List of usage examples for org.apache.hadoop.yarn.client.api YarnClient submitApplication
public abstract ApplicationId submitApplication(ApplicationSubmissionContext appContext) throws YarnException, IOException;
Submit a new application to YARN.
It is a blocking call - it will not return ApplicationId until the submitted application is submitted successfully and accepted by the ResourceManager.
From source file:org.apache.flink.yarn.AbstractYarnClusterDescriptor.java
License:Apache License
public ApplicationReport startAppMaster(JobGraph jobGraph, YarnClient yarnClient, YarnClientApplication yarnApplication) throws Exception { // ------------------ Set default file system scheme ------------------------- try {/*from w w w .ja v a 2 s .c om*/ org.apache.flink.core.fs.FileSystem.setDefaultScheme(flinkConfiguration); } catch (IOException e) { throw new IOException("Error while setting the default " + "filesystem scheme from configuration.", e); } // initialize file system // Copy the application master jar to the filesystem // Create a local resource to point to the destination jar path final FileSystem fs = FileSystem.get(conf); // hard coded check for the GoogleHDFS client because its not overriding the getScheme() method. if (!fs.getClass().getSimpleName().equals("GoogleHadoopFileSystem") && fs.getScheme().startsWith("file")) { LOG.warn("The file system scheme is '" + fs.getScheme() + "'. This indicates that the " + "specified Hadoop configuration path is wrong and the system is using the default Hadoop configuration values." + "The Flink YARN client needs to store its files in a distributed file system"); } ApplicationSubmissionContext appContext = yarnApplication.getApplicationSubmissionContext(); Set<File> effectiveShipFiles = new HashSet<>(shipFiles.size()); for (File file : shipFiles) { effectiveShipFiles.add(file.getAbsoluteFile()); } //check if there is a logback or log4j file File logbackFile = new File(configurationDirectory + File.separator + CONFIG_FILE_LOGBACK_NAME); final boolean hasLogback = logbackFile.exists(); if (hasLogback) { effectiveShipFiles.add(logbackFile); } File log4jFile = new File(configurationDirectory + File.separator + CONFIG_FILE_LOG4J_NAME); final boolean hasLog4j = log4jFile.exists(); if (hasLog4j) { effectiveShipFiles.add(log4jFile); if (hasLogback) { // this means there is already a logback configuration file --> fail LOG.warn("The configuration directory ('" + configurationDirectory + "') contains both LOG4J and " + "Logback configuration files. Please delete or rename one of them."); } } addLibFolderToShipFiles(effectiveShipFiles); // add the user jar to the classpath of the to-be-created cluster if (userJarFiles != null) { effectiveShipFiles.addAll(userJarFiles); } // Set-up ApplicationSubmissionContext for the application final ApplicationId appId = appContext.getApplicationId(); // ------------------ Add Zookeeper namespace to local flinkConfiguraton ------ String zkNamespace = getZookeeperNamespace(); // no user specified cli argument for namespace? if (zkNamespace == null || zkNamespace.isEmpty()) { // namespace defined in config? else use applicationId as default. zkNamespace = flinkConfiguration.getString(HighAvailabilityOptions.HA_CLUSTER_ID, String.valueOf(appId)); setZookeeperNamespace(zkNamespace); } flinkConfiguration.setString(HighAvailabilityOptions.HA_CLUSTER_ID, zkNamespace); if (HighAvailabilityMode.isHighAvailabilityModeActivated(flinkConfiguration)) { // activate re-execution of failed applications appContext.setMaxAppAttempts(flinkConfiguration.getInteger(ConfigConstants.YARN_APPLICATION_ATTEMPTS, YarnConfiguration.DEFAULT_RM_AM_MAX_ATTEMPTS)); activateHighAvailabilitySupport(appContext); } else { // set number of application retries to 1 in the default case appContext .setMaxAppAttempts(flinkConfiguration.getInteger(ConfigConstants.YARN_APPLICATION_ATTEMPTS, 1)); } // local resource map for Yarn final Map<String, LocalResource> localResources = new HashMap<>(2 + effectiveShipFiles.size()); // list of remote paths (after upload) final List<Path> paths = new ArrayList<>(2 + effectiveShipFiles.size()); // classpath assembler final StringBuilder classPathBuilder = new StringBuilder(); // ship list that enables reuse of resources for task manager containers StringBuilder envShipFileList = new StringBuilder(); // upload and register ship files for (File shipFile : effectiveShipFiles) { LocalResource shipResources = Records.newRecord(LocalResource.class); Path shipLocalPath = new Path("file://" + shipFile.getAbsolutePath()); Path remotePath = Utils.setupLocalResource(fs, appId.toString(), shipLocalPath, shipResources, fs.getHomeDirectory()); paths.add(remotePath); localResources.put(shipFile.getName(), shipResources); if (shipFile.isDirectory()) { // add directories to the classpath java.nio.file.Path shipPath = shipFile.toPath(); final java.nio.file.Path parentPath = shipPath.getParent(); Files.walkFileTree(shipPath, new SimpleFileVisitor<java.nio.file.Path>() { @Override public FileVisitResult preVisitDirectory(java.nio.file.Path dir, BasicFileAttributes attrs) throws IOException { super.preVisitDirectory(dir, attrs); java.nio.file.Path relativePath = parentPath.relativize(dir); classPathBuilder.append(relativePath).append(File.separator).append("*") .append(File.pathSeparator); return FileVisitResult.CONTINUE; } }); } else { // add files to the classpath classPathBuilder.append(shipFile.getName()).append(File.pathSeparator); } envShipFileList.append(remotePath).append(","); } // Setup jar for ApplicationMaster LocalResource appMasterJar = Records.newRecord(LocalResource.class); LocalResource flinkConf = Records.newRecord(LocalResource.class); Path remotePathJar = Utils.setupLocalResource(fs, appId.toString(), flinkJarPath, appMasterJar, fs.getHomeDirectory()); Path remotePathConf = Utils.setupLocalResource(fs, appId.toString(), flinkConfigurationPath, flinkConf, fs.getHomeDirectory()); localResources.put("flink.jar", appMasterJar); localResources.put("flink-conf.yaml", flinkConf); paths.add(remotePathJar); classPathBuilder.append("flink.jar").append(File.pathSeparator); paths.add(remotePathConf); classPathBuilder.append("flink-conf.yaml").append(File.pathSeparator); // write job graph to tmp file and add it to local resource // TODO: server use user main method to generate job graph if (jobGraph != null) { try { File fp = File.createTempFile(appId.toString(), null); fp.deleteOnExit(); try (FileOutputStream output = new FileOutputStream(fp); ObjectOutputStream obOutput = new ObjectOutputStream(output);) { obOutput.writeObject(jobGraph); } LocalResource jobgraph = Records.newRecord(LocalResource.class); Path remoteJobGraph = Utils.setupLocalResource(fs, appId.toString(), new Path(fp.toURI()), jobgraph, fs.getHomeDirectory()); localResources.put("job.graph", jobgraph); paths.add(remoteJobGraph); classPathBuilder.append("job.graph").append(File.pathSeparator); } catch (Exception e) { LOG.warn("Add job graph to local resource fail"); throw e; } } sessionFilesDir = new Path(fs.getHomeDirectory(), ".flink/" + appId.toString() + "/"); FsPermission permission = new FsPermission(FsAction.ALL, FsAction.NONE, FsAction.NONE); fs.setPermission(sessionFilesDir, permission); // set permission for path. //To support Yarn Secure Integration Test Scenario //In Integration test setup, the Yarn containers created by YarnMiniCluster does not have the Yarn site XML //and KRB5 configuration files. We are adding these files as container local resources for the container //applications (JM/TMs) to have proper secure cluster setup Path remoteKrb5Path = null; Path remoteYarnSiteXmlPath = null; boolean hasKrb5 = false; if (System.getenv("IN_TESTS") != null) { String krb5Config = System.getProperty("java.security.krb5.conf"); if (krb5Config != null && krb5Config.length() != 0) { File krb5 = new File(krb5Config); LOG.info("Adding KRB5 configuration {} to the AM container local resource bucket", krb5.getAbsolutePath()); LocalResource krb5ConfResource = Records.newRecord(LocalResource.class); Path krb5ConfPath = new Path(krb5.getAbsolutePath()); remoteKrb5Path = Utils.setupLocalResource(fs, appId.toString(), krb5ConfPath, krb5ConfResource, fs.getHomeDirectory()); localResources.put(Utils.KRB5_FILE_NAME, krb5ConfResource); File f = new File(System.getenv("YARN_CONF_DIR"), Utils.YARN_SITE_FILE_NAME); LOG.info("Adding Yarn configuration {} to the AM container local resource bucket", f.getAbsolutePath()); LocalResource yarnConfResource = Records.newRecord(LocalResource.class); Path yarnSitePath = new Path(f.getAbsolutePath()); remoteYarnSiteXmlPath = Utils.setupLocalResource(fs, appId.toString(), yarnSitePath, yarnConfResource, fs.getHomeDirectory()); localResources.put(Utils.YARN_SITE_FILE_NAME, yarnConfResource); hasKrb5 = true; } } // setup security tokens LocalResource keytabResource = null; Path remotePathKeytab = null; String keytab = flinkConfiguration.getString(SecurityOptions.KERBEROS_LOGIN_KEYTAB); if (keytab != null) { LOG.info("Adding keytab {} to the AM container local resource bucket", keytab); keytabResource = Records.newRecord(LocalResource.class); Path keytabPath = new Path(keytab); remotePathKeytab = Utils.setupLocalResource(fs, appId.toString(), keytabPath, keytabResource, fs.getHomeDirectory()); localResources.put(Utils.KEYTAB_FILE_NAME, keytabResource); } final ContainerLaunchContext amContainer = setupApplicationMasterContainer(hasLogback, hasLog4j, hasKrb5); if (UserGroupInformation.isSecurityEnabled() && keytab == null) { //set tokens only when keytab is not provided LOG.info("Adding delegation token to the AM container.."); Utils.setTokensFor(amContainer, paths, conf); } amContainer.setLocalResources(localResources); fs.close(); // Setup CLASSPATH and environment variables for ApplicationMaster final Map<String, String> appMasterEnv = new HashMap<>(); // set user specified app master environment variables appMasterEnv.putAll(Utils.getEnvironmentVariables(ConfigConstants.YARN_APPLICATION_MASTER_ENV_PREFIX, flinkConfiguration)); // set Flink app class path appMasterEnv.put(YarnConfigKeys.ENV_FLINK_CLASSPATH, classPathBuilder.toString()); // set Flink on YARN internal configuration values appMasterEnv.put(YarnConfigKeys.ENV_TM_COUNT, String.valueOf(taskManagerCount)); appMasterEnv.put(YarnConfigKeys.ENV_TM_MEMORY, String.valueOf(taskManagerMemoryMb)); appMasterEnv.put(YarnConfigKeys.FLINK_JAR_PATH, remotePathJar.toString()); appMasterEnv.put(YarnConfigKeys.ENV_APP_ID, appId.toString()); appMasterEnv.put(YarnConfigKeys.ENV_CLIENT_HOME_DIR, fs.getHomeDirectory().toString()); appMasterEnv.put(YarnConfigKeys.ENV_CLIENT_SHIP_FILES, envShipFileList.toString()); appMasterEnv.put(YarnConfigKeys.ENV_SLOTS, String.valueOf(slots)); appMasterEnv.put(YarnConfigKeys.ENV_DETACHED, String.valueOf(detached)); appMasterEnv.put(YarnConfigKeys.ENV_ZOOKEEPER_NAMESPACE, getZookeeperNamespace()); // https://github.com/apache/hadoop/blob/trunk/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-site/src/site/markdown/YarnApplicationSecurity.md#identity-on-an-insecure-cluster-hadoop_user_name appMasterEnv.put(YarnConfigKeys.ENV_HADOOP_USER_NAME, UserGroupInformation.getCurrentUser().getUserName()); if (keytabResource != null) { appMasterEnv.put(YarnConfigKeys.KEYTAB_PATH, remotePathKeytab.toString()); String principal = flinkConfiguration.getString(SecurityOptions.KERBEROS_LOGIN_PRINCIPAL); appMasterEnv.put(YarnConfigKeys.KEYTAB_PRINCIPAL, principal); } //To support Yarn Secure Integration Test Scenario if (remoteYarnSiteXmlPath != null && remoteKrb5Path != null) { appMasterEnv.put(YarnConfigKeys.ENV_YARN_SITE_XML_PATH, remoteYarnSiteXmlPath.toString()); appMasterEnv.put(YarnConfigKeys.ENV_KRB5_PATH, remoteKrb5Path.toString()); } if (dynamicPropertiesEncoded != null) { appMasterEnv.put(YarnConfigKeys.ENV_DYNAMIC_PROPERTIES, dynamicPropertiesEncoded); } // set classpath from YARN configuration Utils.setupYarnClassPath(conf, appMasterEnv); amContainer.setEnvironment(appMasterEnv); // Set up resource type requirements for ApplicationMaster Resource capability = Records.newRecord(Resource.class); capability.setMemory(jobManagerMemoryMb); capability.setVirtualCores(1); String name; if (customName == null) { name = "Flink session with " + taskManagerCount + " TaskManagers"; if (detached) { name += " (detached)"; } } else { name = customName; } appContext.setApplicationName(name); appContext.setApplicationType("Apache Flink"); appContext.setAMContainerSpec(amContainer); appContext.setResource(capability); if (yarnQueue != null) { appContext.setQueue(yarnQueue); } setApplicationTags(appContext); // add a hook to clean up in case deployment fails Thread deploymentFailureHook = new DeploymentFailureHook(yarnClient, yarnApplication); Runtime.getRuntime().addShutdownHook(deploymentFailureHook); LOG.info("Submitting application master " + appId); yarnClient.submitApplication(appContext); LOG.info("Waiting for the cluster to be allocated"); final long startTime = System.currentTimeMillis(); ApplicationReport report; YarnApplicationState lastAppState = YarnApplicationState.NEW; loop: while (true) { try { report = yarnClient.getApplicationReport(appId); } catch (IOException e) { throw new YarnDeploymentException("Failed to deploy the cluster.", e); } YarnApplicationState appState = report.getYarnApplicationState(); LOG.debug("Application State: {}", appState); switch (appState) { case FAILED: case FINISHED: //TODO: the finished state may be valid in flip-6 case KILLED: throw new YarnDeploymentException("The YARN application unexpectedly switched to state " + appState + " during deployment. \n" + "Diagnostics from YARN: " + report.getDiagnostics() + "\n" + "If log aggregation is enabled on your cluster, use this command to further investigate the issue:\n" + "yarn logs -applicationId " + appId); //break .. case RUNNING: LOG.info("YARN application has been deployed successfully."); break loop; default: if (appState != lastAppState) { LOG.info("Deploying cluster, current state " + appState); } if (System.currentTimeMillis() - startTime > 60000) { LOG.info( "Deployment took more than 60 seconds. Please check if the requested resources are available in the YARN cluster"); } } lastAppState = appState; Thread.sleep(250); } // print the application id for user to cancel themselves. if (isDetachedMode()) { LOG.info("The Flink YARN client has been started in detached mode. In order to stop " + "Flink on YARN, use the following command or a YARN web interface to stop " + "it:\nyarn application -kill " + appId + "\nPlease also note that the " + "temporary files of the YARN session in the home directoy will not be removed."); } // since deployment was successful, remove the hook try { Runtime.getRuntime().removeShutdownHook(deploymentFailureHook); } catch (IllegalStateException e) { // we're already in the shut down hook. } return report; }
From source file:org.apache.flink.yarn.Client.java
License:Apache License
public void run(String[] args) throws Exception { if (UserGroupInformation.isSecurityEnabled()) { throw new RuntimeException("Flink YARN client does not have security support right now." + "File a bug, we will fix it asap"); }//from w w w. j av a 2 s. c om //Utils.logFilesInCurrentDirectory(LOG); // // Command Line Options // Options options = new Options(); options.addOption(VERBOSE); options.addOption(FLINK_CONF_DIR); options.addOption(FLINK_JAR); options.addOption(JM_MEMORY); options.addOption(TM_MEMORY); options.addOption(TM_CORES); options.addOption(CONTAINER); options.addOption(GEN_CONF); options.addOption(QUEUE); options.addOption(QUERY); options.addOption(SHIP_PATH); CommandLineParser parser = new PosixParser(); CommandLine cmd = null; try { cmd = parser.parse(options, args); } catch (MissingOptionException moe) { System.out.println(moe.getMessage()); printUsage(); System.exit(1); } if (System.getProperty("log4j.configuration") == null) { Logger root = Logger.getRootLogger(); root.removeAllAppenders(); PatternLayout layout = new PatternLayout("%d{HH:mm:ss,SSS} %-5p %-60c %x - %m%n"); ConsoleAppender appender = new ConsoleAppender(layout, "System.err"); root.addAppender(appender); if (cmd.hasOption(VERBOSE.getOpt())) { root.setLevel(Level.DEBUG); LOG.debug("CLASSPATH: " + System.getProperty("java.class.path")); } else { root.setLevel(Level.INFO); } } // Jar Path Path localJarPath; if (cmd.hasOption(FLINK_JAR.getOpt())) { String userPath = cmd.getOptionValue(FLINK_JAR.getOpt()); if (!userPath.startsWith("file://")) { userPath = "file://" + userPath; } localJarPath = new Path(userPath); } else { localJarPath = new Path( "file://" + Client.class.getProtectionDomain().getCodeSource().getLocation().getPath()); } if (cmd.hasOption(GEN_CONF.getOpt())) { LOG.info("Placing default configuration in current directory"); File outFile = generateDefaultConf(localJarPath); LOG.info("File written to " + outFile.getAbsolutePath()); System.exit(0); } // Conf Path Path confPath = null; String confDirPath = ""; if (cmd.hasOption(FLINK_CONF_DIR.getOpt())) { confDirPath = cmd.getOptionValue(FLINK_CONF_DIR.getOpt()) + "/"; File confFile = new File(confDirPath + CONFIG_FILE_NAME); if (!confFile.exists()) { LOG.fatal("Unable to locate configuration file in " + confFile); System.exit(1); } confPath = new Path(confFile.getAbsolutePath()); } else { System.out.println("No configuration file has been specified"); // no configuration path given. // -> see if there is one in the current directory File currDir = new File("."); File[] candidates = currDir.listFiles(new FilenameFilter() { @Override public boolean accept(final File dir, final String name) { return name != null && name.endsWith(".yaml"); } }); if (candidates == null || candidates.length == 0) { System.out.println( "No configuration file has been found in current directory.\n" + "Copying default."); File outFile = generateDefaultConf(localJarPath); confPath = new Path(outFile.toURI()); } else { if (candidates.length > 1) { System.out.println("Multiple .yaml configuration files were found in the current directory\n" + "Please specify one explicitly"); System.exit(1); } else if (candidates.length == 1) { confPath = new Path(candidates[0].toURI()); } } } List<File> shipFiles = new ArrayList<File>(); // path to directory to ship if (cmd.hasOption(SHIP_PATH.getOpt())) { String shipPath = cmd.getOptionValue(SHIP_PATH.getOpt()); File shipDir = new File(shipPath); if (shipDir.isDirectory()) { shipFiles = new ArrayList<File>(Arrays.asList(shipDir.listFiles(new FilenameFilter() { @Override public boolean accept(File dir, String name) { return !(name.equals(".") || name.equals("..")); } }))); } else { LOG.warn("Ship directory is not a directory!"); } } boolean hasLog4j = false; //check if there is a log4j file if (confDirPath.length() > 0) { File l4j = new File(confDirPath + "/log4j.properties"); if (l4j.exists()) { shipFiles.add(l4j); hasLog4j = true; } } // queue String queue = "default"; if (cmd.hasOption(QUEUE.getOpt())) { queue = cmd.getOptionValue(QUEUE.getOpt()); } // JobManager Memory int jmMemory = 512; if (cmd.hasOption(JM_MEMORY.getOpt())) { jmMemory = Integer.valueOf(cmd.getOptionValue(JM_MEMORY.getOpt())); } // Task Managers memory int tmMemory = 1024; if (cmd.hasOption(TM_MEMORY.getOpt())) { tmMemory = Integer.valueOf(cmd.getOptionValue(TM_MEMORY.getOpt())); } // Task Managers vcores int tmCores = 1; if (cmd.hasOption(TM_CORES.getOpt())) { tmCores = Integer.valueOf(cmd.getOptionValue(TM_CORES.getOpt())); } Utils.getFlinkConfiguration(confPath.toUri().getPath()); int jmPort = GlobalConfiguration.getInteger(ConfigConstants.JOB_MANAGER_IPC_PORT_KEY, 0); if (jmPort == 0) { LOG.warn("Unable to find job manager port in configuration!"); jmPort = ConfigConstants.DEFAULT_JOB_MANAGER_IPC_PORT; } conf = Utils.initializeYarnConfiguration(); // intialize HDFS LOG.info("Copy App Master jar from local filesystem and add to local environment"); // Copy the application master jar to the filesystem // Create a local resource to point to the destination jar path final FileSystem fs = FileSystem.get(conf); if (fs.getScheme().startsWith("file")) { LOG.warn("The file system scheme is '" + fs.getScheme() + "'. This indicates that the " + "specified Hadoop configuration path is wrong and the sytem is using the default Hadoop configuration values." + "The Flink YARN client needs to store its files in a distributed file system"); } // Create yarnClient final YarnClient yarnClient = YarnClient.createYarnClient(); yarnClient.init(conf); yarnClient.start(); // Query cluster for metrics if (cmd.hasOption(QUERY.getOpt())) { showClusterMetrics(yarnClient); } if (!cmd.hasOption(CONTAINER.getOpt())) { LOG.fatal("Missing required argument " + CONTAINER.getOpt()); printUsage(); yarnClient.stop(); System.exit(1); } // TM Count final int taskManagerCount = Integer.valueOf(cmd.getOptionValue(CONTAINER.getOpt())); System.out.println("Using values:"); System.out.println("\tContainer Count = " + taskManagerCount); System.out.println("\tJar Path = " + localJarPath.toUri().getPath()); System.out.println("\tConfiguration file = " + confPath.toUri().getPath()); System.out.println("\tJobManager memory = " + jmMemory); System.out.println("\tTaskManager memory = " + tmMemory); System.out.println("\tTaskManager cores = " + tmCores); // Create application via yarnClient YarnClientApplication app = yarnClient.createApplication(); GetNewApplicationResponse appResponse = app.getNewApplicationResponse(); Resource maxRes = appResponse.getMaximumResourceCapability(); if (tmMemory > maxRes.getMemory() || tmCores > maxRes.getVirtualCores()) { LOG.fatal("The cluster does not have the requested resources for the TaskManagers available!\n" + "Maximum Memory: " + maxRes.getMemory() + ", Maximum Cores: " + tmCores); yarnClient.stop(); System.exit(1); } if (jmMemory > maxRes.getMemory()) { LOG.fatal("The cluster does not have the requested resources for the JobManager available!\n" + "Maximum Memory: " + maxRes.getMemory()); yarnClient.stop(); System.exit(1); } int totalMemoryRequired = jmMemory + tmMemory * taskManagerCount; ClusterResourceDescription freeClusterMem = getCurrentFreeClusterResources(yarnClient); if (freeClusterMem.totalFreeMemory < totalMemoryRequired) { LOG.fatal("This YARN session requires " + totalMemoryRequired + "MB of memory in the cluster. " + "There are currently only " + freeClusterMem.totalFreeMemory + "MB available."); yarnClient.stop(); System.exit(1); } if (tmMemory > freeClusterMem.containerLimit) { LOG.fatal("The requested amount of memory for the TaskManagers (" + tmMemory + "MB) is more than " + "the largest possible YARN container: " + freeClusterMem.containerLimit); yarnClient.stop(); System.exit(1); } if (jmMemory > freeClusterMem.containerLimit) { LOG.fatal("The requested amount of memory for the JobManager (" + jmMemory + "MB) is more than " + "the largest possible YARN container: " + freeClusterMem.containerLimit); yarnClient.stop(); System.exit(1); } // respect custom JVM options in the YAML file final String javaOpts = GlobalConfiguration.getString(ConfigConstants.FLINK_JVM_OPTIONS, ""); // Set up the container launch context for the application master ContainerLaunchContext amContainer = Records.newRecord(ContainerLaunchContext.class); String amCommand = "$JAVA_HOME/bin/java" + " -Xmx" + Utils.calculateHeapSize(jmMemory) + "M " + javaOpts; if (hasLog4j) { amCommand += " -Dlog.file=\"" + ApplicationConstants.LOG_DIR_EXPANSION_VAR + "/jobmanager-log4j.log\" -Dlog4j.configuration=file:log4j.properties"; } amCommand += " org.apache.flink.yarn.ApplicationMaster" + " " + " 1>" + ApplicationConstants.LOG_DIR_EXPANSION_VAR + "/jobmanager-stdout.log" + " 2>" + ApplicationConstants.LOG_DIR_EXPANSION_VAR + "/jobmanager-stderr.log"; amContainer.setCommands(Collections.singletonList(amCommand)); System.err.println("amCommand=" + amCommand); // Set-up ApplicationSubmissionContext for the application ApplicationSubmissionContext appContext = app.getApplicationSubmissionContext(); final ApplicationId appId = appContext.getApplicationId(); // Setup jar for ApplicationMaster LocalResource appMasterJar = Records.newRecord(LocalResource.class); LocalResource flinkConf = Records.newRecord(LocalResource.class); Path remotePathJar = Utils.setupLocalResource(conf, fs, appId.toString(), localJarPath, appMasterJar, fs.getHomeDirectory()); Path remotePathConf = Utils.setupLocalResource(conf, fs, appId.toString(), confPath, flinkConf, fs.getHomeDirectory()); Map<String, LocalResource> localResources = new HashMap<String, LocalResource>(2); localResources.put("flink.jar", appMasterJar); localResources.put("flink-conf.yaml", flinkConf); // setup security tokens (code from apache storm) final Path[] paths = new Path[3 + shipFiles.size()]; StringBuffer envShipFileList = new StringBuffer(); // upload ship files for (int i = 0; i < shipFiles.size(); i++) { File shipFile = shipFiles.get(i); LocalResource shipResources = Records.newRecord(LocalResource.class); Path shipLocalPath = new Path("file://" + shipFile.getAbsolutePath()); paths[3 + i] = Utils.setupLocalResource(conf, fs, appId.toString(), shipLocalPath, shipResources, fs.getHomeDirectory()); localResources.put(shipFile.getName(), shipResources); envShipFileList.append(paths[3 + i]); if (i + 1 < shipFiles.size()) { envShipFileList.append(','); } } paths[0] = remotePathJar; paths[1] = remotePathConf; paths[2] = new Path(fs.getHomeDirectory(), ".flink/" + appId.toString() + "/"); FsPermission permission = new FsPermission(FsAction.ALL, FsAction.ALL, FsAction.ALL); fs.setPermission(paths[2], permission); // set permission for path. Utils.setTokensFor(amContainer, paths, this.conf); amContainer.setLocalResources(localResources); fs.close(); // Setup CLASSPATH for ApplicationMaster Map<String, String> appMasterEnv = new HashMap<String, String>(); Utils.setupEnv(conf, appMasterEnv); // set configuration values appMasterEnv.put(Client.ENV_TM_COUNT, String.valueOf(taskManagerCount)); appMasterEnv.put(Client.ENV_TM_CORES, String.valueOf(tmCores)); appMasterEnv.put(Client.ENV_TM_MEMORY, String.valueOf(tmMemory)); appMasterEnv.put(Client.FLINK_JAR_PATH, remotePathJar.toString()); appMasterEnv.put(Client.ENV_APP_ID, appId.toString()); appMasterEnv.put(Client.ENV_CLIENT_HOME_DIR, fs.getHomeDirectory().toString()); appMasterEnv.put(Client.ENV_CLIENT_SHIP_FILES, envShipFileList.toString()); appMasterEnv.put(Client.ENV_CLIENT_USERNAME, UserGroupInformation.getCurrentUser().getShortUserName()); amContainer.setEnvironment(appMasterEnv); // Set up resource type requirements for ApplicationMaster Resource capability = Records.newRecord(Resource.class); capability.setMemory(jmMemory); capability.setVirtualCores(1); appContext.setApplicationName("Flink"); // application name appContext.setAMContainerSpec(amContainer); appContext.setResource(capability); appContext.setQueue(queue); // file that we write into the conf/ dir containing the jobManager address. final File addrFile = new File(confDirPath + CliFrontend.JOBMANAGER_ADDRESS_FILE); Runtime.getRuntime().addShutdownHook(new Thread() { @Override public void run() { try { LOG.info("Killing the Flink-YARN application."); yarnClient.killApplication(appId); LOG.info("Deleting files in " + paths[2]); FileSystem shutFS = FileSystem.get(conf); shutFS.delete(paths[2], true); // delete conf and jar file. shutFS.close(); } catch (Exception e) { LOG.warn("Exception while killing the YARN application", e); } try { addrFile.delete(); } catch (Exception e) { LOG.warn("Exception while deleting the jobmanager address file", e); } LOG.info("YARN Client is shutting down"); yarnClient.stop(); } }); LOG.info("Submitting application master " + appId); yarnClient.submitApplication(appContext); ApplicationReport appReport = yarnClient.getApplicationReport(appId); YarnApplicationState appState = appReport.getYarnApplicationState(); boolean told = false; char[] el = { '/', '|', '\\', '-' }; int i = 0; while (appState != YarnApplicationState.FINISHED && appState != YarnApplicationState.KILLED && appState != YarnApplicationState.FAILED) { if (!told && appState == YarnApplicationState.RUNNING) { System.err.println("Flink JobManager is now running on " + appReport.getHost() + ":" + jmPort); System.err.println("JobManager Web Interface: " + appReport.getTrackingUrl()); // write jobmanager connect information PrintWriter out = new PrintWriter(addrFile); out.println(appReport.getHost() + ":" + jmPort); out.close(); addrFile.setReadable(true, false); // readable for all. told = true; } if (!told) { System.err.print(el[i++] + "\r"); if (i == el.length) { i = 0; } Thread.sleep(500); // wait for the application to switch to RUNNING } else { Thread.sleep(5000); } appReport = yarnClient.getApplicationReport(appId); appState = appReport.getYarnApplicationState(); } LOG.info("Application " + appId + " finished with" + " state " + appState + " at " + appReport.getFinishTime()); if (appState == YarnApplicationState.FAILED || appState == YarnApplicationState.KILLED) { LOG.warn("Application failed. Diagnostics " + appReport.getDiagnostics()); } }
From source file:org.apache.ignite.yarn.IgniteYarnClient.java
License:Apache License
/** * Main methods has one mandatory parameter and one optional parameter. * * @param args Path to jar mandatory parameter and property file is optional. *//*from w w w . j a va 2 s .c o m*/ public static void main(String[] args) throws Exception { checkArguments(args); // Set path to app master jar. String pathAppMasterJar = args[0]; ClusterProperties props = ClusterProperties.from(args.length == 2 ? args[1] : null); YarnConfiguration conf = new YarnConfiguration(); YarnClient yarnClient = YarnClient.createYarnClient(); yarnClient.init(conf); yarnClient.start(); // Create application via yarnClient YarnClientApplication app = yarnClient.createApplication(); FileSystem fs = FileSystem.get(conf); Path ignite; // Load ignite and jar if (props.ignitePath() == null) ignite = getIgnite(props, fs); else ignite = new Path(props.ignitePath()); // Upload the jar file to HDFS. Path appJar = IgniteYarnUtils.copyLocalToHdfs(fs, pathAppMasterJar, props.igniteWorkDir() + File.separator + IgniteYarnUtils.JAR_NAME); // Set up the container launch context for the application master ContainerLaunchContext amContainer = Records.newRecord(ContainerLaunchContext.class); amContainer.setCommands(Collections .singletonList(Environment.JAVA_HOME.$() + "/bin/java -Xmx512m " + ApplicationMaster.class.getName() + IgniteYarnUtils.SPACE + ignite.toUri() + IgniteYarnUtils.YARN_LOG_OUT)); // Setup jar for ApplicationMaster LocalResource appMasterJar = IgniteYarnUtils.setupFile(appJar, fs, LocalResourceType.FILE); amContainer.setLocalResources(Collections.singletonMap(IgniteYarnUtils.JAR_NAME, appMasterJar)); // Setup CLASSPATH for ApplicationMaster Map<String, String> appMasterEnv = props.toEnvs(); setupAppMasterEnv(appMasterEnv, conf); amContainer.setEnvironment(appMasterEnv); // Setup security tokens if (UserGroupInformation.isSecurityEnabled()) { Credentials creds = new Credentials(); String tokRenewer = conf.get(YarnConfiguration.RM_PRINCIPAL); if (tokRenewer == null || tokRenewer.length() == 0) throw new IOException("Master Kerberos principal for the RM is not set."); log.info("Found RM principal: " + tokRenewer); final Token<?> tokens[] = fs.addDelegationTokens(tokRenewer, creds); if (tokens != null) log.info("File system delegation tokens: " + Arrays.toString(tokens)); amContainer.setTokens(IgniteYarnUtils.createTokenBuffer(creds)); } // Set up resource type requirements for ApplicationMaster Resource capability = Records.newRecord(Resource.class); capability.setMemory(512); capability.setVirtualCores(1); // Finally, set-up ApplicationSubmissionContext for the application ApplicationSubmissionContext appContext = app.getApplicationSubmissionContext(); appContext.setApplicationName("ignition"); // application name appContext.setAMContainerSpec(amContainer); appContext.setResource(capability); appContext.setQueue("default"); // queue // Submit application ApplicationId appId = appContext.getApplicationId(); yarnClient.submitApplication(appContext); log.log(Level.INFO, "Submitted application. Application id: {0}", appId); ApplicationReport appReport = yarnClient.getApplicationReport(appId); YarnApplicationState appState = appReport.getYarnApplicationState(); while (appState == YarnApplicationState.NEW || appState == YarnApplicationState.NEW_SAVING || appState == YarnApplicationState.SUBMITTED || appState == YarnApplicationState.ACCEPTED) { TimeUnit.SECONDS.sleep(1L); appReport = yarnClient.getApplicationReport(appId); if (appState != YarnApplicationState.ACCEPTED && appReport.getYarnApplicationState() == YarnApplicationState.ACCEPTED) log.log(Level.INFO, "Application {0} is ACCEPTED.", appId); appState = appReport.getYarnApplicationState(); } log.log(Level.INFO, "Application {0} is {1}.", new Object[] { appId, appState }); }
From source file:org.apache.reef.runtime.yarn.driver.unmanaged.UnmanagedAmTest.java
License:Apache License
@Test public void testAmShutdown() throws IOException, YarnException { Assume.assumeTrue("This test requires a YARN Resource Manager to connect to", Boolean.parseBoolean(System.getenv("REEF_TEST_YARN"))); final YarnConfiguration yarnConfig = new YarnConfiguration(); // Start YARN client and register the application final YarnClient yarnClient = YarnClient.createYarnClient(); yarnClient.init(yarnConfig);//w ww.j a v a 2 s . co m yarnClient.start(); final ContainerLaunchContext containerContext = Records.newRecord(ContainerLaunchContext.class); containerContext.setCommands(Collections.<String>emptyList()); containerContext.setLocalResources(Collections.<String, LocalResource>emptyMap()); containerContext.setEnvironment(Collections.<String, String>emptyMap()); containerContext.setTokens(getTokens()); final ApplicationSubmissionContext appContext = yarnClient.createApplication() .getApplicationSubmissionContext(); appContext.setApplicationName("REEF_Unmanaged_AM_Test"); appContext.setAMContainerSpec(containerContext); appContext.setUnmanagedAM(true); appContext.setQueue("default"); final ApplicationId applicationId = appContext.getApplicationId(); LOG.log(Level.INFO, "Registered YARN application: {0}", applicationId); yarnClient.submitApplication(appContext); LOG.log(Level.INFO, "YARN application submitted: {0}", applicationId); addToken(yarnClient.getAMRMToken(applicationId)); // Start the AM final AMRMClientAsync<AMRMClient.ContainerRequest> rmClient = AMRMClientAsync.createAMRMClientAsync(1000, this); rmClient.init(yarnConfig); rmClient.start(); final NMClientAsync nmClient = new NMClientAsyncImpl(this); nmClient.init(yarnConfig); nmClient.start(); final RegisterApplicationMasterResponse registration = rmClient .registerApplicationMaster(NetUtils.getHostname(), -1, null); LOG.log(Level.INFO, "Unmanaged AM is running: {0}", registration); rmClient.unregisterApplicationMaster(FinalApplicationStatus.SUCCEEDED, "Success!", null); LOG.log(Level.INFO, "Unregistering AM: state {0}", rmClient.getServiceState()); // Shutdown the AM rmClient.stop(); nmClient.stop(); // Get the final application report final ApplicationReport appReport = yarnClient.getApplicationReport(applicationId); final YarnApplicationState appState = appReport.getYarnApplicationState(); final FinalApplicationStatus finalAttemptStatus = appReport.getFinalApplicationStatus(); LOG.log(Level.INFO, "Application {0} final attempt {1} status: {2}/{3}", new Object[] { applicationId, appReport.getCurrentApplicationAttemptId(), appState, finalAttemptStatus }); Assert.assertEquals("Application must be in FINISHED state", YarnApplicationState.FINISHED, appState); Assert.assertEquals("Final status must be SUCCEEDED", FinalApplicationStatus.SUCCEEDED, finalAttemptStatus); // Shutdown YARN client yarnClient.stop(); }
From source file:proxyyarn.ProxyYarn.java
License:Apache License
public boolean run() throws Exception { Configuration conf = new YarnConfiguration(new Configuration()); YarnClient yarnClient = YarnClient.createYarnClient(); yarnClient.init(conf);/*ww w.j a va 2 s . c o m*/ yarnClient.start(); YarnClusterMetrics clusterMetrics = yarnClient.getYarnClusterMetrics(); log.info("Got Cluster metric info from ASM" + ", numNodeManagers=" + clusterMetrics.getNumNodeManagers()); List<NodeReport> clusterNodeReports = yarnClient.getNodeReports(NodeState.RUNNING); log.info("Got Cluster node info from ASM"); for (NodeReport node : clusterNodeReports) { log.info("Got node report from ASM for" + ", nodeId=" + node.getNodeId() + ", nodeAddress" + node.getHttpAddress() + ", nodeRackName" + node.getRackName() + ", nodeNumContainers" + node.getNumContainers()); } QueueInfo queueInfo = yarnClient.getQueueInfo("default"); log.info("Queue info" + ", queueName=" + queueInfo.getQueueName() + ", queueCurrentCapacity=" + queueInfo.getCurrentCapacity() + ", queueMaxCapacity=" + queueInfo.getMaximumCapacity() + ", queueApplicationCount=" + queueInfo.getApplications().size() + ", queueChildQueueCount=" + queueInfo.getChildQueues().size()); List<QueueUserACLInfo> listAclInfo = yarnClient.getQueueAclsInfo(); for (QueueUserACLInfo aclInfo : listAclInfo) { for (QueueACL userAcl : aclInfo.getUserAcls()) { log.info("User ACL Info for Queue" + ", queueName=" + aclInfo.getQueueName() + ", userAcl=" + userAcl.name()); } } FileSystem fs = FileSystem.get(conf); if (!fs.getClass().equals(DistributedFileSystem.class)) { log.error("Expected DistributedFileSystem, but was {}", fs.getClass().getSimpleName()); System.exit(1); } // ApplicationClientProtocol applicationsManager; // InetSocketAddress rmAddress = NetUtils.createSocketAddr(yarnConf.get(YarnConfiguration.RM_ADDRESS, YarnConfiguration.DEFAULT_RM_ADDRESS)); // log.info("Connecting to ResourceManager at {}", rmAddress); // Configuration appManagerServerConf = new Configuration(conf); // YarnRPC rpc = YarnRPC.create(appManagerServerConf); // ApplicationClientProtocol applicationManager = (ApplicationClientProtocol) rpc.getProxy(ApplicationClientProtocol.class, rmAddress, appManagerServerConf); String appName = "AccumuloProxyYarn"; YarnClientApplication app = yarnClient.createApplication(); // set the application name ApplicationSubmissionContext appContext = app.getApplicationSubmissionContext(); ApplicationId appId = appContext.getApplicationId(); appContext.setApplicationName(appName); // GetNewApplicationRequest request = Records.newRecord(GetNewApplicationRequest.class); // GetNewApplicationResponse response = applicationManager.getNewApplication(request); // log.info("Got new ApplicationId=" + response.getApplicationId()); // ApplicationId appId = response.getApplicationId(); // Create a new ApplicationSubmissionContext // ApplicationSubmissionContext appContext = Records.newRecord(ApplicationSubmissionContext.class); // set the ApplicationId // appContext.setApplicationId(appId); // set the application name // appContext.setApplicationName(appName); // Create a new container launch context for the AM's container ContainerLaunchContext amContainer = Records.newRecord(ContainerLaunchContext.class); // Define the local resources required Map<String, LocalResource> localResources = new HashMap<String, LocalResource>(); // Lets assume the jar we need for our ApplicationMaster is available in // HDFS at a certain known path to us and we want to make it available to // the ApplicationMaster in the launched container Path localJarPath = new Path( "file:///Users/jelser/projects/accumulo-proxy-yarn/target/accumulo-proxy-yarn-0.0.1-SNAPSHOT.jar"); Path jarPath = new Path("hdfs:///accumulo-proxy-yarn-0.0.1-SNAPSHOT.jar"); fs.copyFromLocalFile(false, true, localJarPath, jarPath); FileStatus jarStatus = fs.getFileStatus(jarPath); LocalResource amJarRsrc = Records.newRecord(LocalResource.class); // Set the type of resource - file or archive // archives are untarred at the destination by the framework amJarRsrc.setType(LocalResourceType.FILE); // Set visibility of the resource // Setting to most private option i.e. this file will only // be visible to this instance of the running application amJarRsrc.setVisibility(LocalResourceVisibility.APPLICATION); // Set the location of resource to be copied over into the // working directory amJarRsrc.setResource(ConverterUtils.getYarnUrlFromPath(jarPath)); // Set timestamp and length of file so that the framework // can do basic sanity checks for the local resource // after it has been copied over to ensure it is the same // resource the client intended to use with the application amJarRsrc.setTimestamp(jarStatus.getModificationTime()); amJarRsrc.setSize(jarStatus.getLen()); // The framework will create a symlink called AppMaster.jar in the // working directory that will be linked back to the actual file. // The ApplicationMaster, if needs to reference the jar file, would // need to use the symlink filename. localResources.put("AppMaster.jar", amJarRsrc); // Set the local resources into the launch context amContainer.setLocalResources(localResources); // Set up the environment needed for the launch context Map<String, String> env = new HashMap<String, String>(); // For example, we could setup the classpath needed. // Assuming our classes or jars are available as local resources in the // working directory from which the command will be run, we need to append // "." to the path. // By default, all the hadoop specific classpaths will already be available // in $CLASSPATH, so we should be careful not to overwrite it. String classPathEnv = "$CLASSPATH:./*:/Users/jelser/projects/accumulo-proxy-yarn/target/lib/*"; env.put("CLASSPATH", classPathEnv); amContainer.setEnvironment(env); // Construct the command to be executed on the launched container String command = "${JAVA_HOME}" + "/bin/java" + " proxyyarn.ProxyYarnAppMaster 1>" + ApplicationConstants.LOG_DIR_EXPANSION_VAR + "/stdout" + " 2>" + ApplicationConstants.LOG_DIR_EXPANSION_VAR + "/stderr"; List<String> commands = new ArrayList<String>(); commands.add(command); // add additional commands if needed // Set the command array into the container spec amContainer.setCommands(commands); // Define the resource requirements for the container // For now, YARN only supports memory so we set the memory // requirements. // If the process takes more than its allocated memory, it will // be killed by the framework. // Memory being requested for should be less than max capability // of the cluster and all asks should be a multiple of the min capability. Resource capability = Records.newRecord(Resource.class); capability.setMemory(256); appContext.setResource(capability); // Create the request to send to the ApplicationsManager // SubmitApplicationRequest appRequest = Records.newRecord(SubmitApplicationRequest.class); // appRequest.setApplicationSubmissionContext(appContext); // Submit the application to the ApplicationsManager // Ignore the response as either a valid response object is returned on // success or an exception thrown to denote the failure // applicationManager.submitApplication(appRequest); // Set the container launch content into the ApplicationSubmissionContext appContext.setAMContainerSpec(amContainer); // Set the priority for the application master Priority pri = Records.newRecord(Priority.class); // TODO - what is the range for priority? how to decide? pri.setPriority(0); appContext.setPriority(pri); // Set the queue to which this application is to be submitted in the RM appContext.setQueue("default"); // Submit the application to the applications manager // SubmitApplicationResponse submitResp = applicationsManager.submitApplication(appRequest); // Ignore the response as either a valid response object is returned on success // or an exception thrown to denote some form of a failure log.info("Submitting application to ASM"); yarnClient.submitApplication(appContext); return monitorApplication(yarnClient, appId); /* Thread.sleep(200); boolean running = false; while(true) { GetApplicationReportRequest reportRequest = Records.newRecord(GetApplicationReportRequest.class); reportRequest.setApplicationId(appId); GetApplicationReportResponse reportResponse = applicationManager.getApplicationReport(reportRequest); ApplicationReport report = reportResponse.getApplicationReport(); log.info(report.toString()); YarnApplicationState state = report.getYarnApplicationState(); switch (state) { case NEW: case NEW_SAVING: case SUBMITTED: case ACCEPTED: log.info("State: {}", state); break; case RUNNING: log.info("Running application"); running = true; break; case FINISHED: case FAILED: case KILLED: log.info("State: {}", state); return; default: log.info("Unknown state: {}", state); return; } if (!running) { Thread.sleep(1000); } }*/ }
From source file:runtime.starter.MPJYarnClient.java
License:Open Source License
public void run() throws Exception { Map<String, String> map = System.getenv(); try {/*from w w w . ja va2 s .com*/ mpjHomeDir = map.get("MPJ_HOME"); if (mpjHomeDir == null) { throw new Exception("[MPJRun.java]:MPJ_HOME environment found.."); } } catch (Exception exc) { System.out.println("[MPJRun.java]:" + exc.getMessage()); exc.printStackTrace(); return; } // Copy the application master jar to HDFS // Create a local resource to point to the destination jar path FileSystem fs = FileSystem.get(conf); /* Path dataset = new Path(fs.getHomeDirectory(),"/dataset"); FileStatus datasetFile = fs.getFileStatus(dataset); BlockLocation myBlocks [] = fs.getFileBlockLocations(datasetFile,0,datasetFile.getLen()); for(BlockLocation b : myBlocks){ System.out.println("\n--------------------"); System.out.println("Length "+b.getLength()); for(String host : b.getHosts()){ System.out.println("host "+host); } } */ Path source = new Path(mpjHomeDir + "/lib/mpj-app-master.jar"); String pathSuffix = hdfsFolder + "mpj-app-master.jar"; Path dest = new Path(fs.getHomeDirectory(), pathSuffix); if (debugYarn) { logger.info("Uploading mpj-app-master.jar to: " + dest.toString()); } fs.copyFromLocalFile(false, true, source, dest); FileStatus destStatus = fs.getFileStatus(dest); Path wrapperSource = new Path(mpjHomeDir + "/lib/mpj-yarn-wrapper.jar"); String wrapperSuffix = hdfsFolder + "mpj-yarn-wrapper.jar"; Path wrapperDest = new Path(fs.getHomeDirectory(), wrapperSuffix); if (debugYarn) { logger.info("Uploading mpj-yarn-wrapper.jar to: " + wrapperDest.toString()); } fs.copyFromLocalFile(false, true, wrapperSource, wrapperDest); Path userJar = new Path(jarPath); String userJarSuffix = hdfsFolder + "user-code.jar"; Path userJarDest = new Path(fs.getHomeDirectory(), userJarSuffix); if (debugYarn) { logger.info("Uploading user-code.jar to: " + userJarDest.toString()); } fs.copyFromLocalFile(false, true, userJar, userJarDest); YarnConfiguration conf = new YarnConfiguration(); YarnClient yarnClient = YarnClient.createYarnClient(); yarnClient.init(conf); yarnClient.start(); if (debugYarn) { YarnClusterMetrics metrics = yarnClient.getYarnClusterMetrics(); logger.info("\nNodes Information"); logger.info("Number of NM: " + metrics.getNumNodeManagers() + "\n"); List<NodeReport> nodeReports = yarnClient.getNodeReports(NodeState.RUNNING); for (NodeReport n : nodeReports) { logger.info("NodeId: " + n.getNodeId()); logger.info("RackName: " + n.getRackName()); logger.info("Total Memory: " + n.getCapability().getMemory()); logger.info("Used Memory: " + n.getUsed().getMemory()); logger.info("Total vCores: " + n.getCapability().getVirtualCores()); logger.info("Used vCores: " + n.getUsed().getVirtualCores() + "\n"); } } logger.info("Creating server socket at HOST " + serverName + " PORT " + serverPort + " \nWaiting for " + np + " processes to connect..."); // Creating a server socket for incoming connections try { servSock = new ServerSocket(serverPort); infoSock = new ServerSocket(); TEMP_PORT = findPort(infoSock); } catch (Exception e) { e.printStackTrace(); } // Create application via yarnClient YarnClientApplication app = yarnClient.createApplication(); GetNewApplicationResponse appResponse = app.getNewApplicationResponse(); int maxMem = appResponse.getMaximumResourceCapability().getMemory(); if (debugYarn) { logger.info("Max memory capability resources in cluster: " + maxMem); } if (amMem > maxMem) { amMem = maxMem; logger.info("AM memory specified above threshold of cluster " + "Using maximum memory for AM container: " + amMem); } int maxVcores = appResponse.getMaximumResourceCapability().getVirtualCores(); if (debugYarn) { logger.info("Max vCores capability resources in cluster: " + maxVcores); } if (amCores > maxVcores) { amCores = maxVcores; logger.info("AM virtual cores specified above threshold of cluster " + "Using maximum virtual cores for AM container: " + amCores); } // Set up the container launch context for the application master ContainerLaunchContext amContainer = Records.newRecord(ContainerLaunchContext.class); List<String> commands = new ArrayList<String>(); commands.add("$JAVA_HOME/bin/java"); commands.add("-Xmx" + amMem + "m"); commands.add("runtime.starter.MPJAppMaster"); commands.add("--np"); commands.add(String.valueOf(np)); commands.add("--serverName"); commands.add(serverName); //server name commands.add("--ioServerPort"); commands.add(Integer.toString(serverPort)); //server port commands.add("--deviceName"); commands.add(deviceName); //device name commands.add("--className"); commands.add(className); //class name commands.add("--wdir"); commands.add(workingDirectory); //wdir commands.add("--psl"); commands.add(Integer.toString(psl)); //protocol switch limit commands.add("--wireUpPort"); commands.add(String.valueOf(TEMP_PORT)); //for sharing ports & rank commands.add("--wrapperPath"); commands.add(wrapperDest.toString());//MPJYarnWrapper.jar HDFS path commands.add("--userJarPath"); commands.add(userJarDest.toString());//User Jar File HDFS path commands.add("--mpjContainerPriority"); commands.add(mpjContainerPriority);// priority for mpj containers commands.add("--containerMem"); commands.add(containerMem); commands.add("--containerCores"); commands.add(containerCores); if (debugYarn) { commands.add("--debugYarn"); } if (appArgs != null) { commands.add("--appArgs"); for (int i = 0; i < appArgs.length; i++) { commands.add(appArgs[i]); } } amContainer.setCommands(commands); //set commands // Setup local Resource for ApplicationMaster LocalResource appMasterJar = Records.newRecord(LocalResource.class); appMasterJar.setResource(ConverterUtils.getYarnUrlFromPath(dest)); appMasterJar.setSize(destStatus.getLen()); appMasterJar.setTimestamp(destStatus.getModificationTime()); appMasterJar.setType(LocalResourceType.ARCHIVE); appMasterJar.setVisibility(LocalResourceVisibility.APPLICATION); amContainer.setLocalResources(Collections.singletonMap("mpj-app-master.jar", appMasterJar)); // Setup CLASSPATH for ApplicationMaster // Setting up the environment Map<String, String> appMasterEnv = new HashMap<String, String>(); setupAppMasterEnv(appMasterEnv); amContainer.setEnvironment(appMasterEnv); // Set up resource type requirements for ApplicationMaster Resource capability = Records.newRecord(Resource.class); capability.setMemory(amMem); capability.setVirtualCores(amCores); // Finally, set-up ApplicationSubmissionContext for the application ApplicationSubmissionContext appContext = app.getApplicationSubmissionContext(); appContext.setApplicationName(appName); appContext.setAMContainerSpec(amContainer); appContext.setResource(capability); appContext.setQueue(yarnQueue); // queue Priority priority = Priority.newInstance(amPriority); appContext.setPriority(priority); ApplicationId appId = appContext.getApplicationId(); //Adding ShutDown Hook Runtime.getRuntime().addShutdownHook(new KillYarnApp(appId, yarnClient)); // Submit application System.out.println("Submitting Application: " + appContext.getApplicationName() + "\n"); try { isRunning = true; yarnClient.submitApplication(appContext); } catch (Exception exp) { System.err.println("Error Submitting Application"); exp.printStackTrace(); } // np = number of processes , + 1 for Application Master container IOMessagesThread[] ioThreads = new IOMessagesThread[np + 1]; peers = new String[np]; socketList = new Vector<Socket>(); int wport = 0; int rport = 0; int rank = 0; // np + 1 IOThreads for (int i = 0; i < (np + 1); i++) { try { sock = servSock.accept(); //start IO thread to read STDOUT and STDERR from wrappers IOMessagesThread io = new IOMessagesThread(sock); ioThreads[i] = io; ioThreads[i].start(); } catch (Exception e) { System.err.println("Error accepting connection from peer socket.."); e.printStackTrace(); } } // Loop to read port numbers from Wrapper.java processes // and to create WRAPPER_INFO (containing all IPs and ports) String WRAPPER_INFO = "#Peer Information"; for (int i = np; i > 0; i--) { try { sock = infoSock.accept(); DataOutputStream out = new DataOutputStream(sock.getOutputStream()); DataInputStream in = new DataInputStream(sock.getInputStream()); if (in.readUTF().startsWith("Sending Info")) { wport = in.readInt(); rport = in.readInt(); rank = in.readInt(); peers[rank] = ";" + sock.getInetAddress().getHostAddress() + "@" + rport + "@" + wport + "@" + rank; socketList.add(sock); } } catch (Exception e) { System.err.println("[MPJYarnClient.java]: Error accepting" + " connection from peer socket!"); e.printStackTrace(); } } for (int i = 0; i < np; i++) { WRAPPER_INFO += peers[i]; } // Loop to broadcast WRAPPER_INFO to all Wrappers for (int i = np; i > 0; i--) { try { sock = socketList.get(np - i); DataOutputStream out = new DataOutputStream(sock.getOutputStream()); out.writeUTF(WRAPPER_INFO); out.flush(); sock.close(); } catch (Exception e) { System.err.println("[MPJYarnClient.java]: Error closing" + " connection from peer socket.."); e.printStackTrace(); } } try { infoSock.close(); } catch (IOException exp) { exp.printStackTrace(); } // wait for all IO Threads to complete for (int i = 0; i < (np + 1); i++) { ioThreads[i].join(); } isRunning = true; System.out.println("\nApplication Statistics!"); while (true) { appReport = yarnClient.getApplicationReport(appId); appState = appReport.getYarnApplicationState(); fStatus = appReport.getFinalApplicationStatus(); if (appState == YarnApplicationState.FINISHED) { isRunning = false; if (fStatus == FinalApplicationStatus.SUCCEEDED) { System.out.println("State: " + fStatus); } else { System.out.println("State: " + fStatus); } break; } else if (appState == YarnApplicationState.KILLED) { isRunning = false; System.out.println("State: " + appState); break; } else if (appState == YarnApplicationState.FAILED) { isRunning = false; System.out.println("State: " + appState); break; } Thread.sleep(100); } try { if (debugYarn) { logger.info("Cleaning the files from hdfs: "); logger.info("1) " + dest.toString()); logger.info("2) " + wrapperDest.toString()); logger.info("3) " + userJarDest.toString()); } fs.delete(dest); fs.delete(wrapperDest); fs.delete(userJarDest); } catch (IOException exp) { exp.printStackTrace(); } System.out.println("Application ID: " + appId + "\n" + "Application User: " + appReport.getUser() + "\n" + "RM Queue: " + appReport.getQueue() + "\n" + "Start Time: " + appReport.getStartTime() + "\n" + "Finish Time: " + appReport.getFinishTime()); }
From source file:yrun.YarnRunner.java
License:Apache License
public void execute() throws IOException, YarnException, InterruptedException { LOG.info("Using application path [" + _installPath + "]"); Path jarPath = installThisJar(_installPath, _appJarFile); LOG.info("Driver installed [" + jarPath + "]"); List<Path> installedArchivePathList = install(_installPath, _archivePathList); for (Path p : installedArchivePathList) { LOG.info("Archive installed [" + p + "]"); }/* w w w .j av a 2s . co m*/ YarnRunnerArgs yarnRunnerArgs = new YarnRunnerArgs(); yarnRunnerArgs.setCommand(_command); Path argsPath = installThisArgs(_installPath, yarnRunnerArgs); final YarnClient client = YarnClient.createYarnClient(); _configuration.setInt("yarn.nodemanager.delete.debug-delay-sec", (int) TimeUnit.HOURS.toSeconds(1)); client.init(_configuration); client.start(); YarnClientApplication app = client.createApplication(); ContainerLaunchContext amContainer = Records.newRecord(ContainerLaunchContext.class); Map<String, String> appMasterEnv = new HashMap<String, String>(); setupAppMasterEnv(appMasterEnv, _appJarFile); Map<String, LocalResource> localResources = new HashMap<String, LocalResource>(); { LocalResource appMasterJar = Records.newRecord(LocalResource.class); setupAppMasterJar(jarPath, appMasterJar); localResources.put(jarPath.getName(), appMasterJar); } { LocalResource appMasterArgs = Records.newRecord(LocalResource.class); setupAppMasterArgs(argsPath, appMasterArgs); localResources.put(MASTER_JSON, appMasterArgs); } List<String> vargs = new ArrayList<String>(); vargs.add(Environment.JAVA_HOME.$() + "/bin/java"); vargs.add("-Xmx256m"); vargs.add("-Djava.net.preferIPv4Stack=true"); vargs.add(YarnRunnerApplicationMaster.class.getName()); String strCommand = "(echo ENV && set && echo CURRENT_DIR_LISTING && ls -la && echo PWD && pwd && (" + StringUtils.join(" ", vargs) + "))"; strCommand += " 1>" + ApplicationConstants.LOG_DIR_EXPANSION_VAR + "/stdout"; strCommand += " 2>" + ApplicationConstants.LOG_DIR_EXPANSION_VAR + "/stderr"; LOG.debug("Application Master command [" + strCommand + "]"); amContainer.setCommands(Collections.singletonList(strCommand)); amContainer.setLocalResources(localResources); amContainer.setEnvironment(appMasterEnv); Resource capability = Records.newRecord(Resource.class); capability.setMemory(256); capability.setVirtualCores(1); ApplicationSubmissionContext appContext = app.getApplicationSubmissionContext(); appContext.setApplicationName(_yarnName); appContext.setAMContainerSpec(amContainer); appContext.setResource(capability); if (_queue != null) { appContext.setQueue(_queue); } appContext.setApplicationType("yrun"); ApplicationId appId = appContext.getApplicationId(); AtomicBoolean shutdown = new AtomicBoolean(); if (!_isDaemon) { addShutdownHook(client, appId, shutdown); } LOG.info("Submitting application with id [" + appId + "]"); client.submitApplication(appContext); ApplicationReport report; YarnApplicationState state; do { report = client.getApplicationReport(appId); state = report.getYarnApplicationState(); if (state == YarnApplicationState.RUNNING) { if (_isDaemon) { LOG.info("Application is running. This is a daemon application driver program exiting."); return; } } Thread.sleep(100); } while (isNoLongerRunning(state)); shutdown.set(true); LOG.info("Application has finished with state [" + state + "]"); }