List of usage examples for org.apache.hadoop.yarn.api.records NodeState RUNNING
NodeState RUNNING
To view the source code for org.apache.hadoop.yarn.api.records NodeState RUNNING.
Click Source Link
From source file:io.amient.yarn1.YarnClient.java
License:Open Source License
/** * This method should be called by the implementing application static main * method. It does all the work around creating a yarn application and * submitting the request to the yarn resource manager. The class given in * the appClass argument will be run inside the yarn-allocated master * container.//w ww.j a v a2 s . co m */ public static void submitApplicationMaster(Properties appConfig, Class<? extends YarnMaster> masterClass, String[] args, Boolean awaitCompletion) throws Exception { log.info("Yarn1 App Configuration:"); for (Object param : appConfig.keySet()) { log.info(param.toString() + " = " + appConfig.get(param).toString()); } String yarnConfigPath = appConfig.getProperty("yarn1.site", "/etc/hadoop"); String masterClassName = masterClass.getName(); appConfig.setProperty("yarn1.master.class", masterClassName); String applicationName = appConfig.getProperty("yarn1.application.name", masterClassName); log.info("--------------------------------------------------------------"); if (Boolean.valueOf(appConfig.getProperty("yarn1.local.mode", "false"))) { YarnMaster.run(appConfig, args); return; } int masterPriority = Integer.valueOf( appConfig.getProperty("yarn1.master.priority", String.valueOf(YarnMaster.DEFAULT_MASTER_PRIORITY))); int masterMemoryMb = Integer.valueOf(appConfig.getProperty("yarn1.master.memory.mb", String.valueOf(YarnMaster.DEFAULT_MASTER_MEMORY_MB))); int masterNumCores = Integer.valueOf( appConfig.getProperty("yarn1.master.num.cores", String.valueOf(YarnMaster.DEFAULT_MASTER_CORES))); String queue = appConfig.getProperty("yarn1.queue"); Configuration yarnConfig = new YarnConfiguration(); yarnConfig.addResource(new FileInputStream(yarnConfigPath + "/core-site.xml")); yarnConfig.addResource(new FileInputStream(yarnConfigPath + "/hdfs-site.xml")); yarnConfig.addResource(new FileInputStream(yarnConfigPath + "/yarn-site.xml")); for (Map.Entry<Object, Object> entry : appConfig.entrySet()) { yarnConfig.set(entry.getKey().toString(), entry.getValue().toString()); } final org.apache.hadoop.yarn.client.api.YarnClient yarnClient = org.apache.hadoop.yarn.client.api.YarnClient .createYarnClient(); yarnClient.init(yarnConfig); yarnClient.start(); for (NodeReport report : yarnClient.getNodeReports(NodeState.RUNNING)) { log.debug("Node report:" + report.getNodeId() + " @ " + report.getHttpAddress() + " | " + report.getCapability()); } log.info("Submitting application master class " + masterClassName); YarnClientApplication app = yarnClient.createApplication(); GetNewApplicationResponse appResponse = app.getNewApplicationResponse(); final ApplicationId appId = appResponse.getApplicationId(); if (appId == null) { System.exit(111); } else { appConfig.setProperty("am.timestamp", String.valueOf(appId.getClusterTimestamp())); appConfig.setProperty("am.id", String.valueOf(appId.getId())); } YarnClient.distributeResources(yarnConfig, appConfig, applicationName); String masterJvmArgs = appConfig.getProperty("yarn1.master.jvm.args", ""); YarnContainerContext masterContainer = new YarnContainerContext(yarnConfig, appConfig, masterJvmArgs, masterPriority, masterMemoryMb, masterNumCores, applicationName, YarnMaster.class, args); ApplicationSubmissionContext appContext = app.getApplicationSubmissionContext(); appContext.setApplicationName(masterClassName); appContext.setResource(masterContainer.capability); appContext.setPriority(masterContainer.priority); appContext.setQueue(queue); appContext.setApplicationType(appConfig.getProperty("yarn1.application.type", "YARN")); appContext.setAMContainerSpec(masterContainer.createContainerLaunchContext()); log.info("Master container spec: " + masterContainer.capability); yarnClient.submitApplication(appContext); ApplicationReport report = yarnClient.getApplicationReport(appId); log.info("Tracking URL: " + report.getTrackingUrl()); if (awaitCompletion) { Runtime.getRuntime().addShutdownHook(new Thread() { @Override public void run() { if (!yarnClient.isInState(Service.STATE.STOPPED)) { log.info("Killing yarn application in shutdown hook"); try { yarnClient.killApplication(appId); } catch (Throwable e) { log.error("Failed to kill yarn application - please check YARN Resource Manager", e); } } } }); float lastProgress = -0.0f; while (true) { try { Thread.sleep(10000); report = yarnClient.getApplicationReport(appId); if (lastProgress != report.getProgress()) { lastProgress = report.getProgress(); log.info(report.getApplicationId() + " " + (report.getProgress() * 100.00) + "% " + (System.currentTimeMillis() - report.getStartTime()) + "(ms) " + report.getDiagnostics()); } if (!report.getFinalApplicationStatus().equals(FinalApplicationStatus.UNDEFINED)) { log.info(report.getApplicationId() + " " + report.getFinalApplicationStatus()); log.info("Tracking url: " + report.getTrackingUrl()); log.info("Finish time: " + ((System.currentTimeMillis() - report.getStartTime()) / 1000) + "(s)"); break; } } catch (Throwable e) { log.error("Master Heart Beat Error - terminating", e); yarnClient.killApplication(appId); Thread.sleep(2000); } } yarnClient.stop(); if (!report.getFinalApplicationStatus().equals(FinalApplicationStatus.SUCCEEDED)) { System.exit(112); } } yarnClient.stop(); }
From source file:io.hops.hopsworks.common.jobs.flink.AbstractYarnClusterDescriptor.java
License:Apache License
private ClusterResourceDescription getCurrentFreeClusterResources(YarnClient yarnClient) throws YarnException, IOException { List<NodeReport> nodes = yarnClient.getNodeReports(NodeState.RUNNING); int totalFreeMemory = 0; int containerLimit = 0; int[] nodeManagersFree = new int[nodes.size()]; for (int i = 0; i < nodes.size(); i++) { NodeReport rep = nodes.get(i);//from www . ja va 2s . com int free = rep.getCapability().getMemory() - (rep.getUsed() != null ? rep.getUsed().getMemory() : 0); nodeManagersFree[i] = free; totalFreeMemory += free; if (free > containerLimit) { containerLimit = free; } } return new ClusterResourceDescription(totalFreeMemory, containerLimit, nodeManagersFree); }
From source file:io.hops.hopsworks.common.jobs.flink.AbstractYarnClusterDescriptor.java
License:Apache License
@Override public String getClusterDescription() { try {// w ww. ja v a2 s . c om ByteArrayOutputStream baos = new ByteArrayOutputStream(); PrintStream ps = new PrintStream(baos); YarnClient yarnClient = getYarnClient(); YarnClusterMetrics metrics = yarnClient.getYarnClusterMetrics(); ps.append("NodeManagers in the ClusterClient " + metrics.getNumNodeManagers()); List<NodeReport> nodes = yarnClient.getNodeReports(NodeState.RUNNING); final String format = "|%-16s |%-16s %n"; ps.printf("|Property |Value %n"); ps.println("+---------------------------------------+"); int totalMemory = 0; int totalCores = 0; for (NodeReport rep : nodes) { final Resource res = rep.getCapability(); totalMemory += res.getMemory(); totalCores += res.getVirtualCores(); ps.format(format, "NodeID", rep.getNodeId()); ps.format(format, "Memory", res.getMemory() + " MB"); ps.format(format, "vCores", res.getVirtualCores()); ps.format(format, "HealthReport", rep.getHealthReport()); ps.format(format, "Containers", rep.getNumContainers()); ps.println("+---------------------------------------+"); } ps.println("Summary: totalMemory " + totalMemory + " totalCores " + totalCores); List<QueueInfo> qInfo = yarnClient.getAllQueues(); for (QueueInfo q : qInfo) { ps.println("Queue: " + q.getQueueName() + ", Current Capacity: " + q.getCurrentCapacity() + " Max Capacity: " + q.getMaximumCapacity() + " Applications: " + q.getApplications().size()); } yarnClient.stop(); return baos.toString(); } catch (Exception e) { throw new RuntimeException("Couldn't get cluster description", e); } }
From source file:io.hops.tensorflow.Client.java
License:Apache License
private void logClusterState() throws IOException, YarnException { YarnClusterMetrics clusterMetrics = yarnClient.getYarnClusterMetrics(); LOG.info(// www . j a v a2 s.co m "Got Cluster metric info from ASM" + "\n\t 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" + "\n\t nodeId=" + node.getNodeId() + "\n\t nodeAddress" + node.getHttpAddress() + "\n\t nodeRackName" + node.getRackName() + "\n\t nodeNumContainers" + node.getNumContainers()); } QueueInfo queueInfo = yarnClient.getQueueInfo(this.amQueue); LOG.info("Queue info" + "\n\t queueName=" + queueInfo.getQueueName() + "\n\t queueCurrentCapacity=" + queueInfo.getCurrentCapacity() + "\n\t queueMaxCapacity=" + queueInfo.getMaximumCapacity() + "\n\t queueApplicationCount=" + queueInfo.getApplications().size() + "\n\t 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" + "\n\t queueName=" + aclInfo.getQueueName() + "\n\t userAcl=" + userAcl.name()); } } }
From source file:me.haosdent.noya.Client.java
License:Apache License
/** * Main run function for the client// w ww. j av a 2s . co m * * @return true if application completed successfully * * @throws java.io.IOException * @throws org.apache.hadoop.yarn.exceptions.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()); } } // 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 int maxMem = appResponse.getMaximumResourceCapability().getMemory(); LOG.info("Max mem capabililty 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 capabililty 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); // Set up the container launch context for the application master ContainerLaunchContext amContainer = Records.newRecord(ContainerLaunchContext.class); // 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 local resource info into app master container launch context amContainer.setLocalResources(localResources); // 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)); // 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(ApplicationConstants.Environment.CLASSPATH.$$()) .append(ApplicationConstants.CLASS_PATH_SEPARATOR).append("./*"); for (String c : conf.getStrings(YarnConfiguration.YARN_APPLICATION_CLASSPATH, ApplicationConstants.DEFAULT_YARN_CROSS_PLATFORM_APPLICATION_CLASSPATH)) { classPathEnv.append(ApplicationConstants.CLASS_PATH_SEPARATOR); classPathEnv.append(c.trim()); } classPathEnv.append(ApplicationConstants.CLASS_PATH_SEPARATOR).append("./log4j.properties"); // add the runtime classpath needed for tests to work if (conf.getBoolean(YarnConfiguration.IS_MINI_YARN_CLUSTER, false)) { classPathEnv.append(':'); classPathEnv.append(System.getProperty("java.class.path")); } env.put("CLASSPATH", classPathEnv.toString()); amContainer.setEnvironment(env); // Set the necessary command to execute the application master Vector<CharSequence> vargs = new Vector<CharSequence>(30); // Set java executable command LOG.info("Setting up app master command"); vargs.add(ApplicationConstants.Environment.JAVA_HOME.$$() + "/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)); 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.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()); commands.add("echo 'hello' >/tmp/yarn_test"); amContainer.setCommands(commands); // Set up resource type requirements // For now, both memory and vcores are supported, so we set memory and // vcores requirements Resource capability = Records.newRecord(Resource.class); capability.setMemory(amMemory); capability.setVirtualCores(amVCores); appContext.setResource(capability); // Service data is a binary blob that can be passed to the application // Not needed in this scenario // amContainer.setServiceData(serviceData); // Setup security tokens if (UserGroupInformation.isSecurityEnabled()) { Credentials credentials = new Credentials(); String tokenRenewer = conf.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 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 Priority pri = Records.newRecord(Priority.class); // TODO - what is the range for priority? how to decide? pri.setPriority(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); // TODO // Try submitting the same request again // app submission failure? // Monitor the application return monitorApplication(appId); }
From source file:org.apache.drill.yarn.appMaster.AMYarnFacadeImpl.java
License:Apache License
@Override public List<NodeReport> getNodeReports() throws YarnFacadeException { try {/*from w ww. j av a 2 s.c o m*/ return client.getNodeReports(NodeState.RUNNING); } catch (Exception e) { throw new YarnFacadeException("getNodeReports failed", e); } }
From source file:org.apache.flink.yarn.FlinkYarnClient.java
License:Apache License
public String getClusterDescription() throws Exception { ByteArrayOutputStream baos = new ByteArrayOutputStream(); PrintStream ps = new PrintStream(baos); YarnClusterMetrics metrics = yarnClient.getYarnClusterMetrics(); ps.append("NodeManagers in the Cluster " + metrics.getNumNodeManagers()); List<NodeReport> nodes = yarnClient.getNodeReports(NodeState.RUNNING); final String format = "|%-16s |%-16s %n"; ps.printf("|Property |Value %n"); ps.println("+---------------------------------------+"); int totalMemory = 0; int totalCores = 0; for (NodeReport rep : nodes) { final Resource res = rep.getCapability(); totalMemory += res.getMemory();/*www. jav a2 s.c o m*/ totalCores += res.getVirtualCores(); ps.format(format, "NodeID", rep.getNodeId()); ps.format(format, "Memory", res.getMemory() + " MB"); ps.format(format, "vCores", res.getVirtualCores()); ps.format(format, "HealthReport", rep.getHealthReport()); ps.format(format, "Containers", rep.getNumContainers()); ps.println("+---------------------------------------+"); } ps.println("Summary: totalMemory " + totalMemory + " totalCores " + totalCores); List<QueueInfo> qInfo = yarnClient.getAllQueues(); for (QueueInfo q : qInfo) { ps.println("Queue: " + q.getQueueName() + ", Current Capacity: " + q.getCurrentCapacity() + " Max Capacity: " + q.getMaximumCapacity() + " Applications: " + q.getApplications().size()); } yarnClient.stop(); return baos.toString(); }
From source file:org.apache.flink.yarn.FlinkYarnClientBase.java
License:Apache License
@Override public String getClusterDescription() throws Exception { ByteArrayOutputStream baos = new ByteArrayOutputStream(); PrintStream ps = new PrintStream(baos); YarnClusterMetrics metrics = yarnClient.getYarnClusterMetrics(); ps.append("NodeManagers in the Cluster " + metrics.getNumNodeManagers()); List<NodeReport> nodes = yarnClient.getNodeReports(NodeState.RUNNING); final String format = "|%-16s |%-16s %n"; ps.printf("|Property |Value %n"); ps.println("+---------------------------------------+"); int totalMemory = 0; int totalCores = 0; for (NodeReport rep : nodes) { final Resource res = rep.getCapability(); totalMemory += res.getMemory();//from ww w. j a va2 s .co m totalCores += res.getVirtualCores(); ps.format(format, "NodeID", rep.getNodeId()); ps.format(format, "Memory", res.getMemory() + " MB"); ps.format(format, "vCores", res.getVirtualCores()); ps.format(format, "HealthReport", rep.getHealthReport()); ps.format(format, "Containers", rep.getNumContainers()); ps.println("+---------------------------------------+"); } ps.println("Summary: totalMemory " + totalMemory + " totalCores " + totalCores); List<QueueInfo> qInfo = yarnClient.getAllQueues(); for (QueueInfo q : qInfo) { ps.println("Queue: " + q.getQueueName() + ", Current Capacity: " + q.getCurrentCapacity() + " Max Capacity: " + q.getMaximumCapacity() + " Applications: " + q.getApplications().size()); } yarnClient.stop(); return baos.toString(); }
From source file:org.apache.flink.yarn.YARNSessionCapacitySchedulerITCase.java
License:Apache License
/** * Test TaskManager failure and also if the vcores are set correctly (see issue FLINK-2213). *///from w w w . j ava2 s. c om @Test(timeout = 100000) // timeout after 100 seconds public void testTaskManagerFailure() { LOG.info("Starting testTaskManagerFailure()"); Runner runner = startWithArgs( new String[] { "-j", flinkUberjar.getAbsolutePath(), "-t", flinkLibFolder.getAbsolutePath(), "-n", "1", "-jm", "768", "-tm", "1024", "-s", "3", // set the slots 3 to check if the vCores are set properly! "-nm", "customName", "-Dfancy-configuration-value=veryFancy", "-Dyarn.maximum-failed-containers=3", "-D" + ConfigConstants.YARN_VCORES + "=2" }, "Number of connected TaskManagers changed to 1. Slots available: 3", RunTypes.YARN_SESSION); Assert.assertEquals(2, getRunningContainers()); // ------------------------ Test if JobManager web interface is accessible ------- YarnClient yc = null; try { yc = YarnClient.createYarnClient(); yc.init(yarnConfiguration); yc.start(); List<ApplicationReport> apps = yc.getApplications(EnumSet.of(YarnApplicationState.RUNNING)); Assert.assertEquals(1, apps.size()); // Only one running ApplicationReport app = apps.get(0); Assert.assertEquals("customName", app.getName()); String url = app.getTrackingUrl(); if (!url.endsWith("/")) { url += "/"; } if (!url.startsWith("http://")) { url = "http://" + url; } LOG.info("Got application URL from YARN {}", url); String response = TestBaseUtils.getFromHTTP(url + "taskmanagers/"); JsonNode parsedTMs = new ObjectMapper().readTree(response); ArrayNode taskManagers = (ArrayNode) parsedTMs.get("taskmanagers"); Assert.assertNotNull(taskManagers); Assert.assertEquals(1, taskManagers.size()); Assert.assertEquals(3, taskManagers.get(0).get("slotsNumber").asInt()); // get the configuration from webinterface & check if the dynamic properties from YARN show up there. String jsonConfig = TestBaseUtils.getFromHTTP(url + "jobmanager/config"); Map<String, String> parsedConfig = WebMonitorUtils.fromKeyValueJsonArray(jsonConfig); Assert.assertEquals("veryFancy", parsedConfig.get("fancy-configuration-value")); Assert.assertEquals("3", parsedConfig.get("yarn.maximum-failed-containers")); Assert.assertEquals("2", parsedConfig.get(ConfigConstants.YARN_VCORES)); // -------------- FLINK-1902: check if jobmanager hostname/port are shown in web interface // first, get the hostname/port String oC = outContent.toString(); Pattern p = Pattern.compile("Flink JobManager is now running on ([a-zA-Z0-9.-]+):([0-9]+)"); Matcher matches = p.matcher(oC); String hostname = null; String port = null; while (matches.find()) { hostname = matches.group(1).toLowerCase(); port = matches.group(2); } LOG.info("Extracted hostname:port: {} {}", hostname, port); Assert.assertEquals("unable to find hostname in " + jsonConfig, hostname, parsedConfig.get(ConfigConstants.JOB_MANAGER_IPC_ADDRESS_KEY)); Assert.assertEquals("unable to find port in " + jsonConfig, port, parsedConfig.get(ConfigConstants.JOB_MANAGER_IPC_PORT_KEY)); // test logfile access String logs = TestBaseUtils.getFromHTTP(url + "jobmanager/log"); Assert.assertTrue(logs.contains("Starting YARN ApplicationMaster")); Assert.assertTrue(logs.contains("Starting JobManager")); Assert.assertTrue(logs.contains("Starting JobManager Web Frontend")); } catch (Throwable e) { LOG.warn("Error while running test", e); Assert.fail(e.getMessage()); } // ------------------------ Kill container with TaskManager and check if vcores are set correctly ------- // find container id of taskManager: ContainerId taskManagerContainer = null; NodeManager nodeManager = null; UserGroupInformation remoteUgi = null; NMTokenIdentifier nmIdent = null; try { remoteUgi = UserGroupInformation.getCurrentUser(); } catch (IOException e) { LOG.warn("Unable to get curr user", e); Assert.fail(); } for (int nmId = 0; nmId < NUM_NODEMANAGERS; nmId++) { NodeManager nm = yarnCluster.getNodeManager(nmId); ConcurrentMap<ContainerId, Container> containers = nm.getNMContext().getContainers(); for (Map.Entry<ContainerId, Container> entry : containers.entrySet()) { String command = Joiner.on(" ").join(entry.getValue().getLaunchContext().getCommands()); if (command.contains(YarnTaskManager.class.getSimpleName())) { taskManagerContainer = entry.getKey(); nodeManager = nm; nmIdent = new NMTokenIdentifier(taskManagerContainer.getApplicationAttemptId(), null, "", 0); // allow myself to do stuff with the container // remoteUgi.addCredentials(entry.getValue().getCredentials()); remoteUgi.addTokenIdentifier(nmIdent); } } sleep(500); } Assert.assertNotNull("Unable to find container with TaskManager", taskManagerContainer); Assert.assertNotNull("Illegal state", nodeManager); try { List<NodeReport> nodeReports = yc.getNodeReports(NodeState.RUNNING); // we asked for one node with 2 vcores so we expect 2 vcores int userVcores = 0; for (NodeReport rep : nodeReports) { userVcores += rep.getUsed().getVirtualCores(); } Assert.assertEquals(2, userVcores); } catch (Exception e) { Assert.fail("Test failed: " + e.getMessage()); } yc.stop(); List<ContainerId> toStop = new LinkedList<ContainerId>(); toStop.add(taskManagerContainer); StopContainersRequest scr = StopContainersRequest.newInstance(toStop); try { nodeManager.getNMContext().getContainerManager().stopContainers(scr); } catch (Throwable e) { LOG.warn("Error stopping container", e); Assert.fail("Error stopping container: " + e.getMessage()); } // stateful termination check: // wait until we saw a container being killed and AFTERWARDS a new one launched boolean ok = false; do { LOG.debug("Waiting for correct order of events. Output: {}", errContent.toString()); String o = errContent.toString(); int killedOff = o.indexOf("Container killed by the ApplicationMaster"); if (killedOff != -1) { o = o.substring(killedOff); ok = o.indexOf("Launching TaskManager") > 0; } sleep(1000); } while (!ok); // send "stop" command to command line interface runner.sendStop(); // wait for the thread to stop try { runner.join(1000); } catch (InterruptedException e) { LOG.warn("Interrupted while stopping runner", e); } LOG.warn("stopped"); // ----------- Send output to logger System.setOut(originalStdout); System.setErr(originalStderr); String oC = outContent.toString(); String eC = errContent.toString(); LOG.info("Sending stdout content through logger: \n\n{}\n\n", oC); LOG.info("Sending stderr content through logger: \n\n{}\n\n", eC); // ------ Check if everything happened correctly Assert.assertTrue("Expect to see failed container", eC.contains("New messages from the YARN cluster")); Assert.assertTrue("Expect to see failed container", eC.contains("Container killed by the ApplicationMaster")); Assert.assertTrue("Expect to see new container started", eC.contains("Launching TaskManager") && eC.contains("on host")); // cleanup auth for the subsequent tests. remoteUgi.getTokenIdentifiers().remove(nmIdent); LOG.info("Finished testTaskManagerFailure()"); }
From source file:org.apache.giraph.yarn.GiraphYarnClient.java
License:Apache License
/** * Utility to make sure we have the cluster resources we need to run this * job. If they are not available, we should die here before too much setup. * @param cluster the GetNewApplicationResponse from the YARN RM. *///from w w w. j a va 2s. c o m private void checkPerNodeResourcesAvailable(final GetNewApplicationResponse cluster) throws YarnException, IOException { // are there enough containers to go around for our Giraph job? List<NodeReport> nodes = null; long totalAvailable = 0; try { nodes = yarnClient.getNodeReports(NodeState.RUNNING); } catch (YarnException yre) { throw new RuntimeException("GiraphYarnClient could not connect with " + "the YARN ResourceManager to determine the number of available " + "application containers.", yre); } for (NodeReport node : nodes) { LOG.info("Got node report from ASM for" + ", nodeId=" + node.getNodeId() + ", nodeAddress " + node.getHttpAddress() + ", nodeRackName " + node.getRackName() + ", nodeNumContainers " + node.getNumContainers()); totalAvailable += node.getCapability().getMemory(); } // 1 master + all workers in -w command line arg final int workers = giraphConf.getMaxWorkers() + 1; checkAndAdjustPerTaskHeapSize(cluster); final long totalAsk = giraphConf.getYarnTaskHeapMb() * workers; if (totalAsk > totalAvailable) { throw new IllegalStateException("Giraph's estimated cluster heap " + totalAsk + "MB ask is greater than the current available cluster " + "heap of " + totalAvailable + "MB. Aborting Job."); } }