List of usage examples for org.apache.hadoop.fs FileSystem getWorkingDirectory
public abstract Path getWorkingDirectory();
From source file:root.input.reuters21578.VectorizationJob.java
License:Apache License
/** * {@inheritDoc}// w ww.j ava 2 s .c o m */ @Override public int run(String[] args) throws Exception { constructParameterList(); if (parseArguments(args) == null) { return -1; } initializeConfigurationParameters(); printJobHeader(); Configuration conf = getConf(); URI workingURI = new URI(conf.get("fs.default.name")); URI inputURI = new URI(inputDirectory); FileSystem workingFS = FileSystem.get(workingURI, conf); FileSystem inputFS = FileSystem.get(inputURI, conf); String workingDirectory = workingFS.getWorkingDirectory() + "/"; outputDirectory = workingDirectory + outputDirectory; Path inputDirectoryPath = new Path(inputDirectory); Path outputDirectoryPath = new Path(outputDirectory); if (!inputFS.exists(inputDirectoryPath)) { throw new Exception("Input directory not found."); } if (workingFS.delete(outputDirectoryPath, true)) { System.out.println("Output directory cleaned."); } sequenceFilesDirectory = outputDirectory + sequenceFilesDirectory; vectorDirectory = outputDirectory + vectorDirectory; similarityMatrixDirectory = outputDirectory + similarityMatrixDirectory; filenameDictionaryDirectory = outputDirectory + filenameDictionaryDirectory; renamedInputdirectory = outputDirectory + renamedInputdirectory; // 1: Renames files 1-N System.out.println(); System.out.println("--------------"); System.out.println("Renaming Files"); System.out.println("--------------"); System.out.println("\tInput: " + inputDirectory); System.out.println("\tOutput: " + renamedInputdirectory); System.out.println(); String[] arguments_renameFiles = { "-i", inputDirectory, "-o", renamedInputdirectory, "-f", filenameDictionaryDirectory }; ToolRunner.run(new RenameFilesJob(), arguments_renameFiles); // 2: Converts text to sequence file System.out.println(); System.out.println("--------------------------------------"); System.out.println("Creating Sequence Files From Directory"); System.out.println("--------------------------------------"); System.out.println("\tInput: " + renamedInputdirectory); System.out.println("\tOutput: " + sequenceFilesDirectory); System.out.println(); String[] arguments_SequenceFilesFromDirectory = { "-i", renamedInputdirectory, "-o", sequenceFilesDirectory }; ToolRunner.run(new SequenceFilesFromDirectory(), arguments_SequenceFilesFromDirectory); // 3: Creates vectors of text System.out.println(); System.out.println("----------------"); System.out.println("Creating Vectors"); System.out.println("----------------"); System.out.println("\tInput: " + sequenceFilesDirectory); System.out.println("\tOutput: " + vectorDirectory); System.out.println(); String[] arguments_SparseVectorsFromSequenceFiles = { "-i", sequenceFilesDirectory, "-o", vectorDirectory, "-x", exclusionThreshold, "-md", minimumDocumentFrequency }; ToolRunner.run(new SparseVectorsFromSequenceFiles(), arguments_SparseVectorsFromSequenceFiles); // 4: Create a similarity matrix. System.out.println(); System.out.println("--------------------------"); System.out.println("Creating Similarity Matrix"); System.out.println("--------------------------"); System.out.println("\tInput: " + vectorDirectory + "/" + tf_tfidf + "-vectors"); System.out.println("\tOutput: " + similarityMatrixDirectory); System.out.println("\tLevels: " + numLevels); System.out.println(); String[] arguments_CreateSimilaritySimilarityJob = { "-i", vectorDirectory + "/" + tf_tfidf + "-vectors", "-o", similarityMatrixDirectory, "-dm", distanceMetric, "-l", numLevels, "-smd", diagScale }; ToolRunner.run(new CreateSimilarityMatrixJob(), arguments_CreateSimilaritySimilarityJob); return 0; }
From source file:uk.ac.gla.terrier.probos.controller.ControllerServer.java
License:Open Source License
public ControllerServer(Configuration _hconf) throws IOException { this.yConf = new YarnConfiguration(_hconf); yConf.addResource("yarn-site.xml"); UserGroupInformation.setConfiguration(yConf); this.pConf = new PConfiguration(_hconf); //do the Kerberos authentication if (UserGroupInformation.isSecurityEnabled()) { final String principal = pConf.get(PConfiguration.KEY_CONTROLLER_PRINCIPAL); String keytab = pConf.get(PConfiguration.KEY_CONTROLLER_KEYTAB); File fKeytab = new File(keytab); if (!fKeytab.exists()) { if (!fKeytab.isAbsolute()) { keytab = System.getProperty("probos.conf") + '/' + keytab; fKeytab = new File(keytab); pConf.set(PConfiguration.KEY_CONTROLLER_KEYTAB, keytab); }/* www. j a va 2 s . c om*/ if (!fKeytab.exists()) throw new FileNotFoundException("Could not find keytab file " + keytab); } LOG.debug("Starting login for " + principal + " using keytab " + keytab); SecurityUtil.login(pConf, PConfiguration.KEY_CONTROLLER_KEYTAB, PConfiguration.KEY_CONTROLLER_PRINCIPAL, Utils.getHostname()); LOG.info("Switched principal to " + UserGroupInformation.getCurrentUser().getUserName()); } this.mClient = MailClient.getMailClient(this.pConf); final String bindAddress = pConf.get(PConfiguration.KEY_CONTROLLER_BIND_ADDRESS); if (bindAddress == null) throw new IllegalArgumentException(PConfiguration.KEY_CONTROLLER_BIND_ADDRESS + " cannot be null"); secretManager = new ControllerAPISecretManager( //delegationKeyUpdateInterval //renewal interval for delegation token 7 * 24 * 3600 * 1000, //Yarn default is 7 day //delegationTokenMaxLifetime -- maximum lifetime for which a delegation token is valid //i.e. how long can we keep renewing the token for? 14 * 24 * 3600 * 1000, //Yarn default is 14 days //delegationTokenRenewInterval -- how long should a token last? 7 * 24 * 3600 * 1000, //Yarn default is 7 day //delegationTokenRemoverScanInterval -- how often are expired keys removed? 3600 * 1000); //Yarn default is 1 hour //build the client rpc server: 8027 int port = pConf.getInt(PConfiguration.KEY_CONTROLLER_PORT, 8027); LOG.info("Starting RPC server for " + PBSClient.class.getSimpleName() + " on port " + port); clientRpcserver = new RPC.Builder(yConf).setInstance(this).setBindAddress(bindAddress) .setProtocol(PBSClient.class).setPort(port).setSecretManager(secretManager). //setVerbose(true). build(); System.setProperty("hadoop.policy.file", Constants.PRODUCT_NAME + "-policy.xml"); clientRpcserver.refreshServiceAclWithLoadedConfiguration(yConf, new ControllerPolicyProvider()); //build the master rpc server: 8028 port = Constants.CONTROLLER_MASTER_PORT_OFFSET + pConf.getInt(PConfiguration.KEY_CONTROLLER_PORT, 8027); LOG.info("Starting RPC server for " + PBSMasterClient.class.getSimpleName() + " on port " + port); masterRpcserver = new RPC.Builder(yConf).setInstance(new ApplicationMasterAPI()).setBindAddress(bindAddress) .setProtocol(PBSMasterClient.class).setPort(port).setSecretManager(secretManager). //setVerbose(true). build(); masterRpcserver.refreshServiceAclWithLoadedConfiguration(yConf, new ControllerPolicyProvider()); port = Constants.CONTROLLER_INTERACTIVE_PORT_OFFSET + pConf.getInt(PConfiguration.KEY_CONTROLLER_PORT, 8027); LOG.info("Starting RPC server for " + PBSInteractiveClient.class.getSimpleName() + " on port " + port); //build the interactive rpc server: 8026 interactiveRpcserver = new RPC.Builder(yConf).setInstance(new InteractiveTaskAPI()) .setBindAddress(bindAddress).setProtocol(PBSInteractiveClient.class).setPort(port) .setSecretManager(secretManager). //setVerbose(true). build(); interactiveRpcserver.refreshServiceAclWithLoadedConfiguration(yConf, new ControllerPolicyProvider()); //build the webapp UI server final List<Entry<String, HttpServlet>> controllerServlets = new ArrayList<>(); controllerServlets .add(new MapEntry<String, HttpServlet>("/", new QstatServlet("/", controllerServlets, this))); controllerServlets.add( new MapEntry<String, HttpServlet>("/pbsnodes", new PbsnodesServlet("/", controllerServlets, this))); //metrics is the Servlet from metrics.dropwizard for accessing metrics controllerServlets.add(new MapEntry<String, HttpServlet>("/metrics", new MetricsServlet(metrics))); //this is the hadoop servlet for accessing anything defined in JMX controllerServlets.add(new MapEntry<String, HttpServlet>("/jmx", new JMXJsonServlet())); final int httpport = pConf.getInt(PConfiguration.KEY_CONTROLLER_HTTP_PORT, Constants.DEFAULT_CONTROLLER_PORT + Constants.CONTROLLER_HTTP_PORT_OFFSET); LOG.info("Starting Jetty ProbosControllerHttp on port " + httpport); webServer = new WebServer("ProbosControllerHttp", controllerServlets, httpport); webServer.init(pConf); //this thread detects yarn jobs that have ended watcherThread = new Thread(new ControllerWatcher()); watcherThread.setName(ControllerWatcher.class.getSimpleName()); //ensure we have the directory Path _probosFolder = new Path(pConf.get(PConfiguration.KEY_CONTROLLER_JOBDIR)); FileSystem controllerFS = FileSystem.get(yConf); if (!_probosFolder.isUriPathAbsolute()) { _probosFolder = _probosFolder.makeQualified(controllerFS.getUri(), controllerFS.getWorkingDirectory()); assert _probosFolder.isUriPathAbsolute(); } probosFolder = _probosFolder; if (!controllerFS.exists(probosFolder)) { throw new IllegalArgumentException(probosFolder.toString() + " does not exist"); } //now initialise the metrics //jobs.queued.size metrics.register(MetricRegistry.name(ControllerServer.class, "jobs", "queued.size"), new Gauge<Integer>() { @Override public Integer getValue() { int sum = 0; for (int i : user2QueuedCount.values()) sum += i; return sum; } }); //jobs.size metrics.register(MetricRegistry.name(ControllerServer.class, "jobs", "size"), new Gauge<Integer>() { @Override public Integer getValue() { return jobArray.size(); } }); //jobs.held.size metrics.register(MetricRegistry.name(ControllerServer.class, "jobs", "held.size"), new Gauge<Integer>() { @Override public Integer getValue() { return jobHolds.size(); } }); //nodes.size metrics.register(MetricRegistry.name(ControllerServer.class, "nodes", "size"), new Gauge<Integer>() { @Override public Integer getValue() { try { return getNodesStatus().length; } catch (Exception e) { return 0; } } }); //nodes.free.size metrics.register(MetricRegistry.name(ControllerServer.class, "nodes", "free.size"), new Gauge<Integer>() { @Override public Integer getValue() { try { PBSNodeStatus[] nodes = getNodesStatus(); int count = 0; for (PBSNodeStatus n : nodes) if ("free".equals(n.getState())) count++; return count; } catch (Exception e) { return 0; } } }); runningJobs = metrics.counter(MetricRegistry.name(ControllerServer.class, "jobs", "running.counter")); rejectedJobs = metrics.counter(MetricRegistry.name(ControllerServer.class, "jobs", "rejected.counter")); killedJobs = metrics.counter(MetricRegistry.name(ControllerServer.class, "jobs", "killed.counter")); mailEvents = metrics.counter(MetricRegistry.name(ControllerServer.class, "mails", "counter")); mailFailures = metrics.counter(MetricRegistry.name(ControllerServer.class, "mails", "failure.counter")); }
From source file:yrun.YarnRunner.java
License:Apache License
private Path makeQualified(Path p, FileSystem fileSystem) { return p.makeQualified(fileSystem.getUri(), fileSystem.getWorkingDirectory()); }