectory(job); SequenceFileOutputFormat.setOutputPath(hj, outPath); } else { // normal case, stays in MongoDB hj.setOutputFormatClass((Class<? extends OutputFormat>) Class .forName("com.mongodb.hadoop.MongoOutputFormat", true, child)); } hj.setMapperClass((Class<? extends Mapper>) Class.forName(job.mapper, true, child)); if ((null != job.reducer) && !job.reducer.equalsIgnoreCase("null") && !job.reducer.equalsIgnoreCase("none")) { hj.setReducerClass((Class<? extends Reducer>) Class.forName(job.reducer, true, child)); } else { hj.setNumReduceTasks(0); } if ((null != job.combiner) && !job.combiner.equalsIgnoreCase("null") && !job.combiner.equalsIgnoreCase("none")) { hj.setCombinerClass((Class<? extends Reducer>) Class.forName(job.combiner, true, child)); } hj.setOutputKeyClass(Class.forName(job.outputKey, true, child)); hj.setOutputValueClass(Class.forName(job.outputValue, true, child)); hj.setJobName(job.jobtitle); if (bLocalMode) { hj.waitForCompletion(false); return "local_done"; } else { hj.submit(); String jobId = hj.getJobID().toString(); return jobId; } } catch (Exception e) { e.printStackTrace(); Thread.currentThread().setContextClassLoader(savedClassLoader); return "Error: " + HarvestExceptionUtils.createExceptionMessage(e); } finally { Thread.currentThread().setContextClassLoader(savedClassLoader); } }
From source file:com.ikanow.infinit.e.processing.custom.launcher.CustomHadoopTaskLauncher.java
License:Open Source License
@SuppressWarnings({ "unchecked", "rawtypes" }) public String runHadoopJob(CustomMapReduceJobPojo job, String tempJarLocation) throws IOException, SAXException, ParserConfigurationException { StringWriter xml = new StringWriter(); String outputCollection = job.outputCollectionTemp;// (non-append mode) if ((null != job.appendResults) && job.appendResults) outputCollection = job.outputCollection; // (append mode, write directly in....) else if (null != job.incrementalMode) job.incrementalMode = false; // (not allowed to be in incremental mode and not update mode) createConfigXML(xml, job.jobtitle, job.inputCollection, InfiniteHadoopUtils.getQueryOrProcessing(job.query, InfiniteHadoopUtils.QuerySpec.INPUTFIELDS), job.isCustomTable, job.getOutputDatabase(), job._id.toString(), outputCollection, job.mapper, job.reducer, job.combiner,/*w w w . j a v a 2s.c o m*/ InfiniteHadoopUtils.getQueryOrProcessing(job.query, InfiniteHadoopUtils.QuerySpec.QUERY), job.communityIds, job.outputKey, job.outputValue, job.arguments, job.incrementalMode, job.submitterID, job.selfMerge, job.outputCollection, job.appendResults); ClassLoader savedClassLoader = Thread.currentThread().getContextClassLoader(); URLClassLoader child = new URLClassLoader(new URL[] { new File(tempJarLocation).toURI().toURL() }, savedClassLoader); Thread.currentThread().setContextClassLoader(child); // Check version: for now, any infinit.e.data_model with an VersionTest class is acceptable boolean dataModelLoaded = true; try { URLClassLoader versionTest = new URLClassLoader(new URL[] { new File(tempJarLocation).toURI().toURL() }, null); try { Class.forName("com.ikanow.infinit.e.data_model.custom.InfiniteMongoInputFormat", true, versionTest); } catch (ClassNotFoundException e2) { //(this is fine, will use the cached version) dataModelLoaded = false; } if (dataModelLoaded) Class.forName("com.ikanow.infinit.e.data_model.custom.InfiniteMongoVersionTest", true, versionTest); } catch (ClassNotFoundException e1) { throw new RuntimeException( "This JAR is compiled with too old a version of the data-model, please recompile with Jan 2014 (rc2) onwards"); } // Now load the XML into a configuration object: Configuration config = new Configuration(); // Add the client configuration overrides: if (!bLocalMode) { String hadoopConfigPath = props_custom.getHadoopConfigPath() + "/hadoop/"; config.addResource(new Path(hadoopConfigPath + "core-site.xml")); config.addResource(new Path(hadoopConfigPath + "mapred-site.xml")); config.addResource(new Path(hadoopConfigPath + "hadoop-site.xml")); } //TESTED try { DocumentBuilderFactory dbFactory = DocumentBuilderFactory.newInstance(); DocumentBuilder dBuilder = dbFactory.newDocumentBuilder(); Document doc = dBuilder.parse(new ByteArrayInputStream(xml.toString().getBytes())); NodeList nList = doc.getElementsByTagName("property"); for (int temp = 0; temp < nList.getLength(); temp++) { Node nNode = nList.item(temp); if (nNode.getNodeType() == Node.ELEMENT_NODE) { Element eElement = (Element) nNode; String name = getTagValue("name", eElement); String value = getTagValue("value", eElement); if ((null != name) && (null != value)) { config.set(name, value); } } } } catch (Exception e) { throw new IOException(e.getMessage()); } // Some other config defaults: // (not sure if these are actually applied, or derived from the defaults - for some reason they don't appear in CDH's client config) config.set("mapred.map.tasks.speculative.execution", "false"); config.set("mapred.reduce.tasks.speculative.execution", "false"); // (default security is ignored here, have it set via HADOOP_TASKTRACKER_CONF in cloudera) // Now run the JAR file try { BasicDBObject advancedConfigurationDbo = null; try { advancedConfigurationDbo = (null != job.query) ? ((BasicDBObject) com.mongodb.util.JSON.parse(job.query)) : (new BasicDBObject()); } catch (Exception e) { advancedConfigurationDbo = new BasicDBObject(); } boolean esMode = advancedConfigurationDbo.containsField("qt") && !job.isCustomTable; if (esMode && !job.inputCollection.equals("doc_metadata.metadata")) { throw new RuntimeException( "Infinit.e Queries are only supported on doc_metadata - use MongoDB queries instead."); } config.setBoolean("mapred.used.genericoptionsparser", true); // (just stops an annoying warning from appearing) if (bLocalMode) { // local job tracker and FS mode config.set("mapred.job.tracker", "local"); config.set("fs.default.name", "local"); } else { if (bTestMode) { // run job tracker locally but FS mode remotely config.set("mapred.job.tracker", "local"); } else { // normal job tracker String trackerUrl = HadoopUtils.getXMLProperty( props_custom.getHadoopConfigPath() + "/hadoop/mapred-site.xml", "mapred.job.tracker"); config.set("mapred.job.tracker", trackerUrl); } String fsUrl = HadoopUtils.getXMLProperty( props_custom.getHadoopConfigPath() + "/hadoop/core-site.xml", "fs.default.name"); config.set("fs.default.name", fsUrl); } if (!dataModelLoaded && !(bTestMode || bLocalMode)) { // If running distributed and no data model loaded then add ourselves Path jarToCache = InfiniteHadoopUtils.cacheLocalFile("/opt/infinite-home/lib/", "infinit.e.data_model.jar", config); DistributedCache.addFileToClassPath(jarToCache, config); jarToCache = InfiniteHadoopUtils.cacheLocalFile("/opt/infinite-home/lib/", "infinit.e.processing.custom.library.jar", config); DistributedCache.addFileToClassPath(jarToCache, config); } //TESTED // Debug scripts (only if they exist), and only in non local/test mode if (!bLocalMode && !bTestMode) { try { Path scriptToCache = InfiniteHadoopUtils.cacheLocalFile("/opt/infinite-home/scripts/", "custom_map_error_handler.sh", config); config.set("mapred.map.task.debug.script", "custom_map_error_handler.sh " + job.jobtitle); config.set("mapreduce.map.debug.script", "custom_map_error_handler.sh " + job.jobtitle); DistributedCache.createSymlink(config); DistributedCache.addCacheFile(scriptToCache.toUri(), config); } catch (Exception e) { } // just carry on try { Path scriptToCache = InfiniteHadoopUtils.cacheLocalFile("/opt/infinite-home/scripts/", "custom_reduce_error_handler.sh", config); config.set("mapred.reduce.task.debug.script", "custom_reduce_error_handler.sh " + job.jobtitle); config.set("mapreduce.reduce.debug.script", "custom_reduce_error_handler.sh " + job.jobtitle); DistributedCache.createSymlink(config); DistributedCache.addCacheFile(scriptToCache.toUri(), config); } catch (Exception e) { } // just carry on } //TODO (???): TOTEST // (need to do these 2 things here before the job is created, at which point the config class has been copied across) //1) Class<?> mapperClazz = Class.forName(job.mapper, true, child); if (ICustomInfiniteInternalEngine.class.isAssignableFrom(mapperClazz)) { // Special case: internal custom engine, so gets an additional integration hook ICustomInfiniteInternalEngine preActivities = (ICustomInfiniteInternalEngine) mapperClazz .newInstance(); preActivities.preTaskActivities(job._id, job.communityIds, config, !(bTestMode || bLocalMode)); } //TESTED //2) if (job.inputCollection.equalsIgnoreCase("file.binary_shares")) { // Need to download the GridFSZip file try { Path jarToCache = InfiniteHadoopUtils.cacheLocalFile("/opt/infinite-home/lib/unbundled/", "GridFSZipFile.jar", config); DistributedCache.addFileToClassPath(jarToCache, config); } catch (Throwable t) { } // (this is fine, will already be on the classpath .. otherwise lots of other stuff will be failing all over the place!) } if (job.inputCollection.equals("records")) { InfiniteElasticsearchHadoopUtils.handleElasticsearchInput(job, config, advancedConfigurationDbo); //(won't run under 0.19 so running with "records" should cause all sorts of exceptions) } //TESTED (by hand) if (bTestMode || bLocalMode) { // If running locally, turn "snappy" off - tomcat isn't pointing its native library path in the right place config.set("mapred.map.output.compression.codec", "org.apache.hadoop.io.compress.DefaultCodec"); } // Manually specified caches List<URL> localJarCaches = InfiniteHadoopUtils.handleCacheList(advancedConfigurationDbo.get("$caches"), job, config, props_custom); Job hj = new Job(config); // (NOTE: from here, changes to config are ignored) try { if (null != localJarCaches) { if (bLocalMode || bTestMode) { Method method = URLClassLoader.class.getDeclaredMethod("addURL", new Class[] { URL.class }); method.setAccessible(true); method.invoke(child, localJarCaches.toArray()); } //TOTEST (tested logically) } Class<?> classToLoad = Class.forName(job.mapper, true, child); hj.setJarByClass(classToLoad); if (job.inputCollection.equalsIgnoreCase("filesystem")) { String inputPath = null; try { inputPath = MongoDbUtil.getProperty(advancedConfigurationDbo, "file.url"); if (!inputPath.endsWith("/")) { inputPath = inputPath + "/"; } } catch (Exception e) { } if (null == inputPath) { throw new RuntimeException("Must specify 'file.url' if reading from filesystem."); } inputPath = InfiniteHadoopUtils.authenticateInputDirectory(job, inputPath); InfiniteFileInputFormat.addInputPath(hj, new Path(inputPath + "*/*")); // (that extra bit makes it recursive) InfiniteFileInputFormat.setMaxInputSplitSize(hj, 33554432); // (32MB) InfiniteFileInputFormat.setInfiniteInputPathFilter(hj, config); hj.setInputFormatClass((Class<? extends InputFormat>) Class.forName( "com.ikanow.infinit.e.data_model.custom.InfiniteFileInputFormat", true, child)); } else if (job.inputCollection.equalsIgnoreCase("file.binary_shares")) { String[] oidStrs = null; try { String inputPath = MongoDbUtil.getProperty(advancedConfigurationDbo, "file.url"); Pattern oidExtractor = Pattern.compile("inf://share/([^/]+)"); Matcher m = oidExtractor.matcher(inputPath); if (m.find()) { oidStrs = m.group(1).split("\\s*,\\s*"); } else { throw new RuntimeException( "file.url must be in format inf://share/<oid-list>/<string>: " + inputPath); } InfiniteHadoopUtils.authenticateShareList(job, oidStrs); } catch (Exception e) { throw new RuntimeException( "Authentication error: " + e.getMessage() + ": " + advancedConfigurationDbo, e); } hj.getConfiguration().setStrings("mapred.input.dir", oidStrs); hj.setInputFormatClass((Class<? extends InputFormat>) Class.forName( "com.ikanow.infinit.e.data_model.custom.InfiniteShareInputFormat", true, child)); } else if (job.inputCollection.equals("records")) { hj.setInputFormatClass((Class<? extends InputFormat>) Class .forName("com.ikanow.infinit.e.data_model.custom.InfiniteEsInputFormat", true, child)); } else { if (esMode) { hj.setInputFormatClass((Class<? extends InputFormat>) Class.forName( "com.ikanow.infinit.e.processing.custom.utils.InfiniteElasticsearchMongoInputFormat", true, child)); } else { hj.setInputFormatClass((Class<? extends InputFormat>) Class.forName( "com.ikanow.infinit.e.data_model.custom.InfiniteMongoInputFormat", true, child)); } } if ((null != job.exportToHdfs) && job.exportToHdfs) { //TODO (INF-2469): Also, if the output key is BSON then also run as text (but output as JSON?) Path outPath = InfiniteHadoopUtils.ensureOutputDirectory(job, props_custom); if ((null != job.outputKey) && (null != job.outputValue) && job.outputKey.equalsIgnoreCase("org.apache.hadoop.io.text") && job.outputValue.equalsIgnoreCase("org.apache.hadoop.io.text")) { // (slight hack before I sort out the horrendous job class - if key/val both text and exporting to HDFS then output as Text) hj.setOutputFormatClass((Class<? extends OutputFormat>) Class .forName("org.apache.hadoop.mapreduce.lib.output.TextOutputFormat", true, child)); TextOutputFormat.setOutputPath(hj, outPath); } //TESTED else { hj.setOutputFormatClass((Class<? extends OutputFormat>) Class.forName( "org.apache.hadoop.mapreduce.lib.output.SequenceFileOutputFormat", true, child)); SequenceFileOutputFormat.setOutputPath(hj, outPath); } //TESTED } else { // normal case, stays in MongoDB hj.setOutputFormatClass((Class<? extends OutputFormat>) Class.forName( "com.ikanow.infinit.e.data_model.custom.InfiniteMongoOutputFormat", true, child)); } hj.setMapperClass((Class<? extends Mapper>) mapperClazz); String mapperOutputKeyOverride = advancedConfigurationDbo.getString("$mapper_key_class", null); if (null != mapperOutputKeyOverride) { hj.setMapOutputKeyClass(Class.forName(mapperOutputKeyOverride)); } //TESTED String mapperOutputValueOverride = advancedConfigurationDbo.getString("$mapper_value_class", null); if (null != mapperOutputValueOverride) { hj.setMapOutputValueClass(Class.forName(mapperOutputValueOverride)); } //TESTED if ((null != job.reducer) && !job.reducer.startsWith("#") && !job.reducer.equalsIgnoreCase("null") && !job.reducer.equalsIgnoreCase("none")) { hj.setReducerClass((Class<? extends Reducer>) Class.forName(job.reducer, true, child)); // Variable reducers: if (null != job.query) { try { hj.setNumReduceTasks(advancedConfigurationDbo.getInt("$reducers", 1)); } catch (Exception e) { try { // (just check it's not a string that is a valid int) hj.setNumReduceTasks( Integer.parseInt(advancedConfigurationDbo.getString("$reducers", "1"))); } catch (Exception e2) { } } } //TESTED } else { hj.setNumReduceTasks(0); } if ((null != job.combiner) && !job.combiner.startsWith("#") && !job.combiner.equalsIgnoreCase("null") && !job.combiner.equalsIgnoreCase("none")) { hj.setCombinerClass((Class<? extends Reducer>) Class.forName(job.combiner, true, child)); } hj.setOutputKeyClass(Class.forName(job.outputKey, true, child)); hj.setOutputValueClass(Class.forName(job.outputValue, true, child)); hj.setJobName(job.jobtitle); currJobName = job.jobtitle; } catch (Error e) { // (messing about with class loaders = lots of chances for errors!) throw new RuntimeException(e.getMessage(), e); } if (bTestMode || bLocalMode) { hj.submit(); currThreadId = null; Logger.getRootLogger().addAppender(this); currLocalJobId = hj.getJobID().toString(); currLocalJobErrs.setLength(0); while (!hj.isComplete()) { Thread.sleep(1000); } Logger.getRootLogger().removeAppender(this); if (hj.isSuccessful()) { if (this.currLocalJobErrs.length() > 0) { return "local_done: " + this.currLocalJobErrs.toString(); } else { return "local_done"; } } else { return "Error: " + this.currLocalJobErrs.toString(); } } else { hj.submit(); String jobId = hj.getJobID().toString(); return jobId; } } catch (Exception e) { e.printStackTrace(); Thread.currentThread().setContextClassLoader(savedClassLoader); return "Error: " + InfiniteHadoopUtils.createExceptionMessage(e); } finally { Thread.currentThread().setContextClassLoader(savedClassLoader); } }
From source file:com.inmobi.conduit.distcp.tools.DistCp.java
License:Apache License
/** * Implements the core-execution. Creates the file-list for copy, * and launches the Hadoop-job, to do the copy. * @return Job handle/*w ww . j a va2s . com*/ * @throws Exception, on failure. */ public Job execute() throws Exception { assert inputOptions != null; assert getConf() != null; Job job = null; try { metaFolder = createMetaFolderPath(); jobFS = metaFolder.getFileSystem(getConf()); job = createJob(); createInputFileListing(job); job.submit(); submitted = true; } finally { if (!submitted) { cleanup(); } } String jobID = getJobID(job); job.getConfiguration().set(DistCpConstants.CONF_LABEL_DISTCP_JOB_ID, jobID); LOG.info("DistCp job-id: " + jobID); LOG.info("DistCp job may be tracked at: " + job.getTrackingURL()); LOG.info("To cancel, run the following command:\thadoop job -kill " + jobID); long jobStartTime = System.nanoTime(); if (inputOptions.shouldBlock() && !job.waitForCompletion(true)) { updateJobTimeInNanos(jobStartTime); throw new IOException("DistCp failure: Job " + jobID + " has failed. "); } updateJobTimeInNanos(jobStartTime); return job; }
From source file:com.inmobi.conduit.distcp.tools.mapred.TestCopyCommitter.java
License:Apache License
@Ignore @Test//from w w w . j av a2 s . co m public void testCounterProvider() { try { Job job = MockJobTracker.getJobForClient(); Counters a = EMPTY_COUNTERS; CounterGroup grp = a.getGroup("abc"); Counter cntr = grp.findCounter("counter"); cntr.increment(100); CounterProvider cp = new CounterProvider(a); job.submit(); Assert.assertEquals(job.getCounters(), a); } catch (Exception e) { LOG.error("Exception encountered ", e); } }
From source file:com.kylinolap.job.hadoop.AbstractHadoopJob.java
License:Apache License
protected int waitForCompletion(Job job) throws IOException, InterruptedException, ClassNotFoundException { int retVal = 0; long start = System.nanoTime(); if (isAsync) { job.submit(); } else {/*from www . j a va2s. co m*/ job.waitForCompletion(true); retVal = job.isSuccessful() ? 0 : 1; } log.debug("Job '" + job.getJobName() + "' finished " + (job.isSuccessful() ? "successfully in " : "with failures. Time taken ") + StringUtils.formatTime((System.nanoTime() - start) / 1000000L)); return retVal; }
From source file:com.lakhani.anchorgraph.anchorgraph.java
public int run(String[] args) throws Exception { Configuration conf = new Configuration(); //DistributedCache.addCacheFile(new URI("hdfs://zphdc1n1:8020/user/clakhani/anchorgraph/centroids.txt"), conf); conf.set("numberCentroids", args[3]); conf.set("numberFeatures", args[4]); Job job = new Job(conf, "anchorgraph"); job.addCacheFile(new URI(args[2])); job.setOutputKeyClass(Text.class); job.setOutputValueClass(Text.class); job.setMapperClass(Map.class); job.setInputFormatClass(TextInputFormat.class); job.setOutputFormatClass(TextOutputFormat.class); FileInputFormat.addInputPath(job, new Path(args[0])); FileOutputFormat.setOutputPath(job, new Path(args[2])); job.setJarByClass(anchorgraph.class); job.submit(); int rc = (job.waitForCompletion(true)) ? 1 : 0; return rc;//from w w w . ja v a2 s .c om }
From source file:com.lakhani.anchorgraph.testCache.java
public int run(String[] args) throws Exception { Configuration conf = new Configuration(); Job job = new Job(conf, "testCache"); job.addCacheFile(new URI("hdfs://zphdc1n1:8020/user/clakhani/anchorgraph/centroids.txt")); job.setOutputKeyClass(Text.class); job.setOutputValueClass(Text.class); job.setMapperClass(Map.class); job.setInputFormatClass(TextInputFormat.class); job.setOutputFormatClass(TextOutputFormat.class); FileInputFormat.addInputPath(job, new Path(args[0])); FileOutputFormat.setOutputPath(job, new Path(args[1])); job.setJarByClass(testCache.class); job.submit(); int rc = (job.waitForCompletion(true)) ? 1 : 0; return rc;/* ww w .j a v a 2 s. c om*/ }
From source file:com.metamx.druid.indexer.DeterminePartitionsJob.java
License:Open Source License
public boolean run() { try {//from w w w . j ava 2 s.c o m /* * Group by (timestamp, dimensions) so we can correctly count dimension values as they would appear * in the final segment. */ if (!config.getPartitionsSpec().isAssumeGrouped()) { final Job groupByJob = new Job(new Configuration(), String.format( "%s-determine_partitions_groupby-%s", config.getDataSource(), config.getIntervals())); injectSystemProperties(groupByJob); groupByJob.setInputFormatClass(TextInputFormat.class); groupByJob.setMapperClass(DeterminePartitionsGroupByMapper.class); groupByJob.setMapOutputKeyClass(BytesWritable.class); groupByJob.setMapOutputValueClass(NullWritable.class); groupByJob.setCombinerClass(DeterminePartitionsGroupByReducer.class); groupByJob.setReducerClass(DeterminePartitionsGroupByReducer.class); groupByJob.setOutputKeyClass(BytesWritable.class); groupByJob.setOutputValueClass(NullWritable.class); groupByJob.setOutputFormatClass(SequenceFileOutputFormat.class); groupByJob.setJarByClass(DeterminePartitionsJob.class); config.addInputPaths(groupByJob); config.intoConfiguration(groupByJob); FileOutputFormat.setOutputPath(groupByJob, config.makeGroupedDataDir()); groupByJob.submit(); log.info("Job %s submitted, status available at: %s", groupByJob.getJobName(), groupByJob.getTrackingURL()); if (!groupByJob.waitForCompletion(true)) { log.error("Job failed: %s", groupByJob.getJobID()); return false; } } else { log.info("Skipping group-by job."); } /* * Read grouped data and determine appropriate partitions. */ final Job dimSelectionJob = new Job(new Configuration(), String.format( "%s-determine_partitions_dimselection-%s", config.getDataSource(), config.getIntervals())); dimSelectionJob.getConfiguration().set("io.sort.record.percent", "0.19"); injectSystemProperties(dimSelectionJob); if (!config.getPartitionsSpec().isAssumeGrouped()) { // Read grouped data from the groupByJob. dimSelectionJob.setMapperClass(DeterminePartitionsDimSelectionPostGroupByMapper.class); dimSelectionJob.setInputFormatClass(SequenceFileInputFormat.class); FileInputFormat.addInputPath(dimSelectionJob, config.makeGroupedDataDir()); } else { // Directly read the source data, since we assume it's already grouped. dimSelectionJob.setMapperClass(DeterminePartitionsDimSelectionAssumeGroupedMapper.class); dimSelectionJob.setInputFormatClass(TextInputFormat.class); config.addInputPaths(dimSelectionJob); } SortableBytes.useSortableBytesAsMapOutputKey(dimSelectionJob); dimSelectionJob.setMapOutputValueClass(Text.class); dimSelectionJob.setCombinerClass(DeterminePartitionsDimSelectionCombiner.class); dimSelectionJob.setReducerClass(DeterminePartitionsDimSelectionReducer.class); dimSelectionJob.setOutputKeyClass(BytesWritable.class); dimSelectionJob.setOutputValueClass(Text.class); dimSelectionJob.setOutputFormatClass(DeterminePartitionsDimSelectionOutputFormat.class); dimSelectionJob.setJarByClass(DeterminePartitionsJob.class); config.intoConfiguration(dimSelectionJob); FileOutputFormat.setOutputPath(dimSelectionJob, config.makeIntermediatePath()); dimSelectionJob.submit(); log.info("Job %s submitted, status available at: %s", dimSelectionJob.getJobName(), dimSelectionJob.getTrackingURL()); if (!dimSelectionJob.waitForCompletion(true)) { log.error("Job failed: %s", dimSelectionJob.getJobID().toString()); return false; } /* * Load partitions determined by the previous job. */ log.info("Job completed, loading up partitions for intervals[%s].", config.getSegmentGranularIntervals()); FileSystem fileSystem = null; Map<DateTime, List<HadoopyShardSpec>> shardSpecs = Maps.newTreeMap(DateTimeComparator.getInstance()); int shardCount = 0; for (Interval segmentGranularity : config.getSegmentGranularIntervals()) { DateTime bucket = segmentGranularity.getStart(); final Path partitionInfoPath = config.makeSegmentPartitionInfoPath(new Bucket(0, bucket, 0)); if (fileSystem == null) { fileSystem = partitionInfoPath.getFileSystem(dimSelectionJob.getConfiguration()); } if (fileSystem.exists(partitionInfoPath)) { List<ShardSpec> specs = config.jsonMapper.readValue( Utils.openInputStream(dimSelectionJob, partitionInfoPath), new TypeReference<List<ShardSpec>>() { }); List<HadoopyShardSpec> actualSpecs = Lists.newArrayListWithExpectedSize(specs.size()); for (int i = 0; i < specs.size(); ++i) { actualSpecs.add(new HadoopyShardSpec(specs.get(i), shardCount++)); log.info("DateTime[%s], partition[%d], spec[%s]", bucket, i, actualSpecs.get(i)); } shardSpecs.put(bucket, actualSpecs); } else { log.info("Path[%s] didn't exist!?", partitionInfoPath); } } config.setShardSpecs(shardSpecs); return true; } catch (Exception e) { throw Throwables.propagate(e); } }
From source file:com.metamx.druid.indexer.IndexGeneratorJob.java
License:Open Source License
public boolean run() { try {/*w w w. j a va 2s. com*/ Job job = new Job(new Configuration(), String.format("%s-index-generator-%s", config.getDataSource(), config.getIntervals())); job.getConfiguration().set("io.sort.record.percent", "0.23"); for (String propName : System.getProperties().stringPropertyNames()) { Configuration conf = job.getConfiguration(); if (propName.startsWith("hadoop.")) { conf.set(propName.substring("hadoop.".length()), System.getProperty(propName)); } } job.setInputFormatClass(TextInputFormat.class); job.setMapperClass(IndexGeneratorMapper.class); job.setMapOutputValueClass(Text.class); SortableBytes.useSortableBytesAsMapOutputKey(job); job.setNumReduceTasks(Iterables.size(config.getAllBuckets())); job.setPartitionerClass(IndexGeneratorPartitioner.class); job.setReducerClass(IndexGeneratorReducer.class); job.setOutputKeyClass(BytesWritable.class); job.setOutputValueClass(Text.class); job.setOutputFormatClass(IndexGeneratorOutputFormat.class); FileOutputFormat.setOutputPath(job, config.makeIntermediatePath()); config.addInputPaths(job); config.intoConfiguration(job); job.setJarByClass(IndexGeneratorJob.class); job.submit(); log.info("Job %s submitted, status available at %s", job.getJobName(), job.getTrackingURL()); boolean success = job.waitForCompletion(true); Counter invalidRowCount = job.getCounters() .findCounter(HadoopDruidIndexerConfig.IndexJobCounters.INVALID_ROW_COUNTER); jobStats.setInvalidRowCount(invalidRowCount.getValue()); return success; } catch (Exception e) { throw new RuntimeException(e); } }
From source file:com.mongodb.hadoop.examples.ufos.UfoSightings.java
License:Apache License
public int run(final String[] args) throws Exception { final Configuration conf = getConf(); final Job job = new Job(conf, "ufo-sightings"); job.setMapperClass(UfoSightingsMapper.class); job.setReducerClass(UfoSightingsReducer.class); job.setOutputFormatClass(com.mongodb.hadoop.MongoOutputFormat.class); job.setOutputKeyClass(org.apache.hadoop.io.Text.class); job.setOutputValueClass(org.apache.hadoop.io.IntWritable.class); job.setInputFormatClass(com.mongodb.hadoop.MongoInputFormat.class); final boolean verbose = true; try {//from w ww . jav a 2s . c o m if (BACKGROUND) { LOG.info("Setting up and running MapReduce job in background."); job.submit(); return 0; } else { LOG.info("Setting up and running MapReduce job in foreground, will wait for results. {Verbose? " + verbose + "}"); return job.waitForCompletion(true) ? 0 : 1; } } catch (final Exception e) { LOG.error("Exception while executing job... ", e); return 1; } }