Search in sources :

Example 41 with JobId

use of org.apache.hadoop.mapreduce.v2.api.records.JobId 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 42 with JobId

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

the class TestJobHistoryParsing method testJobHistoryMethods.

/**
   * Simple test some methods of JobHistory
   */
@Test(timeout = 20000)
public void testJobHistoryMethods() throws Exception {
    LOG.info("STARTING testJobHistoryMethods");
    try {
        Configuration configuration = new Configuration();
        configuration.setClass(NET_TOPOLOGY_NODE_SWITCH_MAPPING_IMPL_KEY, MyResolver.class, DNSToSwitchMapping.class);
        RackResolver.init(configuration);
        MRApp app = new MRAppWithHistory(1, 1, true, this.getClass().getName(), true);
        app.submit(configuration);
        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 job history events are handled
        app.waitForState(Service.STATE.STOPPED);
        JobHistory jobHistory = new JobHistory();
        jobHistory.init(configuration);
        // Method getAllJobs
        Assert.assertEquals(1, jobHistory.getAllJobs().size());
        // and with ApplicationId
        Assert.assertEquals(1, jobHistory.getAllJobs(app.getAppID()).size());
        JobsInfo jobsinfo = jobHistory.getPartialJobs(0L, 10L, null, "default", 0L, System.currentTimeMillis() + 1, 0L, System.currentTimeMillis() + 1, JobState.SUCCEEDED);
        Assert.assertEquals(1, jobsinfo.getJobs().size());
        Assert.assertNotNull(jobHistory.getApplicationAttemptId());
        // test Application Id
        Assert.assertEquals("application_0_0000", jobHistory.getApplicationID().toString());
        Assert.assertEquals("Job History Server", jobHistory.getApplicationName());
        // method does not work
        Assert.assertNull(jobHistory.getEventHandler());
        // method does not work
        Assert.assertNull(jobHistory.getClock());
        // method does not work
        Assert.assertNull(jobHistory.getClusterInfo());
    } finally {
        LOG.info("FINISHED testJobHistoryMethods");
    }
}
Also used : MRAppWithHistory(org.apache.hadoop.mapreduce.v2.hs.TestJobHistoryEvents.MRAppWithHistory) Configuration(org.apache.hadoop.conf.Configuration) JobsInfo(org.apache.hadoop.mapreduce.v2.hs.webapp.dao.JobsInfo) 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) Test(org.junit.Test)

Example 43 with JobId

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

the class TestJobHistoryParsing method testDeleteFileInfo.

/**
   * Test clean old history files. Files should be deleted after 1 week by
   * default.
   */
@Test(timeout = 15000)
public void testDeleteFileInfo() throws Exception {
    LOG.info("STARTING testDeleteFileInfo");
    try {
        Configuration conf = new Configuration();
        conf.setClass(NET_TOPOLOGY_NODE_SWITCH_MAPPING_IMPL_KEY, MyResolver.class, DNSToSwitchMapping.class);
        RackResolver.init(conf);
        MRApp app = new MRAppWithHistory(1, 1, true, this.getClass().getName(), true);
        app.submit(conf);
        Job job = app.getContext().getAllJobs().values().iterator().next();
        JobId jobId = job.getID();
        app.waitForState(job, JobState.SUCCEEDED);
        // make sure all events are flushed
        app.waitForState(Service.STATE.STOPPED);
        HistoryFileManager hfm = new HistoryFileManager();
        hfm.init(conf);
        HistoryFileInfo fileInfo = hfm.getFileInfo(jobId);
        hfm.initExisting();
        // directory
        while (fileInfo.isMovePending()) {
            Thread.sleep(300);
        }
        Assert.assertNotNull(hfm.jobListCache.values());
        // try to remove fileInfo
        hfm.clean();
        // check that fileInfo does not deleted
        Assert.assertFalse(fileInfo.isDeleted());
        // correct live time
        hfm.setMaxHistoryAge(-1);
        hfm.clean();
        hfm.stop();
        Assert.assertTrue("Thread pool shutdown", hfm.moveToDoneExecutor.isTerminated());
        // should be deleted !
        Assert.assertTrue("file should be deleted ", fileInfo.isDeleted());
    } finally {
        LOG.info("FINISHED testDeleteFileInfo");
    }
}
Also used : MRAppWithHistory(org.apache.hadoop.mapreduce.v2.hs.TestJobHistoryEvents.MRAppWithHistory) HistoryFileInfo(org.apache.hadoop.mapreduce.v2.hs.HistoryFileManager.HistoryFileInfo) Configuration(org.apache.hadoop.conf.Configuration) 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) Test(org.junit.Test)

Example 44 with JobId

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

the class TestJobHistoryEntities method testGetTaskAttemptCompletionEvent.

/**
   * Simple test of some methods of CompletedJob
   * @throws Exception
   */
@Test(timeout = 30000)
public void testGetTaskAttemptCompletionEvent() throws Exception {
    HistoryFileInfo info = mock(HistoryFileInfo.class);
    when(info.getConfFile()).thenReturn(fullConfPath);
    completedJob = new CompletedJob(conf, jobId, fullHistoryPath, loadTasks, "user", info, jobAclsManager);
    TaskCompletionEvent[] events = completedJob.getMapAttemptCompletionEvents(0, 1000);
    assertEquals(10, completedJob.getMapAttemptCompletionEvents(0, 10).length);
    int currentEventId = 0;
    for (TaskCompletionEvent taskAttemptCompletionEvent : events) {
        int eventId = taskAttemptCompletionEvent.getEventId();
        assertTrue(eventId >= currentEventId);
        currentEventId = eventId;
    }
    assertNull(completedJob.loadConfFile());
    // job name
    assertEquals("Sleep job", completedJob.getName());
    // queue name
    assertEquals("default", completedJob.getQueueName());
    // progress
    assertEquals(1.0, completedJob.getProgress(), 0.001);
    // 12 rows in answer
    assertEquals(12, completedJob.getTaskAttemptCompletionEvents(0, 1000).length);
    // select first 10 rows
    assertEquals(10, completedJob.getTaskAttemptCompletionEvents(0, 10).length);
    // select 5-10 rows include 5th
    assertEquals(7, completedJob.getTaskAttemptCompletionEvents(5, 10).length);
    // without errors
    assertEquals(1, completedJob.getDiagnostics().size());
    assertEquals("", completedJob.getDiagnostics().get(0));
    assertEquals(0, completedJob.getJobACLs().size());
}
Also used : HistoryFileInfo(org.apache.hadoop.mapreduce.v2.hs.HistoryFileManager.HistoryFileInfo) TaskCompletionEvent(org.apache.hadoop.mapred.TaskCompletionEvent) Test(org.junit.Test)

Example 45 with JobId

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

the class TestJobHistoryEntities method testCompletedJobWithDiagnostics.

@Test(timeout = 30000)
public void testCompletedJobWithDiagnostics() throws Exception {
    final String jobError = "Job Diagnostics";
    JobInfo jobInfo = spy(new JobInfo());
    when(jobInfo.getErrorInfo()).thenReturn(jobError);
    when(jobInfo.getJobStatus()).thenReturn(JobState.FAILED.toString());
    when(jobInfo.getAMInfos()).thenReturn(Collections.<JobHistoryParser.AMInfo>emptyList());
    final JobHistoryParser mockParser = mock(JobHistoryParser.class);
    when(mockParser.parse()).thenReturn(jobInfo);
    HistoryFileInfo info = mock(HistoryFileInfo.class);
    when(info.getConfFile()).thenReturn(fullConfPath);
    when(info.getHistoryFile()).thenReturn(fullHistoryPath);
    CompletedJob job = new CompletedJob(conf, jobId, fullHistoryPath, loadTasks, "user", info, jobAclsManager) {

        @Override
        protected JobHistoryParser createJobHistoryParser(Path historyFileAbsolute) throws IOException {
            return mockParser;
        }
    };
    assertEquals(jobError, job.getReport().getDiagnostics());
}
Also used : Path(org.apache.hadoop.fs.Path) HistoryFileInfo(org.apache.hadoop.mapreduce.v2.hs.HistoryFileManager.HistoryFileInfo) JobHistoryParser(org.apache.hadoop.mapreduce.jobhistory.JobHistoryParser) JobInfo(org.apache.hadoop.mapreduce.jobhistory.JobHistoryParser.JobInfo) Test(org.junit.Test)

Aggregations

JobId (org.apache.hadoop.mapreduce.v2.api.records.JobId)276 Test (org.junit.Test)238 Job (org.apache.hadoop.mapreduce.v2.app.job.Job)212 ClientResponse (com.sun.jersey.api.client.ClientResponse)103 WebResource (com.sun.jersey.api.client.WebResource)103 Task (org.apache.hadoop.mapreduce.v2.app.job.Task)88 JSONObject (org.codehaus.jettison.json.JSONObject)81 Configuration (org.apache.hadoop.conf.Configuration)77 TaskId (org.apache.hadoop.mapreduce.v2.api.records.TaskId)61 ApplicationAttemptId (org.apache.hadoop.yarn.api.records.ApplicationAttemptId)55 Path (org.apache.hadoop.fs.Path)52 TaskAttemptId (org.apache.hadoop.mapreduce.v2.api.records.TaskAttemptId)51 ApplicationId (org.apache.hadoop.yarn.api.records.ApplicationId)47 TaskAttempt (org.apache.hadoop.mapreduce.v2.app.job.TaskAttempt)41 YarnConfiguration (org.apache.hadoop.yarn.conf.YarnConfiguration)36 AppContext (org.apache.hadoop.mapreduce.v2.app.AppContext)35 IOException (java.io.IOException)32 JobConf (org.apache.hadoop.mapred.JobConf)28 HistoryFileInfo (org.apache.hadoop.mapreduce.v2.hs.HistoryFileManager.HistoryFileInfo)25 JobID (org.apache.hadoop.mapreduce.JobID)23