本文整理汇总了Java中org.apache.hadoop.mapreduce.v2.app.commit.CommitterJobAbortEvent类的典型用法代码示例。如果您正苦于以下问题：Java CommitterJobAbortEvent类的具体用法？Java CommitterJobAbortEvent怎么用？Java CommitterJobAbortEvent使用的例子？那么恭喜您, 这里精选的类代码示例或许可以为您提供帮助。
CommitterJobAbortEvent类属于org.apache.hadoop.mapreduce.v2.app.commit包，在下文中一共展示了CommitterJobAbortEvent类的3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: checkJobAfterTaskCompletion
import org.apache.hadoop.mapreduce.v2.app.commit.CommitterJobAbortEvent; //导入依赖的package包/类
protected JobStateInternal checkJobAfterTaskCompletion(JobImpl job) {
//check for Job failure
if (job.failedMapTaskCount*100 >
job.allowedMapFailuresPercent*job.numMapTasks ||
job.failedReduceTaskCount*100 >
job.allowedReduceFailuresPercent*job.numReduceTasks) {
job.setFinishTime();
String diagnosticMsg = "Job failed as tasks failed. " +
"failedMaps:" + job.failedMapTaskCount +
" failedReduces:" + job.failedReduceTaskCount;
LOG.info(diagnosticMsg);
job.addDiagnostic(diagnosticMsg);
job.eventHandler.handle(new CommitterJobAbortEvent(job.jobId,
job.jobContext,
org.apache.hadoop.mapreduce.JobStatus.State.FAILED));
return JobStateInternal.FAIL_ABORT;
}
return job.checkReadyForCommit();
}
示例2: transition
import org.apache.hadoop.mapreduce.v2.app.commit.CommitterJobAbortEvent; //导入依赖的package包/类
@Override
public void transition(JobImpl job, JobEvent event) {
job.metrics.endRunningJob(job);
job.addDiagnostic("Job setup failed : "
+ ((JobSetupFailedEvent) event).getMessage());
job.eventHandler.handle(new CommitterJobAbortEvent(job.jobId,
job.jobContext,
org.apache.hadoop.mapreduce.JobStatus.State.FAILED));
}
示例3: checkJobAfterTaskCompletion
import org.apache.hadoop.mapreduce.v2.app.commit.CommitterJobAbortEvent; //导入依赖的package包/类
@Override
protected JobStateInternal checkJobAfterTaskCompletion(JobImpl job) {
if (job.completedTaskCount == job.tasks.size()) {
job.setFinishTime();
job.eventHandler.handle(new CommitterJobAbortEvent(job.jobId,
job.jobContext,
org.apache.hadoop.mapreduce.JobStatus.State.KILLED));
return JobStateInternal.KILL_ABORT;
}
//return the current state, Job not finished yet
return job.getInternalState();
}