当前位置: 首页>>代码示例>>Java>>正文


Java ApplicationStateData.newInstance方法代码示例

本文整理汇总了Java中org.apache.hadoop.yarn.server.resourcemanager.recovery.records.ApplicationStateData.newInstance方法的典型用法代码示例。如果您正苦于以下问题:Java ApplicationStateData.newInstance方法的具体用法?Java ApplicationStateData.newInstance怎么用?Java ApplicationStateData.newInstance使用的例子?那么恭喜您, 这里精选的方法代码示例或许可以为您提供帮助。您也可以进一步了解该方法所在org.apache.hadoop.yarn.server.resourcemanager.recovery.records.ApplicationStateData的用法示例。


在下文中一共展示了ApplicationStateData.newInstance方法的11个代码示例,这些例子默认根据受欢迎程度排序。您可以为喜欢或者感觉有用的代码点赞,您的评价将有助于系统推荐出更棒的Java代码示例。

示例1: testCreateAppSubmittedRecovery

import org.apache.hadoop.yarn.server.resourcemanager.recovery.records.ApplicationStateData; //导入方法依赖的package包/类
protected RMApp testCreateAppSubmittedRecovery(
    ApplicationSubmissionContext submissionContext) throws IOException {
  RMApp application = createNewTestApp(submissionContext);
  // NEW => SUBMITTED event RMAppEventType.RECOVER
  RMState state = new RMState();
  ApplicationStateData appState =
      ApplicationStateData.newInstance(123, 123, null, "user");
  state.getApplicationState().put(application.getApplicationId(), appState);
  RMAppEvent event =
      new RMAppRecoverEvent(application.getApplicationId(), state);


  application.handle(event);
  assertStartTimeSet(application);
  assertAppState(RMAppState.SUBMITTED, application);
  return application;
}
 
开发者ID:naver,项目名称:hadoop,代码行数:18,代码来源:TestRMAppTransitions.java

示例2: testCreateAppSubmittedRecovery

import org.apache.hadoop.yarn.server.resourcemanager.recovery.records.ApplicationStateData; //导入方法依赖的package包/类
protected RMApp testCreateAppSubmittedRecovery(
    ApplicationSubmissionContext submissionContext) throws IOException {
  RMApp application = createNewTestApp(submissionContext);
  // NEW => SUBMITTED event RMAppEventType.RECOVER
  RMState state = new RMState();
  ApplicationStateData appState =
      ApplicationStateData.newInstance(123, 123, null, "user", null);
  state.getApplicationState().put(application.getApplicationId(), appState);
  RMAppEvent event =
      new RMAppRecoverEvent(application.getApplicationId(), state);


  application.handle(event);
  assertStartTimeSet(application);
  assertAppState(RMAppState.SUBMITTED, application);
  return application;
}
 
开发者ID:aliyun-beta,项目名称:aliyun-oss-hadoop-fs,代码行数:18,代码来源:TestRMAppTransitions.java

示例3: rememberTargetTransitionsAndStoreState

import org.apache.hadoop.yarn.server.resourcemanager.recovery.records.ApplicationStateData; //导入方法依赖的package包/类
private void rememberTargetTransitionsAndStoreState(RMAppEvent event,
    Object transitionToDo, RMAppState targetFinalState,
    RMAppState stateToBeStored) {
  rememberTargetTransitions(event, transitionToDo, targetFinalState);
  this.stateBeforeFinalSaving = getState();
  this.storedFinishTime = this.systemClock.getTime();

  LOG.info("Updating application " + this.applicationId
      + " with final state: " + this.targetedFinalState);
  // we lost attempt_finished diagnostics in app, because attempt_finished
  // diagnostics is sent after app final state is saved. Later on, we will
  // create GetApplicationAttemptReport specifically for getting per attempt
  // info.
  String diags = null;
  switch (event.getType()) {
  case APP_REJECTED:
    RMAppRejectedEvent rejectedEvent = (RMAppRejectedEvent) event;
    diags = rejectedEvent.getMessage();
    break;
  case ATTEMPT_FINISHED:
    RMAppFinishedAttemptEvent finishedEvent =
        (RMAppFinishedAttemptEvent) event;
    diags = finishedEvent.getDiagnostics();
    break;
  case ATTEMPT_FAILED:
    RMAppFailedAttemptEvent failedEvent = (RMAppFailedAttemptEvent) event;
    diags = getAppAttemptFailedDiagnostics(failedEvent);
    break;
  case ATTEMPT_KILLED:
    diags = getAppKilledDiagnostics();
    break;
  default:
    break;
  }
  ApplicationStateData appState =
      ApplicationStateData.newInstance(this.submitTime, this.startTime,
          this.user, this.submissionContext,
          stateToBeStored, diags, this.storedFinishTime);
  this.rmContext.getStateStore().updateApplicationState(appState);
}
 
开发者ID:naver,项目名称:hadoop,代码行数:41,代码来源:RMAppImpl.java

示例4: removeApplication

import org.apache.hadoop.yarn.server.resourcemanager.recovery.records.ApplicationStateData; //导入方法依赖的package包/类
/**
 * Non-blocking API
 * ResourceManager services call this to remove an application from the state
 * store
 * This does not block the dispatcher threads
 * There is no notification of completion for this operation.
 */
@SuppressWarnings("unchecked")
public synchronized void removeApplication(RMApp app) {
  ApplicationStateData appState =
      ApplicationStateData.newInstance(
          app.getSubmitTime(), app.getStartTime(),
          app.getApplicationSubmissionContext(), app.getUser());
  for(RMAppAttempt appAttempt : app.getAppAttempts().values()) {
    appState.attempts.put(appAttempt.getAppAttemptId(), null);
  }
  
  dispatcher.getEventHandler().handle(new RMStateStoreRemoveAppEvent(appState));
}
 
开发者ID:naver,项目名称:hadoop,代码行数:20,代码来源:RMStateStore.java

示例5: createRMStateForApplications

import org.apache.hadoop.yarn.server.resourcemanager.recovery.records.ApplicationStateData; //导入方法依赖的package包/类
public void createRMStateForApplications(
    Map<ApplicationId, ApplicationStateData> applicationState,
    RMAppState rmAppState) {
  RMApp app = createNewTestApp(null);
  ApplicationStateData appState =
      ApplicationStateData.newInstance(app.getSubmitTime(), app.getStartTime(),
          app.getUser(), app.getApplicationSubmissionContext(), rmAppState,
          null, app.getFinishTime());
  applicationState.put(app.getApplicationId(), appState);
}
 
开发者ID:naver,项目名称:hadoop,代码行数:11,代码来源:TestRMAppTransitions.java

示例6: testDuplicateRMAppDeletion

import org.apache.hadoop.yarn.server.resourcemanager.recovery.records.ApplicationStateData; //导入方法依赖的package包/类
@Test
public void testDuplicateRMAppDeletion() throws Exception {
  TestZKRMStateStoreTester zkTester = new TestZKRMStateStoreTester();
  long submitTime = System.currentTimeMillis();
  long startTime = System.currentTimeMillis() + 1234;
  RMStateStore store = zkTester.getRMStateStore();
  TestDispatcher dispatcher = new TestDispatcher();
  store.setRMDispatcher(dispatcher);

  ApplicationAttemptId attemptIdRemoved = ConverterUtils
      .toApplicationAttemptId("appattempt_1352994193343_0002_000001");
  ApplicationId appIdRemoved = attemptIdRemoved.getApplicationId();
  storeApp(store, appIdRemoved, submitTime, startTime);
  storeAttempt(store, attemptIdRemoved,
      "container_1352994193343_0002_01_000001", null, null, dispatcher);

  ApplicationSubmissionContext context =
      new ApplicationSubmissionContextPBImpl();
  context.setApplicationId(appIdRemoved);
  ApplicationStateData appStateRemoved =
      ApplicationStateData.newInstance(
          submitTime, startTime, context, "user1");
  appStateRemoved.attempts.put(attemptIdRemoved, null);
  store.removeApplicationStateInternal(appStateRemoved);
  try {
    store.removeApplicationStateInternal(appStateRemoved);
  } catch (KeeperException.NoNodeException nne) {
    Assert.fail("NoNodeException should not happen.");
  }
  store.close();
}
 
开发者ID:naver,项目名称:hadoop,代码行数:32,代码来源:TestZKRMStateStore.java

示例7: rememberTargetTransitionsAndStoreState

import org.apache.hadoop.yarn.server.resourcemanager.recovery.records.ApplicationStateData; //导入方法依赖的package包/类
private void rememberTargetTransitionsAndStoreState(RMAppEvent event,
    Object transitionToDo, RMAppState targetFinalState,
    RMAppState stateToBeStored) {
  rememberTargetTransitions(event, transitionToDo, targetFinalState);
  this.stateBeforeFinalSaving = getState();
  this.storedFinishTime = this.systemClock.getTime();

  LOG.info("Updating application " + this.applicationId
      + " with final state: " + this.targetedFinalState);
  // we lost attempt_finished diagnostics in app, because attempt_finished
  // diagnostics is sent after app final state is saved. Later on, we will
  // create GetApplicationAttemptReport specifically for getting per attempt
  // info.
  String diags = null;
  switch (event.getType()) {
  case APP_REJECTED:
  case ATTEMPT_FINISHED:
  case ATTEMPT_KILLED:
    diags = event.getDiagnosticMsg();
    break;
  case ATTEMPT_FAILED:
    RMAppFailedAttemptEvent failedEvent = (RMAppFailedAttemptEvent) event;
    diags = getAppAttemptFailedDiagnostics(failedEvent);
    break;
  default:
    break;
  }

  ApplicationStateData appState =
      ApplicationStateData.newInstance(this.submitTime, this.startTime,
          this.user, this.submissionContext,
          stateToBeStored, diags, this.storedFinishTime, this.callerContext);
  this.rmContext.getStateStore().updateApplicationState(appState);
}
 
开发者ID:aliyun-beta,项目名称:aliyun-oss-hadoop-fs,代码行数:35,代码来源:RMAppImpl.java

示例8: removeApplication

import org.apache.hadoop.yarn.server.resourcemanager.recovery.records.ApplicationStateData; //导入方法依赖的package包/类
/**
 * Non-blocking API
 * ResourceManager services call this to remove an application from the state
 * store
 * This does not block the dispatcher threads
 * There is no notification of completion for this operation.
 */
@SuppressWarnings("unchecked")
public void removeApplication(RMApp app) {
  ApplicationStateData appState =
      ApplicationStateData.newInstance(app.getSubmitTime(),
          app.getStartTime(), app.getApplicationSubmissionContext(),
          app.getUser(), app.getCallerContext());
  for(RMAppAttempt appAttempt : app.getAppAttempts().values()) {
    appState.attempts.put(appAttempt.getAppAttemptId(), null);
  }
  
  dispatcher.getEventHandler().handle(new RMStateStoreRemoveAppEvent(appState));
}
 
开发者ID:aliyun-beta,项目名称:aliyun-oss-hadoop-fs,代码行数:20,代码来源:RMStateStore.java

示例9: createRMStateForApplications

import org.apache.hadoop.yarn.server.resourcemanager.recovery.records.ApplicationStateData; //导入方法依赖的package包/类
public void createRMStateForApplications(
    Map<ApplicationId, ApplicationStateData> applicationState,
    RMAppState rmAppState) {
  RMApp app = createNewTestApp(null);
  ApplicationStateData appState =
      ApplicationStateData.newInstance(app.getSubmitTime(), app.getStartTime(),
          app.getUser(), app.getApplicationSubmissionContext(), rmAppState,
          null, app.getFinishTime(), null);
  applicationState.put(app.getApplicationId(), appState);
}
 
开发者ID:aliyun-beta,项目名称:aliyun-oss-hadoop-fs,代码行数:11,代码来源:TestRMAppTransitions.java

示例10: testDuplicateRMAppDeletion

import org.apache.hadoop.yarn.server.resourcemanager.recovery.records.ApplicationStateData; //导入方法依赖的package包/类
@Test
public void testDuplicateRMAppDeletion() throws Exception {
  TestZKRMStateStoreTester zkTester = new TestZKRMStateStoreTester();
  long submitTime = System.currentTimeMillis();
  long startTime = System.currentTimeMillis() + 1234;
  RMStateStore store = zkTester.getRMStateStore();
  TestDispatcher dispatcher = new TestDispatcher();
  store.setRMDispatcher(dispatcher);

  ApplicationAttemptId attemptIdRemoved = ConverterUtils
      .toApplicationAttemptId("appattempt_1352994193343_0002_000001");
  ApplicationId appIdRemoved = attemptIdRemoved.getApplicationId();
  storeApp(store, appIdRemoved, submitTime, startTime);
  storeAttempt(store, attemptIdRemoved,
      "container_1352994193343_0002_01_000001", null, null, dispatcher);

  ApplicationSubmissionContext context =
      new ApplicationSubmissionContextPBImpl();
  context.setApplicationId(appIdRemoved);

  ApplicationStateData appStateRemoved =
      ApplicationStateData.newInstance(
          submitTime, startTime, context, "user1");
  appStateRemoved.attempts.put(attemptIdRemoved, null);
  store.removeApplicationStateInternal(appStateRemoved);
  try {
    store.removeApplicationStateInternal(appStateRemoved);
  } catch (KeeperException.NoNodeException nne) {
    Assert.fail("NoNodeException should not happen.");
  }
  store.close();
}
 
开发者ID:aliyun-beta,项目名称:aliyun-oss-hadoop-fs,代码行数:33,代码来源:TestZKRMStateStore.java

示例11: updateApplicationPriority

import org.apache.hadoop.yarn.server.resourcemanager.recovery.records.ApplicationStateData; //导入方法依赖的package包/类
@Override
public void updateApplicationPriority(Priority newPriority,
    ApplicationId applicationId) throws YarnException {
  Priority appPriority = null;
  SchedulerApplication<FiCaSchedulerApp> application = applications
      .get(applicationId);

  if (application == null) {
    throw new YarnException("Application '" + applicationId
        + "' is not present, hence could not change priority.");
  }

  RMApp rmApp = rmContext.getRMApps().get(applicationId);
  appPriority = checkAndGetApplicationPriority(newPriority, rmApp.getUser(),
      rmApp.getQueue(), applicationId);

  if (application.getPriority().equals(appPriority)) {
    return;
  }

  // Update new priority in Submission Context to keep track in HA
  rmApp.getApplicationSubmissionContext().setPriority(appPriority);

  // Update to state store
  ApplicationStateData appState =
      ApplicationStateData.newInstance(rmApp.getSubmitTime(),
          rmApp.getStartTime(), rmApp.getApplicationSubmissionContext(),
          rmApp.getUser(), rmApp.getCallerContext());
  rmContext.getStateStore().updateApplicationStateSynchronously(appState,
      false);

  // As we use iterator over a TreeSet for OrderingPolicy, once we change
  // priority then reinsert back to make order correct.
  LeafQueue queue = (LeafQueue) getQueue(rmApp.getQueue());
  synchronized (queue) {
    queue.getOrderingPolicy().removeSchedulableEntity(
        application.getCurrentAppAttempt());

    // Update new priority in SchedulerApplication
    application.setPriority(appPriority);

    queue.getOrderingPolicy().addSchedulableEntity(
        application.getCurrentAppAttempt());
  }

  // Update the changed application state to timeline server
  rmContext.getSystemMetricsPublisher().appUpdated(rmApp,
      System.currentTimeMillis());

  LOG.info("Priority '" + appPriority + "' is updated in queue :"
      + rmApp.getQueue() + " for application: " + applicationId
      + " for the user: " + rmApp.getUser());
}
 
开发者ID:aliyun-beta,项目名称:aliyun-oss-hadoop-fs,代码行数:54,代码来源:CapacityScheduler.java


注:本文中的org.apache.hadoop.yarn.server.resourcemanager.recovery.records.ApplicationStateData.newInstance方法示例由纯净天空整理自Github/MSDocs等开源代码及文档管理平台,相关代码片段筛选自各路编程大神贡献的开源项目,源码版权归原作者所有,传播和使用请参考对应项目的License;未经允许,请勿转载。