use of org.apache.hadoop.yarn.api.records.NodeReport in project hadoop by apache.
the class BaseAMRMProxyE2ETest method createAllocateRequest.
protected AllocateRequest createAllocateRequest(List<NodeReport> listNode) {
// The test needs AMRMClient to create a real allocate request
AMRMClientImpl<AMRMClient.ContainerRequest> amClient = new AMRMClientImpl<>();
Resource capability = Resource.newInstance(1024, 2);
Priority priority = Priority.newInstance(1);
List<NodeReport> nodeReports = listNode;
String node = nodeReports.get(0).getNodeId().getHost();
String[] nodes = new String[] { node };
AMRMClient.ContainerRequest storedContainer1 = new AMRMClient.ContainerRequest(capability, nodes, null, priority);
amClient.addContainerRequest(storedContainer1);
amClient.addContainerRequest(storedContainer1);
List<ResourceRequest> resourceAsk = new ArrayList<>();
for (ResourceRequest rr : amClient.ask) {
resourceAsk.add(rr);
}
ResourceBlacklistRequest resourceBlacklistRequest = ResourceBlacklistRequest.newInstance(new ArrayList<>(), new ArrayList<>());
int responseId = 1;
return AllocateRequest.newInstance(responseId, 0, resourceAsk, new ArrayList<>(), resourceBlacklistRequest);
}
use of org.apache.hadoop.yarn.api.records.NodeReport in project hadoop by apache.
the class TestDistributedScheduling method testAMOpportunistic.
/**
* Check if an AM can ask for opportunistic containers and get them.
* @throws Exception
*/
@Test
public void testAMOpportunistic() throws Exception {
// Basic container to request
Resource capability = Resource.newInstance(1024, 1);
Priority priority = Priority.newInstance(1);
// Get the cluster topology
List<NodeReport> nodeReports = rmClient.getNodeReports(NodeState.RUNNING);
String node = nodeReports.get(0).getNodeId().getHost();
String rack = nodeReports.get(0).getRackName();
String[] nodes = new String[] { node };
String[] racks = new String[] { rack };
// Create an AM to request resources
AMRMClient<AMRMClient.ContainerRequest> amClient = null;
try {
amClient = new AMRMClientImpl<AMRMClient.ContainerRequest>(client);
amClient.init(yarnConf);
amClient.start();
amClient.registerApplicationMaster(NetUtils.getHostname(), 1024, "");
// AM requests an opportunistic container
ExecutionTypeRequest execTypeRequest = ExecutionTypeRequest.newInstance(ExecutionType.OPPORTUNISTIC, true);
ContainerRequest containerRequest = new AMRMClient.ContainerRequest(capability, nodes, racks, priority, 0, true, null, execTypeRequest);
amClient.addContainerRequest(containerRequest);
// Wait until the container is allocated
ContainerId opportunisticContainerId = null;
for (int i = 0; i < 10 && opportunisticContainerId == null; i++) {
AllocateResponse allocResponse = amClient.allocate(0.1f);
List<Container> allocatedContainers = allocResponse.getAllocatedContainers();
for (Container allocatedContainer : allocatedContainers) {
// Check that this is the container we required
assertEquals(ExecutionType.OPPORTUNISTIC, allocatedContainer.getExecutionType());
opportunisticContainerId = allocatedContainer.getId();
}
sleep(100);
}
assertNotNull(opportunisticContainerId);
// The RM sees the container as OPPORTUNISTIC
ResourceScheduler scheduler = cluster.getResourceManager().getResourceScheduler();
RMContainer rmContainer = scheduler.getRMContainer(opportunisticContainerId);
assertEquals(ExecutionType.OPPORTUNISTIC, rmContainer.getExecutionType());
// Release the opportunistic container
amClient.releaseAssignedContainer(opportunisticContainerId);
// Wait for the release container to appear
boolean released = false;
for (int i = 0; i < 10 && !released; i++) {
AllocateResponse allocResponse = amClient.allocate(0.1f);
List<ContainerStatus> completedContainers = allocResponse.getCompletedContainersStatuses();
for (ContainerStatus completedContainer : completedContainers) {
ContainerId completedContainerId = completedContainer.getContainerId();
assertEquals(completedContainerId, opportunisticContainerId);
released = true;
}
if (!released) {
sleep(100);
}
}
assertTrue(released);
// The RM shouldn't see the container anymore
rmContainer = scheduler.getRMContainer(opportunisticContainerId);
assertNull(rmContainer);
// Clean the AM
amClient.unregisterApplicationMaster(FinalApplicationStatus.SUCCEEDED, null, null);
} finally {
if (amClient != null && amClient.getServiceState() == Service.STATE.STARTED) {
amClient.close();
}
}
}
use of org.apache.hadoop.yarn.api.records.NodeReport in project hadoop by apache.
the class TestDistributedScheduling method testAMRMClient.
/**
* Validates if AMRMClient can be used with Distributed Scheduling turned on.
*
* @throws Exception
*/
@Test(timeout = 120000)
@SuppressWarnings("unchecked")
public void testAMRMClient() throws Exception {
AMRMClientImpl<AMRMClient.ContainerRequest> amClient = null;
try {
Priority priority = Priority.newInstance(1);
Priority priority2 = Priority.newInstance(2);
Resource capability = Resource.newInstance(1024, 1);
List<NodeReport> nodeReports = rmClient.getNodeReports(NodeState.RUNNING);
String node = nodeReports.get(0).getNodeId().getHost();
String rack = nodeReports.get(0).getRackName();
String[] nodes = new String[] { node };
String[] racks = new String[] { rack };
// start am rm client
amClient = new AMRMClientImpl(client);
amClient.init(yarnConf);
amClient.start();
amClient.registerApplicationMaster(NetUtils.getHostname(), 1024, "");
assertEquals(0, amClient.ask.size());
assertEquals(0, amClient.release.size());
amClient.addContainerRequest(new AMRMClient.ContainerRequest(capability, nodes, racks, priority));
amClient.addContainerRequest(new AMRMClient.ContainerRequest(capability, nodes, racks, priority));
amClient.addContainerRequest(new AMRMClient.ContainerRequest(capability, nodes, racks, priority));
amClient.addContainerRequest(new AMRMClient.ContainerRequest(capability, nodes, racks, priority));
amClient.addContainerRequest(new AMRMClient.ContainerRequest(capability, null, null, priority2, 0, true, null, ExecutionTypeRequest.newInstance(ExecutionType.OPPORTUNISTIC, true)));
amClient.addContainerRequest(new AMRMClient.ContainerRequest(capability, null, null, priority2, 0, true, null, ExecutionTypeRequest.newInstance(ExecutionType.OPPORTUNISTIC, true)));
amClient.removeContainerRequest(new AMRMClient.ContainerRequest(capability, nodes, racks, priority));
amClient.removeContainerRequest(new AMRMClient.ContainerRequest(capability, nodes, racks, priority));
amClient.removeContainerRequest(new AMRMClient.ContainerRequest(capability, null, null, priority2, 0, true, null, ExecutionTypeRequest.newInstance(ExecutionType.OPPORTUNISTIC, true)));
RemoteRequestsTable<ContainerRequest> remoteRequestsTable = amClient.getTable(0);
int containersRequestedNode = remoteRequestsTable.get(priority, node, ExecutionType.GUARANTEED, capability).remoteRequest.getNumContainers();
int containersRequestedRack = remoteRequestsTable.get(priority, rack, ExecutionType.GUARANTEED, capability).remoteRequest.getNumContainers();
int containersRequestedAny = remoteRequestsTable.get(priority, ResourceRequest.ANY, ExecutionType.GUARANTEED, capability).remoteRequest.getNumContainers();
int oppContainersRequestedAny = remoteRequestsTable.get(priority2, ResourceRequest.ANY, ExecutionType.OPPORTUNISTIC, capability).remoteRequest.getNumContainers();
assertEquals(2, containersRequestedNode);
assertEquals(2, containersRequestedRack);
assertEquals(2, containersRequestedAny);
assertEquals(1, oppContainersRequestedAny);
assertEquals(4, amClient.ask.size());
assertEquals(0, amClient.release.size());
// RM should allocate container within 2 calls to allocate()
int allocatedContainerCount = 0;
int iterationsLeft = 10;
Set<ContainerId> releases = new TreeSet<>();
amClient.getNMTokenCache().clearCache();
Assert.assertEquals(0, amClient.getNMTokenCache().numberOfTokensInCache());
HashMap<String, Token> receivedNMTokens = new HashMap<>();
while (allocatedContainerCount < (containersRequestedAny + oppContainersRequestedAny) && iterationsLeft-- > 0) {
AllocateResponse allocResponse = amClient.allocate(0.1f);
assertEquals(0, amClient.ask.size());
assertEquals(0, amClient.release.size());
allocatedContainerCount += allocResponse.getAllocatedContainers().size();
for (Container container : allocResponse.getAllocatedContainers()) {
ContainerId rejectContainerId = container.getId();
releases.add(rejectContainerId);
}
for (NMToken token : allocResponse.getNMTokens()) {
String nodeID = token.getNodeId().toString();
receivedNMTokens.put(nodeID, token.getToken());
}
if (allocatedContainerCount < containersRequestedAny) {
// sleep to let NM's heartbeat to RM and trigger allocations
sleep(100);
}
}
assertEquals(allocatedContainerCount, containersRequestedAny + oppContainersRequestedAny);
for (ContainerId rejectContainerId : releases) {
amClient.releaseAssignedContainer(rejectContainerId);
}
assertEquals(3, amClient.release.size());
assertEquals(0, amClient.ask.size());
// need to tell the AMRMClient that we dont need these resources anymore
amClient.removeContainerRequest(new AMRMClient.ContainerRequest(capability, nodes, racks, priority));
amClient.removeContainerRequest(new AMRMClient.ContainerRequest(capability, nodes, racks, priority));
amClient.removeContainerRequest(new AMRMClient.ContainerRequest(capability, nodes, racks, priority2, 0, true, null, ExecutionTypeRequest.newInstance(ExecutionType.OPPORTUNISTIC, true)));
assertEquals(4, amClient.ask.size());
// test RPC exception handling
amClient.addContainerRequest(new AMRMClient.ContainerRequest(capability, nodes, racks, priority));
amClient.addContainerRequest(new AMRMClient.ContainerRequest(capability, nodes, racks, priority));
amClient.addContainerRequest(new AMRMClient.ContainerRequest(capability, nodes, racks, priority2, 0, true, null, ExecutionTypeRequest.newInstance(ExecutionType.OPPORTUNISTIC, true)));
final AMRMClient amc = amClient;
ApplicationMasterProtocol realRM = amClient.rmClient;
try {
ApplicationMasterProtocol mockRM = mock(ApplicationMasterProtocol.class);
when(mockRM.allocate(any(AllocateRequest.class))).thenAnswer(new Answer<AllocateResponse>() {
public AllocateResponse answer(InvocationOnMock invocation) throws Exception {
amc.removeContainerRequest(new AMRMClient.ContainerRequest(capability, nodes, racks, priority));
amc.removeContainerRequest(new AMRMClient.ContainerRequest(capability, nodes, racks, priority));
amc.removeContainerRequest(new AMRMClient.ContainerRequest(capability, null, null, priority2, 0, true, null, ExecutionTypeRequest.newInstance(ExecutionType.OPPORTUNISTIC, true)));
throw new Exception();
}
});
amClient.rmClient = mockRM;
amClient.allocate(0.1f);
} catch (Exception ioe) {
} finally {
amClient.rmClient = realRM;
}
assertEquals(3, amClient.release.size());
assertEquals(6, amClient.ask.size());
iterationsLeft = 3;
// do a few iterations to ensure RM is not going send new containers
while (iterationsLeft-- > 0) {
// inform RM of rejection
AllocateResponse allocResponse = amClient.allocate(0.1f);
// RM did not send new containers because AM does not need any
assertEquals(0, allocResponse.getAllocatedContainers().size());
if (allocResponse.getCompletedContainersStatuses().size() > 0) {
for (ContainerStatus cStatus : allocResponse.getCompletedContainersStatuses()) {
if (releases.contains(cStatus.getContainerId())) {
assertEquals(cStatus.getState(), ContainerState.COMPLETE);
assertEquals(-100, cStatus.getExitStatus());
releases.remove(cStatus.getContainerId());
}
}
}
if (iterationsLeft > 0) {
// sleep to make sure NM's heartbeat
sleep(100);
}
}
assertEquals(0, amClient.ask.size());
assertEquals(0, amClient.release.size());
amClient.unregisterApplicationMaster(FinalApplicationStatus.SUCCEEDED, null, null);
} finally {
if (amClient != null && amClient.getServiceState() == Service.STATE.STARTED) {
amClient.stop();
}
}
}
use of org.apache.hadoop.yarn.api.records.NodeReport in project hadoop by apache.
the class ClientRMService method getClusterNodes.
@Override
public GetClusterNodesResponse getClusterNodes(GetClusterNodesRequest request) throws YarnException {
GetClusterNodesResponse response = recordFactory.newRecordInstance(GetClusterNodesResponse.class);
EnumSet<NodeState> nodeStates = request.getNodeStates();
if (nodeStates == null || nodeStates.isEmpty()) {
nodeStates = EnumSet.allOf(NodeState.class);
}
Collection<RMNode> nodes = RMServerUtils.queryRMNodes(rmContext, nodeStates);
List<NodeReport> nodeReports = new ArrayList<NodeReport>(nodes.size());
for (RMNode nodeInfo : nodes) {
nodeReports.add(createNodeReports(nodeInfo));
}
response.setNodeReports(nodeReports);
return response;
}
use of org.apache.hadoop.yarn.api.records.NodeReport 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);
}
Aggregations