Examples with YarnClientApplication org.apache.hadoop.yarn.client.api.YarnClientApplication used on opensource projects

Example 1 with YarnClientApplication

use of org.apache.hadoop.yarn.client.api.YarnClientApplication in project flink by apache.

the class AbstractYarnClusterDescriptor method deployInternal.

/**
	 * This method will block until the ApplicationMaster/JobManager have been
	 * deployed on YARN.
	 */
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();
    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);
}

Example 2 with YarnClientApplication

use of org.apache.hadoop.yarn.client.api.YarnClientApplication in project hadoop by apache.

the class Client method run.

/**
   * Main run function for the client
   * @return true if application completed successfully
   * @throws IOException
   * @throws YarnException
   */
public boolean run() throws IOException, YarnException {
    LOG.info("Running Client");
    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(this.amQueue);
    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());
        }
    }
    if (domainId != null && domainId.length() > 0 && toCreateDomain) {
        prepareTimelineDomain();
    }
    // Get a new application id
    YarnClientApplication app = yarnClient.createApplication();
    GetNewApplicationResponse appResponse = app.getNewApplicationResponse();
    // TODO get min/max resource capabilities from RM and change memory ask if needed
    // If we do not have min/max, we may not be able to correctly request 
    // the required resources from the RM for the app master
    // Memory ask has to be a multiple of min and less than max. 
    // Dump out information about cluster capability as seen by the resource manager
    long maxMem = appResponse.getMaximumResourceCapability().getMemorySize();
    LOG.info("Max mem capability of resources in this cluster " + maxMem);
    // A resource ask cannot exceed the max. 
    if (amMemory > maxMem) {
        LOG.info("AM memory specified above max threshold of cluster. Using max value." + ", specified=" + amMemory + ", max=" + maxMem);
        amMemory = maxMem;
    }
    int maxVCores = appResponse.getMaximumResourceCapability().getVirtualCores();
    LOG.info("Max virtual cores capability of resources in this cluster " + maxVCores);
    if (amVCores > maxVCores) {
        LOG.info("AM virtual cores specified above max threshold of cluster. " + "Using max value." + ", specified=" + amVCores + ", max=" + maxVCores);
        amVCores = maxVCores;
    }
    // set the application name
    ApplicationSubmissionContext appContext = app.getApplicationSubmissionContext();
    ApplicationId appId = appContext.getApplicationId();
    appContext.setKeepContainersAcrossApplicationAttempts(keepContainers);
    appContext.setApplicationName(appName);
    if (attemptFailuresValidityInterval >= 0) {
        appContext.setAttemptFailuresValidityInterval(attemptFailuresValidityInterval);
    }
    Set<String> tags = new HashSet<String>();
    if (flowName != null) {
        tags.add(TimelineUtils.generateFlowNameTag(flowName));
    }
    if (flowVersion != null) {
        tags.add(TimelineUtils.generateFlowVersionTag(flowVersion));
    }
    if (flowRunId != 0) {
        tags.add(TimelineUtils.generateFlowRunIdTag(flowRunId));
    }
    appContext.setApplicationTags(tags);
    // set local resources for the application master
    // local files or archives as needed
    // In this scenario, the jar file for the application master is part of the local resources			
    Map<String, LocalResource> localResources = new HashMap<String, LocalResource>();
    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 
    FileSystem fs = FileSystem.get(conf);
    addToLocalResources(fs, appMasterJar, appMasterJarPath, appId.toString(), localResources, null);
    // Set the log4j properties if needed 
    if (!log4jPropFile.isEmpty()) {
        addToLocalResources(fs, log4jPropFile, log4jPath, appId.toString(), localResources, null);
    }
    // The shell script has to be made available on the final container(s)
    // where it will be executed. 
    // To do this, we need to first copy into the filesystem that is visible 
    // to the yarn framework. 
    // We do not need to set this as a local resource for the application 
    // master as the application master does not need it. 		
    String hdfsShellScriptLocation = "";
    long hdfsShellScriptLen = 0;
    long hdfsShellScriptTimestamp = 0;
    if (!shellScriptPath.isEmpty()) {
        Path shellSrc = new Path(shellScriptPath);
        String shellPathSuffix = appName + "/" + appId.toString() + "/" + SCRIPT_PATH;
        Path shellDst = new Path(fs.getHomeDirectory(), shellPathSuffix);
        fs.copyFromLocalFile(false, true, shellSrc, shellDst);
        hdfsShellScriptLocation = shellDst.toUri().toString();
        FileStatus shellFileStatus = fs.getFileStatus(shellDst);
        hdfsShellScriptLen = shellFileStatus.getLen();
        hdfsShellScriptTimestamp = shellFileStatus.getModificationTime();
    }
    if (!shellCommand.isEmpty()) {
        addToLocalResources(fs, null, shellCommandPath, appId.toString(), localResources, shellCommand);
    }
    if (shellArgs.length > 0) {
        addToLocalResources(fs, null, shellArgsPath, appId.toString(), localResources, StringUtils.join(shellArgs, " "));
    }
    // Set the necessary security tokens as needed
    //amContainer.setContainerTokens(containerToken);
    // Set the env variables to be setup in the env where the application master will be run
    LOG.info("Set the environment for the application master");
    Map<String, String> env = new HashMap<String, String>();
    // put location of shell script into env
    // using the env info, the application master will create the correct local resource for the 
    // eventual containers that will be launched to execute the shell scripts
    env.put(DSConstants.DISTRIBUTEDSHELLSCRIPTLOCATION, hdfsShellScriptLocation);
    env.put(DSConstants.DISTRIBUTEDSHELLSCRIPTTIMESTAMP, Long.toString(hdfsShellScriptTimestamp));
    env.put(DSConstants.DISTRIBUTEDSHELLSCRIPTLEN, Long.toString(hdfsShellScriptLen));
    if (domainId != null && domainId.length() > 0) {
        env.put(DSConstants.DISTRIBUTEDSHELLTIMELINEDOMAIN, domainId);
    }
    // Add AppMaster.jar location to classpath 		
    // At some point we should not be required to add 
    // the hadoop specific classpaths to the env. 
    // It should be provided out of the box. 
    // For now setting all required classpaths including
    // the classpath to "." for the application jar
    StringBuilder classPathEnv = new StringBuilder(Environment.CLASSPATH.$$()).append(ApplicationConstants.CLASS_PATH_SEPARATOR).append("./*");
    for (String c : conf.getStrings(YarnConfiguration.YARN_APPLICATION_CLASSPATH, YarnConfiguration.DEFAULT_YARN_CROSS_PLATFORM_APPLICATION_CLASSPATH)) {
        classPathEnv.append(ApplicationConstants.CLASS_PATH_SEPARATOR);
        classPathEnv.append(c.trim());
    }
    classPathEnv.append(ApplicationConstants.CLASS_PATH_SEPARATOR).append("./log4j.properties");
    // add the runtime classpath needed for tests to work
    if (conf.getBoolean(YarnConfiguration.IS_MINI_YARN_CLUSTER, false)) {
        classPathEnv.append(':');
        classPathEnv.append(System.getProperty("java.class.path"));
    }
    env.put("CLASSPATH", classPathEnv.toString());
    // Set the necessary command to execute the application master 
    Vector<CharSequence> vargs = new Vector<CharSequence>(30);
    // Set java executable command 
    LOG.info("Setting up app master command");
    vargs.add(Environment.JAVA_HOME.$$() + "/bin/java");
    // Set Xmx based on am memory size
    vargs.add("-Xmx" + amMemory + "m");
    // Set class name 
    vargs.add(appMasterMainClass);
    // Set params for Application Master
    vargs.add("--container_memory " + String.valueOf(containerMemory));
    vargs.add("--container_vcores " + String.valueOf(containerVirtualCores));
    vargs.add("--num_containers " + String.valueOf(numContainers));
    if (null != nodeLabelExpression) {
        appContext.setNodeLabelExpression(nodeLabelExpression);
    }
    vargs.add("--priority " + String.valueOf(shellCmdPriority));
    for (Map.Entry<String, String> entry : shellEnv.entrySet()) {
        vargs.add("--shell_env " + entry.getKey() + "=" + entry.getValue());
    }
    if (debugFlag) {
        vargs.add("--debug");
    }
    vargs.addAll(containerRetryOptions);
    vargs.add("1>" + ApplicationConstants.LOG_DIR_EXPANSION_VAR + "/AppMaster.stdout");
    vargs.add("2>" + ApplicationConstants.LOG_DIR_EXPANSION_VAR + "/AppMaster.stderr");
    // Get final commmand
    StringBuilder command = new StringBuilder();
    for (CharSequence str : vargs) {
        command.append(str).append(" ");
    }
    LOG.info("Completed setting up app master command " + command.toString());
    List<String> commands = new ArrayList<String>();
    commands.add(command.toString());
    // Set up the container launch context for the application master
    ContainerLaunchContext amContainer = ContainerLaunchContext.newInstance(localResources, env, commands, null, null, null);
    // Set up resource type requirements
    // For now, both memory and vcores are supported, so we set memory and 
    // vcores requirements
    Resource capability = Resource.newInstance(amMemory, amVCores);
    appContext.setResource(capability);
    // Setup security tokens
    if (UserGroupInformation.isSecurityEnabled()) {
        // Note: Credentials class is marked as LimitedPrivate for HDFS and MapReduce
        Credentials credentials = new Credentials();
        String tokenRenewer = YarnClientUtils.getRmPrincipal(conf);
        if (tokenRenewer == null || tokenRenewer.length() == 0) {
            throw new IOException("Can't get Master Kerberos principal for the RM to use as renewer");
        }
        // For now, only getting tokens for the default file-system.
        final Token<?>[] tokens = fs.addDelegationTokens(tokenRenewer, credentials);
        if (tokens != null) {
            for (Token<?> token : tokens) {
                LOG.info("Got dt for " + fs.getUri() + "; " + token);
            }
        }
        DataOutputBuffer dob = new DataOutputBuffer();
        credentials.writeTokenStorageToStream(dob);
        ByteBuffer fsTokens = ByteBuffer.wrap(dob.getData(), 0, dob.getLength());
        amContainer.setTokens(fsTokens);
    }
    appContext.setAMContainerSpec(amContainer);
    // Set the priority for the application master
    // TODO - what is the range for priority? how to decide? 
    Priority pri = Priority.newInstance(amPriority);
    appContext.setPriority(pri);
    // Set the queue to which this application is to be submitted in the RM
    appContext.setQueue(amQueue);
    // 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);
    // Monitor the application
    return monitorApplication(appId);
}

Example 3 with YarnClientApplication

use of org.apache.hadoop.yarn.client.api.YarnClientApplication in project kitten by cloudera.

the class YarnClientServiceImpl method startUp.

@Override
protected void startUp() throws IOException {
    ByteBuffer serializedTokens = null;
    if (UserGroupInformation.isSecurityEnabled()) {
        Configuration conf = this.yarnClientFactory.getConfig();
        FileSystem fs = FileSystem.get(conf);
        Credentials credentials = new Credentials();
        String tokenRenewer = this.yarnClientFactory.getConfig().get(YarnConfiguration.RM_PRINCIPAL);
        if (tokenRenewer == null || tokenRenewer.length() == 0) {
            throw new IOException("Can't get Master Kerberos principal for the RM to use as renewer");
        }
        // For now, only getting tokens for the default file-system.
        final Token<?>[] tokens = fs.addDelegationTokens(tokenRenewer, credentials);
        if (tokens != null) {
            for (Token<?> token : tokens) {
                LOG.info("Got delegation token for " + fs.getUri() + "; " + token);
            }
        }
        DataOutputBuffer dob = new DataOutputBuffer();
        credentials.writeTokenStorageToStream(dob);
        serializedTokens = ByteBuffer.wrap(dob.getData(), 0, dob.getLength());
    }
    this.yarnClient = yarnClientFactory.connect();
    YarnClientApplication clientApp = getNewApplication();
    GetNewApplicationResponse newApp = clientApp.getNewApplicationResponse();
    ContainerLaunchContextFactory clcFactory = new ContainerLaunchContextFactory(newApp.getMaximumResourceCapability(), serializedTokens);
    ApplicationSubmissionContext appContext = clientApp.getApplicationSubmissionContext();
    this.applicationId = appContext.getApplicationId();
    appContext.setApplicationName(parameters.getApplicationName());
    // Setup the container for the application master.
    ContainerLaunchParameters appMasterParams = parameters.getApplicationMasterParameters(applicationId);
    ContainerLaunchContext clc = clcFactory.create(appMasterParams);
    LOG.debug("Master context: " + clc);
    appContext.setResource(clcFactory.createResource(appMasterParams));
    appContext.setQueue(parameters.getQueue());
    appContext.setPriority(clcFactory.createPriority(appMasterParams.getPriority()));
    appContext.setAMContainerSpec(clc);
    submitApplication(appContext);
    // Make sure we stop the application in the case that it isn't done already.
    Runtime.getRuntime().addShutdownHook(new Thread() {

        @Override
        public void run() {
            if (YarnClientServiceImpl.this.isRunning()) {
                YarnClientServiceImpl.this.stop();
            }
        }
    });
    stopwatch.start();
}

Example 4 with YarnClientApplication

use of org.apache.hadoop.yarn.client.api.YarnClientApplication in project alluxio by Alluxio.

the class Client method submitApplication.

/**
   * Submits an application to the ResourceManager to run ApplicationMaster.
   *
   * The stable Yarn API provides a convenience method (YarnClient#createApplication) for creating
   * applications and setting up the application submission context. This was not available in the
   * alpha API.
   */
private void submitApplication() throws YarnException, IOException {
    // Initialize a YarnClient
    mYarnClient = YarnClient.createYarnClient();
    mYarnClient.init(mYarnConf);
    mYarnClient.start();
    // Create an application, get and check the information about the cluster
    YarnClientApplication app = mYarnClient.createApplication();
    // Get a response of this application, containing information of the cluster
    GetNewApplicationResponse appResponse = app.getNewApplicationResponse();
    // Check if the cluster has enough resource to launch the ApplicationMaster
    checkClusterResource(appResponse);
    // Check that there are enough hosts in the cluster to support the desired number of workers
    checkNodesAvailable();
    // Set up the container launch context for the application master
    mAmContainer = Records.newRecord(ContainerLaunchContext.class);
    setupContainerLaunchContext();
    // Finally, set-up ApplicationSubmissionContext for the application
    mAppContext = app.getApplicationSubmissionContext();
    setupApplicationSubmissionContext();
    // Submit the application to the applications manager.
    // Ignore the response as either a valid response object is returned on success
    // or an exception thrown to denote some form of a failure
    mAppId = mAppContext.getApplicationId();
    System.out.println("Submitting application of id " + mAppId + " to ResourceManager");
    mYarnClient.submitApplication(mAppContext);
    monitorApplication();
}

Example 5 with YarnClientApplication

use of org.apache.hadoop.yarn.client.api.YarnClientApplication in project alluxio by Alluxio.

the class ClientTest method generateMaxAllocation.

private void generateMaxAllocation(Resource resource) throws Exception {
    YarnClientApplication mockYarnClientApplication = mock(YarnClientApplication.class);
    doReturn(mockYarnClientApplication).when(mYarnClient).createApplication();
    GetNewApplicationResponse mockNewApplicationResponse = mock(GetNewApplicationResponse.class);
    doReturn(mockNewApplicationResponse).when(mockYarnClientApplication).getNewApplicationResponse();
    doReturn(resource).when(mockNewApplicationResponse).getMaximumResourceCapability();
}