List of usage examples for org.apache.hadoop.yarn.api.records ApplicationReport getRpcPort
@Public @Stable public abstract int getRpcPort();
ApplicationMaster
. From source file:com.sogou.dockeronyarn.client.DockerClient.java
License:Apache License
/** * Monitor the submitted application for completion. * Kill application if time expires. /* w w w . j a va 2s . c o m*/ * @param appId Application Id of application to be monitored * @return true if application completed successfully * @throws YarnException * @throws IOException */ private boolean monitorApplication(ApplicationId appId) throws YarnException, IOException { while (true) { // Check app status every 1 second. try { Thread.sleep(1000); } catch (InterruptedException e) { LOG.debug("Thread sleep in monitoring loop interrupted"); } // Get application report for the appId we are interested in ApplicationReport report = yarnClient.getApplicationReport(appId); LOG.info("Got application report from ASM for" + ", appId=" + appId.getId() + ", clientToAMToken=" + report.getClientToAMToken() + ", appDiagnostics=" + report.getDiagnostics() + ", appMasterHost=" + report.getHost() + ", appQueue=" + report.getQueue() + ", appMasterRpcPort=" + report.getRpcPort() + ", appStartTime=" + report.getStartTime() + ", yarnAppState=" + report.getYarnApplicationState().toString() + ", distributedFinalState=" + report.getFinalApplicationStatus().toString() + ", appTrackingUrl=" + report.getTrackingUrl() + ", appUser=" + report.getUser()); YarnApplicationState state = report.getYarnApplicationState(); FinalApplicationStatus dsStatus = report.getFinalApplicationStatus(); if (YarnApplicationState.FINISHED == state) { if (FinalApplicationStatus.SUCCEEDED == dsStatus) { LOG.info("Application has completed successfully. Breaking monitoring loop"); return true; } else { LOG.info("Application did finished unsuccessfully." + " YarnState=" + state.toString() + ", DSFinalStatus=" + dsStatus.toString() + ". Breaking monitoring loop"); return false; } } else if (YarnApplicationState.KILLED == state || YarnApplicationState.FAILED == state) { LOG.info("Application did not shutdown." + " YarnState=" + state.toString() + ", DSFinalStatus=" + dsStatus.toString() + ". Breaking monitoring loop"); return false; } if (System.currentTimeMillis() > (clientStartTime + clientTimeout)) { LOG.info("Reached client specified timeout for application. Killing application"); forceKillApplication(appId); return false; } } }
From source file:com.yahoo.storm.yarn.StormOnYarn.java
License:Open Source License
@SuppressWarnings("unchecked") public synchronized StormMaster.Client getClient() throws YarnException, IOException { if (_client == null) { String host = null;//from w w w .ja v a 2 s. c o m int port = 0; //wait for application to be ready int max_wait_for_report = Utils.getInt(_stormConf.get(Config.YARN_REPORT_WAIT_MILLIS)); int waited = 0; while (waited < max_wait_for_report) { ApplicationReport report = _yarn.getApplicationReport(_appId); host = report.getHost(); port = report.getRpcPort(); if (host == null || port == 0) { try { Thread.sleep(1000); } catch (InterruptedException e) { } waited += 1000; } else { break; } } if (host == null || port == 0) { LOG.info("No host/port returned for Application Master " + _appId); return null; } LOG.info("application report for " + _appId + " :" + host + ":" + port); if (_stormConf == null) { _stormConf = new HashMap<Object, Object>(); } _stormConf.put(Config.MASTER_HOST, host); _stormConf.put(Config.MASTER_THRIFT_PORT, port); LOG.info("Attaching to " + host + ":" + port + " to talk to app master " + _appId); _client = MasterClient.getConfiguredClient(_stormConf); } return _client.getClient(); }
From source file:com.zqh.hadoop.moya.core.yarn.Client.java
License:Apache License
/** * Monitor the submitted application for completion. Kill application if * time expires./*from ww w .j a v a2s . c o m*/ * * @param appId * Application Id of application to be monitored * @return true if application completed successfully * @throws org.apache.hadoop.yarn.exceptions.YarnException * @throws java.io.IOException */ private boolean monitorApplication(ApplicationId appId) throws YarnException, IOException { while (true) { // Check app status every 1 second. try { Thread.sleep(1000); } catch (InterruptedException e) { LOG.debug("Thread sleep in monitoring loop interrupted"); } // Get application report for the appId we are interested in ApplicationReport report = yarnClient.getApplicationReport(appId); LOG.info("Got application report from ASM for" + ", appId=" + appId.getId() + ", clientToAMToken=" + report.getClientToAMToken() + ", appDiagnostics=" + report.getDiagnostics() + ", appMasterHost=" + report.getHost() + ", appQueue=" + report.getQueue() + ", appMasterRpcPort=" + report.getRpcPort() + ", appStartTime=" + report.getStartTime() + ", yarnAppState=" + report.getYarnApplicationState().toString() + ", distributedFinalState=" + report.getFinalApplicationStatus().toString() + ", appTrackingUrl=" + report.getTrackingUrl() + ", appUser=" + report.getUser()); YarnApplicationState state = report.getYarnApplicationState(); FinalApplicationStatus dsStatus = report.getFinalApplicationStatus(); if (YarnApplicationState.FINISHED == state) { if (FinalApplicationStatus.SUCCEEDED == dsStatus) { LOG.info("Application has completed successfully. Breaking monitoring loop"); return true; } else { LOG.info("Application did finished unsuccessfully." + " YarnState=" + state.toString() + ", DSFinalStatus=" + dsStatus.toString() + ". Breaking monitoring loop"); return false; } } else if (YarnApplicationState.KILLED == state || YarnApplicationState.FAILED == state) { LOG.info("Application did not finish." + " YarnState=" + state.toString() + ", DSFinalStatus=" + dsStatus.toString() + ". Breaking monitoring loop"); return false; } } }
From source file:io.hops.hopsworks.common.jobs.flink.AbstractYarnClusterDescriptor.java
License:Apache License
@Override public YarnClusterClient retrieve(String applicationID) { try {//from w ww. j a va 2 s . c o m // check if required Hadoop environment variables are set. If not, warn user if (System.getenv("HADOOP_CONF_DIR") == null && System.getenv("YARN_CONF_DIR") == null) { LOG.warn("Neither the HADOOP_CONF_DIR nor the YARN_CONF_DIR environment variable is set." + "The Flink YARN Client needs one of these to be set to properly load the Hadoop " + "configuration for accessing YARN."); } final ApplicationId yarnAppId = ConverterUtils.toApplicationId(applicationID); final YarnClient yarnClient = getYarnClient(); final ApplicationReport appReport = yarnClient.getApplicationReport(yarnAppId); if (appReport.getFinalApplicationStatus() != FinalApplicationStatus.UNDEFINED) { // Flink cluster is not running anymore LOG.error( "The application {} doesn't run anymore. It has previously completed with final status: {}", applicationID, appReport.getFinalApplicationStatus()); throw new RuntimeException("The Yarn application " + applicationID + " doesn't run anymore."); } LOG.info("Found application JobManager host name '{}' and port '{}' from supplied application id '{}'", appReport.getHost(), appReport.getRpcPort(), applicationID); flinkConfiguration.setString(ConfigConstants.JOB_MANAGER_IPC_ADDRESS_KEY, appReport.getHost()); flinkConfiguration.setInteger(ConfigConstants.JOB_MANAGER_IPC_PORT_KEY, appReport.getRpcPort()); return createYarnClusterClient(this, yarnClient, appReport, flinkConfiguration, sessionFilesDir, false); } catch (Exception e) { throw new RuntimeException("Couldn't retrieve Yarn cluster", e); } }
From source file:io.hops.hopsworks.common.jobs.flink.AbstractYarnClusterDescriptor.java
License:Apache License
/** * This method will block until the ApplicationMaster/JobManager have been * deployed on YARN./*from w w w. j a v a 2 s .c om*/ */ protected YarnClusterClient deployInternal() throws Exception { isReadyForDeployment(); LOG.info("Using values:"); LOG.info("\tTaskManager count = {}", taskManagerCount); LOG.info("\tJobManager memory = {}", jobManagerMemoryMb); LOG.info("\tTaskManager memory = {}", taskManagerMemoryMb); final YarnClient yarnClient = getYarnClient(); // ------------------ Check if the specified queue exists -------------------- try { List<QueueInfo> queues = yarnClient.getAllQueues(); // check only if there are queues configured in yarn and for this session. if (queues.size() > 0 && this.yarnQueue != null) { boolean queueFound = false; for (QueueInfo queue : queues) { if (queue.getQueueName().equals(this.yarnQueue)) { queueFound = true; break; } } if (!queueFound) { String queueNames = ""; for (QueueInfo queue : queues) { queueNames += queue.getQueueName() + ", "; } LOG.warn("The specified queue '" + this.yarnQueue + "' does not exist. " + "Available queues: " + queueNames); } } else { LOG.debug("The YARN cluster does not have any queues configured"); } } catch (Throwable e) { LOG.warn("Error while getting queue information from YARN: " + e.getMessage()); if (LOG.isDebugEnabled()) { LOG.debug("Error details", e); } } // Create application via yarnClient final YarnClientApplication yarnApplication = yarnClient.createApplication(); GetNewApplicationResponse appResponse = yarnApplication.getNewApplicationResponse(); Map<String, String> jobSystemProperties = new HashMap<>(2); // Certificates are materialized locally so DFSClient can be set to null // LocalResources are not used by Flink, so set it null HopsUtils.copyUserKafkaCerts(services.getUserCerts(), project, username, services.getSettings().getHopsworksTmpCertDir(), services.getSettings().getHdfsTmpCertDir(), JobType.FLINK, null, null, jobSystemProperties, services.getSettings().getFlinkKafkaCertDir(), appResponse.getApplicationId().toString()); StringBuilder tmpBuilder = new StringBuilder(); for (Map.Entry<String, String> prop : jobSystemProperties.entrySet()) { String option = YarnRunner.escapeForShell("-D" + prop.getKey() + "=" + prop.getValue()); javaOptions.add(option); addHopsworksParam(option); tmpBuilder.append(prop.getKey()).append("=").append(prop.getValue()).append("@@"); } dynamicPropertiesEncoded += tmpBuilder.toString(); // ------------------ Add dynamic properties to local flinkConfiguraton ------ Map<String, String> dynProperties = getDynamicProperties(dynamicPropertiesEncoded); for (Map.Entry<String, String> dynProperty : dynProperties.entrySet()) { flinkConfiguration.setString(dynProperty.getKey(), dynProperty.getValue()); } // ------------------ Set default file system scheme ------------------------- try { 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"); } // ------ Check if the YARN ClusterClient has the requested resources --- // the yarnMinAllocationMB specifies the smallest possible container // allocation size. all allocations below this value are automatically // set to this value. final int yarnMinAllocationMB = conf.getInt("yarn.scheduler.minimum-allocation-mb", 0); if (jobManagerMemoryMb < yarnMinAllocationMB || taskManagerMemoryMb < yarnMinAllocationMB) { LOG.warn("The JobManager or TaskManager memory is below the smallest possible " + "YARN Container size. The value of 'yarn.scheduler.minimum-allocation-mb'" + " is " + yarnMinAllocationMB + "'. Please increase the memory size." + "YARN will allocate the smaller containers but the scheduler will" + " account for the minimum-allocation-mb, maybe not all instances " + "you requested will start."); } // set the memory to minAllocationMB to do the next checks correctly if (jobManagerMemoryMb < yarnMinAllocationMB) { jobManagerMemoryMb = yarnMinAllocationMB; } if (taskManagerMemoryMb < yarnMinAllocationMB) { taskManagerMemoryMb = yarnMinAllocationMB; } Resource maxRes = appResponse.getMaximumResourceCapability(); final String NOTE = "Please check the 'yarn.scheduler.maximum-allocation-mb' and the " + "'yarn.nodemanager.resource.memory-mb' configuration values\n"; if (jobManagerMemoryMb > maxRes.getMemory()) { failSessionDuringDeployment(yarnClient, yarnApplication); throw new YarnDeploymentException("The cluster does not have the requested resources for the JobManager" + " available!\n" + "Maximum Memory: " + maxRes.getMemory() + "MB Requested: " + jobManagerMemoryMb + "MB. " + NOTE); } if (taskManagerMemoryMb > maxRes.getMemory()) { failSessionDuringDeployment(yarnClient, yarnApplication); throw new YarnDeploymentException( "The cluster does not have the requested resources for the TaskManagers available!\n" + "Maximum Memory: " + maxRes.getMemory() + " Requested: " + taskManagerMemoryMb + "MB. " + NOTE); } final String NOTE_RSC = "\nThe Flink YARN client will try to allocate the YARN session, " + "but maybe not all TaskManagers are connecting from the beginning " + "because the resources are currently not available in the cluster. " + "The allocation might take more time than usual because the Flink " + "YARN client needs to wait until the resources become available."; int totalMemoryRequired = jobManagerMemoryMb + taskManagerMemoryMb * taskManagerCount; ClusterResourceDescription freeClusterMem = getCurrentFreeClusterResources(yarnClient); if (freeClusterMem.totalFreeMemory < totalMemoryRequired) { LOG.warn("This YARN session requires " + totalMemoryRequired + "MB of memory in the cluster. " + "There are currently only " + freeClusterMem.totalFreeMemory + "MB available." + NOTE_RSC); } if (taskManagerMemoryMb > freeClusterMem.containerLimit) { LOG.warn("The requested amount of memory for the TaskManagers (" + taskManagerMemoryMb + "MB) is more than " + "the largest possible YARN container: " + freeClusterMem.containerLimit + NOTE_RSC); } if (jobManagerMemoryMb > freeClusterMem.containerLimit) { LOG.warn( "The requested amount of memory for the JobManager (" + jobManagerMemoryMb + "MB) is more than " + "the largest possible YARN container: " + freeClusterMem.containerLimit + NOTE_RSC); } // ----------------- check if the requested containers fit into the cluster. int[] nmFree = Arrays.copyOf(freeClusterMem.nodeManagersFree, freeClusterMem.nodeManagersFree.length); // first, allocate the jobManager somewhere. if (!allocateResource(nmFree, jobManagerMemoryMb)) { LOG.warn("Unable to find a NodeManager that can fit the JobManager/Application master. " + "The JobManager requires " + jobManagerMemoryMb + "MB. NodeManagers available: " + Arrays.toString(freeClusterMem.nodeManagersFree) + NOTE_RSC); } // allocate TaskManagers for (int i = 0; i < taskManagerCount; i++) { if (!allocateResource(nmFree, taskManagerMemoryMb)) { LOG.warn("There is not enough memory available in the YARN cluster. " + "The TaskManager(s) require " + taskManagerMemoryMb + "MB each. " + "NodeManagers available: " + Arrays.toString(freeClusterMem.nodeManagersFree) + "\n" + "After allocating the JobManager (" + jobManagerMemoryMb + "MB) and (" + i + "/" + taskManagerCount + ") TaskManagers, " + "the following NodeManagers are available: " + Arrays.toString(nmFree) + NOTE_RSC); } } 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); final ContainerLaunchContext amContainer = setupApplicationMasterContainer(hasLogback, hasLog4j); // Set-up ApplicationSubmissionContext for the application ApplicationSubmissionContext appContext = yarnApplication.getApplicationSubmissionContext(); 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(ConfigConstants.ZOOKEEPER_NAMESPACE_KEY, String.valueOf(appId)); setZookeeperNamespace(zkNamespace); } flinkConfiguration.setString(ConfigConstants.ZOOKEEPER_NAMESPACE_KEY, zkNamespace); if (RecoveryMode.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); classPathBuilder.append(shipFile.getName()); if (shipFile.isDirectory()) { // add directories to the classpath classPathBuilder.append(File.separator).append("*"); } classPathBuilder.append(File.pathSeparator); envShipFileList.append(remotePath).append(","); } //////////////////////////////////////////////////////////////////////////// /* * Add Hops LocalResources paths here * */ //Add it to localResources for (Entry<String, LocalResource> entry : hopsworksResources.entrySet()) { localResources.put(entry.getKey(), entry.getValue()); //Append name to classPathBuilder classPathBuilder.append(entry.getKey()); classPathBuilder.append(File.pathSeparator); } //////////////////////////////////////////////////////////////////////////// // 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); 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. // setup security tokens 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_CLIENT_USERNAME, UserGroupInformation.getCurrentUser().getShortUserName()); appMasterEnv.put(YarnConfigKeys.ENV_SLOTS, String.valueOf(slots)); appMasterEnv.put(YarnConfigKeys.ENV_DETACHED, String.valueOf(detached)); appMasterEnv.put(YarnConfigKeys.ENV_ZOOKEEPER_NAMESPACE, getZookeeperNamespace()); 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); // application name appContext.setApplicationType("Apache Flink"); appContext.setAMContainerSpec(amContainer); appContext.setResource(capability); if (yarnQueue != null) { appContext.setQueue(yarnQueue); } // 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.getMessage()); } YarnApplicationState appState = report.getYarnApplicationState(); switch (appState) { case FAILED: case FINISHED: 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. } String host = report.getHost(); int port = report.getRpcPort(); // Correctly initialize the Flink config flinkConfiguration.setString(ConfigConstants.JOB_MANAGER_IPC_ADDRESS_KEY, host); flinkConfiguration.setInteger(ConfigConstants.JOB_MANAGER_IPC_PORT_KEY, port); // the Flink cluster is deployed in YARN. Represent cluster return createYarnClusterClient(this, yarnClient, report, flinkConfiguration, sessionFilesDir, true); }
From source file:io.hops.tensorflow.Client.java
License:Apache License
/** * Monitor the submitted application for completion. * Kill application if time expires./* w w w . j av a 2 s . co m*/ * * @param appId * Application Id of application to be monitored * @return true if application completed successfully * @throws YarnException * @throws IOException */ public boolean monitorApplication(ApplicationId appId) throws YarnException, IOException { YarnApplicationState oldState = null; while (true) { // Check app status every 1 second. try { Thread.sleep(1000); } catch (InterruptedException e) { LOG.debug("Thread sleep in monitoring loop interrupted"); } ApplicationReport report = yarnClient.getApplicationReport(appId); YarnApplicationState state = report.getYarnApplicationState(); FinalApplicationStatus dsStatus = report.getFinalApplicationStatus(); if (oldState != state) { LOG.info("Got application report from ASM for" + "\n\t appId=" + appId.getId() + "\n\t clientToAMToken=" + report.getClientToAMToken() + "\n\t appDiagnostics=" + report.getDiagnostics() + "\n\t appMasterHost=" + report.getHost() + "\n\t appQueue=" + report.getQueue() + "\n\t appMasterRpcPort=" + report.getRpcPort() + "\n\t appStartTime=" + report.getStartTime() + "\n\t yarnAppState=" + report.getYarnApplicationState().toString() + "\n\t distributedFinalState=" + report.getFinalApplicationStatus().toString() + "\n\t appTrackingUrl=" + report.getTrackingUrl() + "\n\t appUser=" + report.getUser()); oldState = state; } else { LOG.info("Got application report from ASM for " + appId + " (state: " + state + ")"); } if (YarnApplicationState.FINISHED == state) { if (FinalApplicationStatus.SUCCEEDED == dsStatus) { LOG.info("Application has completed successfully. Breaking monitoring loop"); return true; } else { LOG.info("Application did finished unsuccessfully." + " YarnState=" + state.toString() + ", DSFinalStatus=" + dsStatus.toString() + ". Breaking monitoring loop"); return false; } } else if (YarnApplicationState.KILLED == state || YarnApplicationState.FAILED == state) { LOG.info("Application did not finish." + " YarnState=" + state.toString() + ", DSFinalStatus=" + dsStatus.toString() + ". Breaking monitoring loop"); return false; } if (System.currentTimeMillis() > (clientStartTime + clientTimeout)) { LOG.info("Reached client specified timeout for application. Killing application"); forceKillApplication(appId); return false; } } }
From source file:ml.shifu.guagua.yarn.GuaguaYarnClient.java
License:Apache License
/** * Assess whether job is already finished/failed and 'done' flag needs to be * set, prints progress display for client if all is going well. * //from w ww . j a v a 2 s.com * @param report * the application report to assess. * @return true if job report indicates the job run is over. */ private boolean checkProgress(final ApplicationReport report) { YarnApplicationState jobState = report.getYarnApplicationState(); LOG.info( "Got applicaton report for appId={}, state={}, progress={}%, amDiag={}, masterHost={}, masterRpcPort={}, queue={}, startTime={}, clientToken={}, finalState={}, trackingUrl={}, user={}", this.appId.getId(), report.getYarnApplicationState().toString(), DF.format(report.getProgress() * 100), report.getDiagnostics(), report.getHost(), report.getRpcPort(), report.getQueue(), report.getStartTime(), report.getClientToAMToken(), report.getFinalApplicationStatus().toString(), report.getTrackingUrl(), report.getUser()); switch (jobState) { case FINISHED: LOG.info("Application finished in {} ms", (System.currentTimeMillis() - getStartTime())); return true; case KILLED: LOG.error("{} reports KILLED state, diagnostics show: {}", getAppName(), report.getDiagnostics()); return true; case FAILED: LOG.error("{} reports FAILED state, diagnostics show: {}", getAppName(), report.getDiagnostics()); return true; default: if (this.reportCounter++ % 5 == 0) { displayJobReport(report); } return false; } }
From source file:org.apache.flink.yarn.AbstractYarnClusterDescriptor.java
License:Apache License
/** * This method will block until the ApplicationMaster/JobManager have been * deployed on YARN.//from w ww .ja v a 2s .c o m */ protected YarnClusterClient deployInternal() throws Exception { isReadyForDeployment(); LOG.info("Using values:"); LOG.info("\tTaskManager count = {}", taskManagerCount); LOG.info("\tJobManager memory = {}", jobManagerMemoryMb); LOG.info("\tTaskManager memory = {}", taskManagerMemoryMb); final YarnClient yarnClient = getYarnClient(); // ------------------ Check if the specified queue exists -------------------- try { List<QueueInfo> queues = yarnClient.getAllQueues(); if (queues.size() > 0 && this.yarnQueue != null) { // check only if there are queues configured in yarn and for this session. boolean queueFound = false; for (QueueInfo queue : queues) { if (queue.getQueueName().equals(this.yarnQueue)) { queueFound = true; break; } } if (!queueFound) { String queueNames = ""; for (QueueInfo queue : queues) { queueNames += queue.getQueueName() + ", "; } LOG.warn("The specified queue '" + this.yarnQueue + "' does not exist. " + "Available queues: " + queueNames); } } else { LOG.debug("The YARN cluster does not have any queues configured"); } } catch (Throwable e) { LOG.warn("Error while getting queue information from YARN: " + e.getMessage()); if (LOG.isDebugEnabled()) { LOG.debug("Error details", e); } } // ------------------ Add dynamic properties to local flinkConfiguraton ------ Map<String, String> dynProperties = getDynamicProperties(dynamicPropertiesEncoded); for (Map.Entry<String, String> dynProperty : dynProperties.entrySet()) { flinkConfiguration.setString(dynProperty.getKey(), dynProperty.getValue()); } // ------------------ Check if the YARN ClusterClient has the requested resources -------------- // the yarnMinAllocationMB specifies the smallest possible container allocation size. // all allocations below this value are automatically set to this value. final int yarnMinAllocationMB = conf.getInt("yarn.scheduler.minimum-allocation-mb", 0); if (jobManagerMemoryMb < yarnMinAllocationMB || taskManagerMemoryMb < yarnMinAllocationMB) { LOG.warn("The JobManager or TaskManager memory is below the smallest possible YARN Container size. " + "The value of 'yarn.scheduler.minimum-allocation-mb' is '" + yarnMinAllocationMB + "'. Please increase the memory size." + "YARN will allocate the smaller containers but the scheduler will account for the minimum-allocation-mb, maybe not all instances " + "you requested will start."); } // set the memory to minAllocationMB to do the next checks correctly if (jobManagerMemoryMb < yarnMinAllocationMB) { jobManagerMemoryMb = yarnMinAllocationMB; } if (taskManagerMemoryMb < yarnMinAllocationMB) { taskManagerMemoryMb = yarnMinAllocationMB; } // Create application via yarnClient final YarnClientApplication yarnApplication = yarnClient.createApplication(); GetNewApplicationResponse appResponse = yarnApplication.getNewApplicationResponse(); Resource maxRes = appResponse.getMaximumResourceCapability(); final String NOTE = "Please check the 'yarn.scheduler.maximum-allocation-mb' and the 'yarn.nodemanager.resource.memory-mb' configuration values\n"; if (jobManagerMemoryMb > maxRes.getMemory()) { failSessionDuringDeployment(yarnClient, yarnApplication); throw new YarnDeploymentException( "The cluster does not have the requested resources for the JobManager available!\n" + "Maximum Memory: " + maxRes.getMemory() + "MB Requested: " + jobManagerMemoryMb + "MB. " + NOTE); } if (taskManagerMemoryMb > maxRes.getMemory()) { failSessionDuringDeployment(yarnClient, yarnApplication); throw new YarnDeploymentException( "The cluster does not have the requested resources for the TaskManagers available!\n" + "Maximum Memory: " + maxRes.getMemory() + " Requested: " + taskManagerMemoryMb + "MB. " + NOTE); } final String NOTE_RSC = "\nThe Flink YARN client will try to allocate the YARN session, but maybe not all TaskManagers are " + "connecting from the beginning because the resources are currently not available in the cluster. " + "The allocation might take more time than usual because the Flink YARN client needs to wait until " + "the resources become available."; int totalMemoryRequired = jobManagerMemoryMb + taskManagerMemoryMb * taskManagerCount; ClusterResourceDescription freeClusterMem = getCurrentFreeClusterResources(yarnClient); if (freeClusterMem.totalFreeMemory < totalMemoryRequired) { LOG.warn("This YARN session requires " + totalMemoryRequired + "MB of memory in the cluster. " + "There are currently only " + freeClusterMem.totalFreeMemory + "MB available." + NOTE_RSC); } if (taskManagerMemoryMb > freeClusterMem.containerLimit) { LOG.warn("The requested amount of memory for the TaskManagers (" + taskManagerMemoryMb + "MB) is more than " + "the largest possible YARN container: " + freeClusterMem.containerLimit + NOTE_RSC); } if (jobManagerMemoryMb > freeClusterMem.containerLimit) { LOG.warn( "The requested amount of memory for the JobManager (" + jobManagerMemoryMb + "MB) is more than " + "the largest possible YARN container: " + freeClusterMem.containerLimit + NOTE_RSC); } // ----------------- check if the requested containers fit into the cluster. int[] nmFree = Arrays.copyOf(freeClusterMem.nodeManagersFree, freeClusterMem.nodeManagersFree.length); // first, allocate the jobManager somewhere. if (!allocateResource(nmFree, jobManagerMemoryMb)) { LOG.warn("Unable to find a NodeManager that can fit the JobManager/Application master. " + "The JobManager requires " + jobManagerMemoryMb + "MB. NodeManagers available: " + Arrays.toString(freeClusterMem.nodeManagersFree) + NOTE_RSC); } // allocate TaskManagers for (int i = 0; i < taskManagerCount; i++) { if (!allocateResource(nmFree, taskManagerMemoryMb)) { LOG.warn("There is not enough memory available in the YARN cluster. " + "The TaskManager(s) require " + taskManagerMemoryMb + "MB each. " + "NodeManagers available: " + Arrays.toString(freeClusterMem.nodeManagersFree) + "\n" + "After allocating the JobManager (" + jobManagerMemoryMb + "MB) and (" + i + "/" + taskManagerCount + ") TaskManagers, " + "the following NodeManagers are available: " + Arrays.toString(nmFree) + NOTE_RSC); } } ApplicationReport report = startAppMaster(null, yarnClient, yarnApplication); String host = report.getHost(); int port = report.getRpcPort(); // Correctly initialize the Flink config flinkConfiguration.setString(ConfigConstants.JOB_MANAGER_IPC_ADDRESS_KEY, host); flinkConfiguration.setInteger(ConfigConstants.JOB_MANAGER_IPC_PORT_KEY, port); // the Flink cluster is deployed in YARN. Represent cluster return createYarnClusterClient(this, yarnClient, report, flinkConfiguration, sessionFilesDir, true); }
From source file:org.apache.helix.provisioning.yarn.AppLauncher.java
License:Apache License
/** * TODO: kill the app only in dev mode. In prod, its ok for the app to continue running if the * launcher dies after launching// ww w . ja va 2 s . c om */ private String generateReport(ApplicationReport report) { return "Got application report from ASM for" + ", appId=" + _appId.getId() + ", clientToAMToken=" + report.getClientToAMToken() + ", appDiagnostics=" + report.getDiagnostics() + ", appMasterHost=" + report.getHost() + ", appQueue=" + report.getQueue() + ", appMasterRpcPort=" + report.getRpcPort() + ", appStartTime=" + report.getStartTime() + ", yarnAppState=" + report.getYarnApplicationState().toString() + ", distributedFinalState=" + report.getFinalApplicationStatus().toString() + ", appTrackingUrl=" + report.getTrackingUrl() + ", appUser=" + report.getUser(); }
From source file:org.apache.hoya.tools.HoyaUtils.java
License:Apache License
public static String appReportToString(ApplicationReport r, String separator) { StringBuilder builder = new StringBuilder(512); builder.append("application ").append(r.getName()).append("/").append(r.getApplicationType()); builder.append(separator).append("state: ").append(r.getYarnApplicationState()); builder.append(separator).append("URL: ").append(r.getTrackingUrl()); builder.append(separator).append("Started ").append(new Date(r.getStartTime()).toGMTString()); long finishTime = r.getFinishTime(); if (finishTime > 0) { builder.append(separator).append("Finished ").append(new Date(finishTime).toGMTString()); }//from www . j ava 2 s . c o m builder.append(separator).append("RPC :").append(r.getHost()).append(':').append(r.getRpcPort()); String diagnostics = r.getDiagnostics(); if (!diagnostics.isEmpty()) { builder.append(separator).append("Diagnostics :").append(diagnostics); } return builder.toString(); }