Search in sources :

Example 6 with TaskId

use of org.apache.hadoop.mapreduce.v2.api.records.TaskId in project hadoop by apache.

the class TestMRApps method testTaskIDtoString.

//TODO_get.set
@Test(timeout = 120000)
public void testTaskIDtoString() {
    TaskId tid = RecordFactoryProvider.getRecordFactory(null).newRecordInstance(TaskId.class);
    tid.setJobId(RecordFactoryProvider.getRecordFactory(null).newRecordInstance(JobId.class));
    tid.getJobId().setAppId(ApplicationId.newInstance(0, 0));
    tid.setTaskType(TaskType.MAP);
    TaskType type = tid.getTaskType();
    System.err.println(type);
    type = TaskType.REDUCE;
    System.err.println(type);
    System.err.println(tid.getTaskType());
    assertEquals("task_0_0000_m_000000", MRApps.toString(tid));
    tid.setTaskType(TaskType.REDUCE);
    assertEquals("task_0_0000_r_000000", MRApps.toString(tid));
}
Also used : TaskId(org.apache.hadoop.mapreduce.v2.api.records.TaskId) TaskType(org.apache.hadoop.mapreduce.v2.api.records.TaskType) JobId(org.apache.hadoop.mapreduce.v2.api.records.JobId) Test(org.junit.Test)

Example 7 with TaskId

use of org.apache.hadoop.mapreduce.v2.api.records.TaskId in project hadoop by apache.

the class TestJobHistoryParsing method checkHistoryParsing.

private void checkHistoryParsing(final int numMaps, final int numReduces, final int numSuccessfulMaps) throws Exception {
    Configuration conf = new Configuration();
    conf.set(MRJobConfig.USER_NAME, System.getProperty("user.name"));
    long amStartTimeEst = System.currentTimeMillis();
    conf.setClass(NET_TOPOLOGY_NODE_SWITCH_MAPPING_IMPL_KEY, MyResolver.class, DNSToSwitchMapping.class);
    RackResolver.init(conf);
    MRApp app = new MRAppWithHistory(numMaps, numReduces, true, this.getClass().getName(), true);
    app.submit(conf);
    Job job = app.getContext().getAllJobs().values().iterator().next();
    JobId jobId = job.getID();
    LOG.info("JOBID is " + TypeConverter.fromYarn(jobId).toString());
    app.waitForState(job, JobState.SUCCEEDED);
    // make sure all events are flushed
    app.waitForState(Service.STATE.STOPPED);
    String jobhistoryDir = JobHistoryUtils.getHistoryIntermediateDoneDirForUser(conf);
    FileContext fc = null;
    try {
        fc = FileContext.getFileContext(conf);
    } catch (IOException ioe) {
        LOG.info("Can not get FileContext", ioe);
        throw (new Exception("Can not get File Context"));
    }
    if (numMaps == numSuccessfulMaps) {
        String summaryFileName = JobHistoryUtils.getIntermediateSummaryFileName(jobId);
        Path summaryFile = new Path(jobhistoryDir, summaryFileName);
        String jobSummaryString = getJobSummary(fc, summaryFile);
        Assert.assertNotNull(jobSummaryString);
        Assert.assertTrue(jobSummaryString.contains("resourcesPerMap=100"));
        Assert.assertTrue(jobSummaryString.contains("resourcesPerReduce=100"));
        Map<String, String> jobSummaryElements = new HashMap<String, String>();
        StringTokenizer strToken = new StringTokenizer(jobSummaryString, ",");
        while (strToken.hasMoreTokens()) {
            String keypair = strToken.nextToken();
            jobSummaryElements.put(keypair.split("=")[0], keypair.split("=")[1]);
        }
        Assert.assertEquals("JobId does not match", jobId.toString(), jobSummaryElements.get("jobId"));
        Assert.assertEquals("JobName does not match", "test", jobSummaryElements.get("jobName"));
        Assert.assertTrue("submitTime should not be 0", Long.parseLong(jobSummaryElements.get("submitTime")) != 0);
        Assert.assertTrue("launchTime should not be 0", Long.parseLong(jobSummaryElements.get("launchTime")) != 0);
        Assert.assertTrue("firstMapTaskLaunchTime should not be 0", Long.parseLong(jobSummaryElements.get("firstMapTaskLaunchTime")) != 0);
        Assert.assertTrue("firstReduceTaskLaunchTime should not be 0", Long.parseLong(jobSummaryElements.get("firstReduceTaskLaunchTime")) != 0);
        Assert.assertTrue("finishTime should not be 0", Long.parseLong(jobSummaryElements.get("finishTime")) != 0);
        Assert.assertEquals("Mismatch in num map slots", numSuccessfulMaps, Integer.parseInt(jobSummaryElements.get("numMaps")));
        Assert.assertEquals("Mismatch in num reduce slots", numReduces, Integer.parseInt(jobSummaryElements.get("numReduces")));
        Assert.assertEquals("User does not match", System.getProperty("user.name"), jobSummaryElements.get("user"));
        Assert.assertEquals("Queue does not match", "default", jobSummaryElements.get("queue"));
        Assert.assertEquals("Status does not match", "SUCCEEDED", jobSummaryElements.get("status"));
    }
    JobHistory jobHistory = new JobHistory();
    jobHistory.init(conf);
    HistoryFileInfo fileInfo = jobHistory.getJobFileInfo(jobId);
    JobInfo jobInfo;
    long numFinishedMaps;
    synchronized (fileInfo) {
        Path historyFilePath = fileInfo.getHistoryFile();
        FSDataInputStream in = null;
        LOG.info("JobHistoryFile is: " + historyFilePath);
        try {
            in = fc.open(fc.makeQualified(historyFilePath));
        } catch (IOException ioe) {
            LOG.info("Can not open history file: " + historyFilePath, ioe);
            throw (new Exception("Can not open History File"));
        }
        JobHistoryParser parser = new JobHistoryParser(in);
        final EventReader realReader = new EventReader(in);
        EventReader reader = Mockito.mock(EventReader.class);
        if (numMaps == numSuccessfulMaps) {
            reader = realReader;
        } else {
            // Hack!
            final AtomicInteger numFinishedEvents = new AtomicInteger(0);
            Mockito.when(reader.getNextEvent()).thenAnswer(new Answer<HistoryEvent>() {

                public HistoryEvent answer(InvocationOnMock invocation) throws IOException {
                    HistoryEvent event = realReader.getNextEvent();
                    if (event instanceof TaskFinishedEvent) {
                        numFinishedEvents.incrementAndGet();
                    }
                    if (numFinishedEvents.get() <= numSuccessfulMaps) {
                        return event;
                    } else {
                        throw new IOException("test");
                    }
                }
            });
        }
        jobInfo = parser.parse(reader);
        numFinishedMaps = computeFinishedMaps(jobInfo, numMaps, numSuccessfulMaps);
        if (numFinishedMaps != numMaps) {
            Exception parseException = parser.getParseException();
            Assert.assertNotNull("Didn't get expected parse exception", parseException);
        }
    }
    Assert.assertEquals("Incorrect username ", System.getProperty("user.name"), jobInfo.getUsername());
    Assert.assertEquals("Incorrect jobName ", "test", jobInfo.getJobname());
    Assert.assertEquals("Incorrect queuename ", "default", jobInfo.getJobQueueName());
    Assert.assertEquals("incorrect conf path", "test", jobInfo.getJobConfPath());
    Assert.assertEquals("incorrect finishedMap ", numSuccessfulMaps, numFinishedMaps);
    Assert.assertEquals("incorrect finishedReduces ", numReduces, jobInfo.getFinishedReduces());
    Assert.assertEquals("incorrect uberized ", job.isUber(), jobInfo.getUberized());
    Map<TaskID, TaskInfo> allTasks = jobInfo.getAllTasks();
    int totalTasks = allTasks.size();
    Assert.assertEquals("total number of tasks is incorrect  ", (numMaps + numReduces), totalTasks);
    // Verify aminfo
    Assert.assertEquals(1, jobInfo.getAMInfos().size());
    Assert.assertEquals(MRApp.NM_HOST, jobInfo.getAMInfos().get(0).getNodeManagerHost());
    AMInfo amInfo = jobInfo.getAMInfos().get(0);
    Assert.assertEquals(MRApp.NM_PORT, amInfo.getNodeManagerPort());
    Assert.assertEquals(MRApp.NM_HTTP_PORT, amInfo.getNodeManagerHttpPort());
    Assert.assertEquals(1, amInfo.getAppAttemptId().getAttemptId());
    Assert.assertEquals(amInfo.getAppAttemptId(), amInfo.getContainerId().getApplicationAttemptId());
    Assert.assertTrue(amInfo.getStartTime() <= System.currentTimeMillis() && amInfo.getStartTime() >= amStartTimeEst);
    ContainerId fakeCid = MRApp.newContainerId(-1, -1, -1, -1);
    // Assert at taskAttempt level
    for (TaskInfo taskInfo : allTasks.values()) {
        int taskAttemptCount = taskInfo.getAllTaskAttempts().size();
        Assert.assertEquals("total number of task attempts ", 1, taskAttemptCount);
        TaskAttemptInfo taInfo = taskInfo.getAllTaskAttempts().values().iterator().next();
        Assert.assertNotNull(taInfo.getContainerId());
        // Verify the wrong ctor is not being used. Remove after mrv1 is removed.
        Assert.assertFalse(taInfo.getContainerId().equals(fakeCid));
    }
    // Deep compare Job and JobInfo
    for (Task task : job.getTasks().values()) {
        TaskInfo taskInfo = allTasks.get(TypeConverter.fromYarn(task.getID()));
        Assert.assertNotNull("TaskInfo not found", taskInfo);
        for (TaskAttempt taskAttempt : task.getAttempts().values()) {
            TaskAttemptInfo taskAttemptInfo = taskInfo.getAllTaskAttempts().get(TypeConverter.fromYarn((taskAttempt.getID())));
            Assert.assertNotNull("TaskAttemptInfo not found", taskAttemptInfo);
            Assert.assertEquals("Incorrect shuffle port for task attempt", taskAttempt.getShufflePort(), taskAttemptInfo.getShufflePort());
            if (numMaps == numSuccessfulMaps) {
                Assert.assertEquals(MRApp.NM_HOST, taskAttemptInfo.getHostname());
                Assert.assertEquals(MRApp.NM_PORT, taskAttemptInfo.getPort());
                // Verify rack-name
                Assert.assertEquals("rack-name is incorrect", taskAttemptInfo.getRackname(), RACK_NAME);
            }
        }
    }
    // test output for HistoryViewer
    PrintStream stdps = System.out;
    try {
        System.setOut(new PrintStream(outContent));
        HistoryViewer viewer;
        synchronized (fileInfo) {
            viewer = new HistoryViewer(fc.makeQualified(fileInfo.getHistoryFile()).toString(), conf, true);
        }
        viewer.print();
        for (TaskInfo taskInfo : allTasks.values()) {
            String test = (taskInfo.getTaskStatus() == null ? "" : taskInfo.getTaskStatus()) + " " + taskInfo.getTaskType() + " task list for " + taskInfo.getTaskId().getJobID();
            Assert.assertTrue(outContent.toString().indexOf(test) > 0);
            Assert.assertTrue(outContent.toString().indexOf(taskInfo.getTaskId().toString()) > 0);
        }
    } finally {
        System.setOut(stdps);
    }
}
Also used : MRAppWithHistory(org.apache.hadoop.mapreduce.v2.hs.TestJobHistoryEvents.MRAppWithHistory) Task(org.apache.hadoop.mapreduce.v2.app.job.Task) Configuration(org.apache.hadoop.conf.Configuration) HashMap(java.util.HashMap) TaskInfo(org.apache.hadoop.mapreduce.jobhistory.JobHistoryParser.TaskInfo) HistoryViewer(org.apache.hadoop.mapreduce.jobhistory.HistoryViewer) JobInfo(org.apache.hadoop.mapreduce.jobhistory.JobHistoryParser.JobInfo) ContainerId(org.apache.hadoop.yarn.api.records.ContainerId) TaskAttempt(org.apache.hadoop.mapreduce.v2.app.job.TaskAttempt) Job(org.apache.hadoop.mapreduce.v2.app.job.Job) JobId(org.apache.hadoop.mapreduce.v2.api.records.JobId) MRApp(org.apache.hadoop.mapreduce.v2.app.MRApp) Path(org.apache.hadoop.fs.Path) EventReader(org.apache.hadoop.mapreduce.jobhistory.EventReader) PrintStream(java.io.PrintStream) HistoryFileInfo(org.apache.hadoop.mapreduce.v2.hs.HistoryFileManager.HistoryFileInfo) TaskID(org.apache.hadoop.mapreduce.TaskID) IOException(java.io.IOException) HistoryEvent(org.apache.hadoop.mapreduce.jobhistory.HistoryEvent) IOException(java.io.IOException) AMInfo(org.apache.hadoop.mapreduce.jobhistory.JobHistoryParser.AMInfo) StringTokenizer(java.util.StringTokenizer) TaskFinishedEvent(org.apache.hadoop.mapreduce.jobhistory.TaskFinishedEvent) JobHistoryParser(org.apache.hadoop.mapreduce.jobhistory.JobHistoryParser) AtomicInteger(java.util.concurrent.atomic.AtomicInteger) InvocationOnMock(org.mockito.invocation.InvocationOnMock) TaskAttemptInfo(org.apache.hadoop.mapreduce.jobhistory.JobHistoryParser.TaskAttemptInfo) FSDataInputStream(org.apache.hadoop.fs.FSDataInputStream) FileContext(org.apache.hadoop.fs.FileContext)

Example 8 with TaskId

use of org.apache.hadoop.mapreduce.v2.api.records.TaskId in project hadoop by apache.

the class TestJobHistoryEntities method testCompletedTaskAttempt.

@Test(timeout = 10000)
public void testCompletedTaskAttempt() throws Exception {
    HistoryFileInfo info = mock(HistoryFileInfo.class);
    when(info.getConfFile()).thenReturn(fullConfPath);
    completedJob = new CompletedJob(conf, jobId, fullHistoryPath, loadTasks, "user", info, jobAclsManager);
    TaskId mt1Id = MRBuilderUtils.newTaskId(jobId, 0, TaskType.MAP);
    TaskId rt1Id = MRBuilderUtils.newTaskId(jobId, 0, TaskType.REDUCE);
    TaskAttemptId mta1Id = MRBuilderUtils.newTaskAttemptId(mt1Id, 0);
    TaskAttemptId rta1Id = MRBuilderUtils.newTaskAttemptId(rt1Id, 0);
    Task mt1 = completedJob.getTask(mt1Id);
    Task rt1 = completedJob.getTask(rt1Id);
    TaskAttempt mta1 = mt1.getAttempt(mta1Id);
    assertEquals(TaskAttemptState.SUCCEEDED, mta1.getState());
    assertEquals("localhost:45454", mta1.getAssignedContainerMgrAddress());
    assertEquals("localhost:9999", mta1.getNodeHttpAddress());
    TaskAttemptReport mta1Report = mta1.getReport();
    assertEquals(TaskAttemptState.SUCCEEDED, mta1Report.getTaskAttemptState());
    assertEquals("localhost", mta1Report.getNodeManagerHost());
    assertEquals(45454, mta1Report.getNodeManagerPort());
    assertEquals(9999, mta1Report.getNodeManagerHttpPort());
    TaskAttempt rta1 = rt1.getAttempt(rta1Id);
    assertEquals(TaskAttemptState.SUCCEEDED, rta1.getState());
    assertEquals("localhost:45454", rta1.getAssignedContainerMgrAddress());
    assertEquals("localhost:9999", rta1.getNodeHttpAddress());
    TaskAttemptReport rta1Report = rta1.getReport();
    assertEquals(TaskAttemptState.SUCCEEDED, rta1Report.getTaskAttemptState());
    assertEquals("localhost", rta1Report.getNodeManagerHost());
    assertEquals(45454, rta1Report.getNodeManagerPort());
    assertEquals(9999, rta1Report.getNodeManagerHttpPort());
}
Also used : HistoryFileInfo(org.apache.hadoop.mapreduce.v2.hs.HistoryFileManager.HistoryFileInfo) Task(org.apache.hadoop.mapreduce.v2.app.job.Task) TaskId(org.apache.hadoop.mapreduce.v2.api.records.TaskId) TaskAttemptReport(org.apache.hadoop.mapreduce.v2.api.records.TaskAttemptReport) TaskAttemptId(org.apache.hadoop.mapreduce.v2.api.records.TaskAttemptId) TaskAttempt(org.apache.hadoop.mapreduce.v2.app.job.TaskAttempt) Test(org.junit.Test)

Example 9 with TaskId

use of org.apache.hadoop.mapreduce.v2.api.records.TaskId in project hadoop by apache.

the class TestJobHistoryEntities method testCompletedTask.

@Test(timeout = 10000)
public void testCompletedTask() throws Exception {
    HistoryFileInfo info = mock(HistoryFileInfo.class);
    when(info.getConfFile()).thenReturn(fullConfPath);
    completedJob = new CompletedJob(conf, jobId, fullHistoryPath, loadTasks, "user", info, jobAclsManager);
    TaskId mt1Id = MRBuilderUtils.newTaskId(jobId, 0, TaskType.MAP);
    TaskId rt1Id = MRBuilderUtils.newTaskId(jobId, 0, TaskType.REDUCE);
    Map<TaskId, Task> mapTasks = completedJob.getTasks(TaskType.MAP);
    Map<TaskId, Task> reduceTasks = completedJob.getTasks(TaskType.REDUCE);
    assertEquals(10, mapTasks.size());
    assertEquals(2, reduceTasks.size());
    Task mt1 = mapTasks.get(mt1Id);
    assertEquals(1, mt1.getAttempts().size());
    assertEquals(TaskState.SUCCEEDED, mt1.getState());
    TaskReport mt1Report = mt1.getReport();
    assertEquals(TaskState.SUCCEEDED, mt1Report.getTaskState());
    assertEquals(mt1Id, mt1Report.getTaskId());
    Task rt1 = reduceTasks.get(rt1Id);
    assertEquals(1, rt1.getAttempts().size());
    assertEquals(TaskState.SUCCEEDED, rt1.getState());
    TaskReport rt1Report = rt1.getReport();
    assertEquals(TaskState.SUCCEEDED, rt1Report.getTaskState());
    assertEquals(rt1Id, rt1Report.getTaskId());
}
Also used : HistoryFileInfo(org.apache.hadoop.mapreduce.v2.hs.HistoryFileManager.HistoryFileInfo) Task(org.apache.hadoop.mapreduce.v2.app.job.Task) TaskId(org.apache.hadoop.mapreduce.v2.api.records.TaskId) TaskReport(org.apache.hadoop.mapreduce.v2.api.records.TaskReport) Test(org.junit.Test)

Example 10 with TaskId

use of org.apache.hadoop.mapreduce.v2.api.records.TaskId in project hadoop by apache.

the class HsWebServices method getJobTask.

@GET
@Path("/mapreduce/jobs/{jobid}/tasks/{taskid}")
@Produces({ MediaType.APPLICATION_JSON + "; " + JettyUtils.UTF_8, MediaType.APPLICATION_XML + "; " + JettyUtils.UTF_8 })
public TaskInfo getJobTask(@Context HttpServletRequest hsr, @PathParam("jobid") String jid, @PathParam("taskid") String tid) {
    init();
    Job job = AMWebServices.getJobFromJobIdString(jid, ctx);
    checkAccess(job, hsr);
    Task task = AMWebServices.getTaskFromTaskIdString(tid, job);
    return new TaskInfo(task);
}
Also used : TaskInfo(org.apache.hadoop.mapreduce.v2.app.webapp.dao.TaskInfo) Task(org.apache.hadoop.mapreduce.v2.app.job.Task) Job(org.apache.hadoop.mapreduce.v2.app.job.Job) Path(javax.ws.rs.Path) Produces(javax.ws.rs.Produces) GET(javax.ws.rs.GET)

Aggregations

TaskId (org.apache.hadoop.mapreduce.v2.api.records.TaskId)113 Test (org.junit.Test)75 Task (org.apache.hadoop.mapreduce.v2.app.job.Task)69 Job (org.apache.hadoop.mapreduce.v2.app.job.Job)60 JobId (org.apache.hadoop.mapreduce.v2.api.records.JobId)58 TaskAttemptId (org.apache.hadoop.mapreduce.v2.api.records.TaskAttemptId)56 TaskAttempt (org.apache.hadoop.mapreduce.v2.app.job.TaskAttempt)42 Configuration (org.apache.hadoop.conf.Configuration)29 AppContext (org.apache.hadoop.mapreduce.v2.app.AppContext)24 TaskAttemptEvent (org.apache.hadoop.mapreduce.v2.app.job.event.TaskAttemptEvent)24 Path (org.apache.hadoop.fs.Path)23 ApplicationId (org.apache.hadoop.yarn.api.records.ApplicationId)22 HashMap (java.util.HashMap)20 JobConf (org.apache.hadoop.mapred.JobConf)17 ApplicationAttemptId (org.apache.hadoop.yarn.api.records.ApplicationAttemptId)17 ContainerId (org.apache.hadoop.yarn.api.records.ContainerId)17 MapTaskAttemptImpl (org.apache.hadoop.mapred.MapTaskAttemptImpl)16 TaskAttemptListener (org.apache.hadoop.mapreduce.v2.app.TaskAttemptListener)16 InetSocketAddress (java.net.InetSocketAddress)15 TaskSplitMetaInfo (org.apache.hadoop.mapreduce.split.JobSplit.TaskSplitMetaInfo)15