use of org.apache.tez.hadoop.shim.HadoopShim in project tez by apache.
the class TestOrderedWordCount method run.
@Override
public int run(String[] args) throws Exception {
Configuration conf = getConf();
String[] otherArgs = new GenericOptionsParser(conf, args).getRemainingArgs();
boolean generateSplitsInClient;
SplitsInClientOptionParser splitCmdLineParser = new SplitsInClientOptionParser();
try {
generateSplitsInClient = splitCmdLineParser.parse(otherArgs, false);
otherArgs = splitCmdLineParser.getRemainingArgs();
} catch (ParseException e1) {
System.err.println("Invalid options");
printUsage();
return 2;
}
boolean useTezSession = conf.getBoolean("USE_TEZ_SESSION", true);
long interJobSleepTimeout = conf.getInt("INTER_JOB_SLEEP_INTERVAL", 0) * 1000;
boolean retainStagingDir = conf.getBoolean("RETAIN_STAGING_DIR", false);
boolean useMRSettings = conf.getBoolean("USE_MR_CONFIGS", true);
// TODO needs to use auto reduce parallelism
int intermediateNumReduceTasks = conf.getInt("IREDUCE_NUM_TASKS", 2);
int maxDataLengthThroughIPC = conf.getInt(MAX_IPC_DATA_LENGTH, -1);
int exceedDataLimit = conf.getInt(EXCEED_IPC_DATA_LIMIT, 3);
if (maxDataLengthThroughIPC > 0) {
conf.setInt(CommonConfigurationKeys.IPC_MAXIMUM_DATA_LENGTH, maxDataLengthThroughIPC * 1024 * 1024);
}
if (((otherArgs.length % 2) != 0) || (!useTezSession && otherArgs.length != 2)) {
printUsage();
return 2;
}
List<String> inputPaths = new ArrayList<String>();
List<String> outputPaths = new ArrayList<String>();
TezConfiguration tezConf = new TezConfiguration(conf);
for (int i = 0; i < otherArgs.length; i += 2) {
FileSystem inputPathFs = new Path(otherArgs[i]).getFileSystem(tezConf);
inputPaths.add(inputPathFs.makeQualified(new Path(otherArgs[i])).toString());
FileSystem outputPathFs = new Path(otherArgs[i + 1]).getFileSystem(tezConf);
outputPaths.add(outputPathFs.makeQualified(new Path(otherArgs[i + 1])).toString());
}
UserGroupInformation.setConfiguration(conf);
HadoopShim hadoopShim = new HadoopShimsLoader(tezConf).getHadoopShim();
TestOrderedWordCount instance = new TestOrderedWordCount();
FileSystem fs = FileSystem.get(conf);
String stagingDirStr = conf.get(TezConfiguration.TEZ_AM_STAGING_DIR, TezConfiguration.TEZ_AM_STAGING_DIR_DEFAULT) + Path.SEPARATOR + Long.toString(System.currentTimeMillis());
Path stagingDir = new Path(stagingDirStr);
FileSystem pathFs = stagingDir.getFileSystem(tezConf);
pathFs.mkdirs(new Path(stagingDirStr));
tezConf.set(TezConfiguration.TEZ_AM_STAGING_DIR, stagingDirStr);
stagingDir = pathFs.makeQualified(new Path(stagingDirStr));
TokenCache.obtainTokensForNamenodes(instance.credentials, new Path[] { stagingDir }, conf);
TezClientUtils.ensureStagingDirExists(tezConf, stagingDir);
if (useTezSession) {
LOG.info("Creating Tez Session");
tezConf.setBoolean(TezConfiguration.TEZ_AM_SESSION_MODE, true);
} else {
tezConf.setBoolean(TezConfiguration.TEZ_AM_SESSION_MODE, false);
}
TezClient tezSession = TezClient.create("OrderedWordCountSession", tezConf, null, instance.credentials);
tezSession.start();
if (tezSession.getAppMasterApplicationId() != null) {
TezUtilsInternal.setHadoopCallerContext(hadoopShim, tezSession.getAppMasterApplicationId());
}
DAGStatus dagStatus = null;
DAGClient dagClient = null;
String[] vNames = { "initialmap", "intermediate_reducer", "finalreduce" };
Set<StatusGetOpts> statusGetOpts = EnumSet.of(StatusGetOpts.GET_COUNTERS);
try {
for (int dagIndex = 1; dagIndex <= inputPaths.size(); ++dagIndex) {
if (dagIndex != 1 && interJobSleepTimeout > 0) {
try {
LOG.info("Sleeping between jobs, sleepInterval=" + (interJobSleepTimeout / 1000));
Thread.sleep(interJobSleepTimeout);
} catch (InterruptedException e) {
LOG.info("Main thread interrupted. Breaking out of job loop");
break;
}
}
String inputPath = inputPaths.get(dagIndex - 1);
String outputPath = outputPaths.get(dagIndex - 1);
if (fs.exists(new Path(outputPath))) {
throw new FileAlreadyExistsException("Output directory " + outputPath + " already exists");
}
LOG.info("Running OrderedWordCount DAG" + ", dagIndex=" + dagIndex + ", inputPath=" + inputPath + ", outputPath=" + outputPath);
Map<String, LocalResource> localResources = new TreeMap<String, LocalResource>();
DAG dag = instance.createDAG(fs, tezConf, localResources, stagingDir, dagIndex, inputPath, outputPath, generateSplitsInClient, useMRSettings, intermediateNumReduceTasks, maxDataLengthThroughIPC, exceedDataLimit);
String callerType = "TestOrderedWordCount";
String callerId = tezSession.getAppMasterApplicationId() == null ? ("UnknownApp_" + System.currentTimeMillis() + dagIndex) : (tezSession.getAppMasterApplicationId().toString() + "_" + dagIndex);
dag.setCallerContext(CallerContext.create("Tez", callerId, callerType, "TestOrderedWordCount Job"));
boolean doPreWarm = dagIndex == 1 && useTezSession && conf.getBoolean("PRE_WARM_SESSION", true);
int preWarmNumContainers = 0;
if (doPreWarm) {
preWarmNumContainers = conf.getInt("PRE_WARM_NUM_CONTAINERS", 0);
if (preWarmNumContainers <= 0) {
doPreWarm = false;
}
}
if (doPreWarm) {
LOG.info("Pre-warming Session");
PreWarmVertex preWarmVertex = PreWarmVertex.create("PreWarm", preWarmNumContainers, dag.getVertex("initialmap").getTaskResource());
preWarmVertex.addTaskLocalFiles(dag.getVertex("initialmap").getTaskLocalFiles());
preWarmVertex.setTaskEnvironment(dag.getVertex("initialmap").getTaskEnvironment());
preWarmVertex.setTaskLaunchCmdOpts(dag.getVertex("initialmap").getTaskLaunchCmdOpts());
tezSession.preWarm(preWarmVertex);
}
if (useTezSession) {
LOG.info("Waiting for TezSession to get into ready state");
waitForTezSessionReady(tezSession);
LOG.info("Submitting DAG to Tez Session, dagIndex=" + dagIndex);
dagClient = tezSession.submitDAG(dag);
LOG.info("Submitted DAG to Tez Session, dagIndex=" + dagIndex);
} else {
LOG.info("Submitting DAG as a new Tez Application");
dagClient = tezSession.submitDAG(dag);
}
while (true) {
dagStatus = dagClient.getDAGStatus(statusGetOpts);
if (dagStatus.getState() == DAGStatus.State.RUNNING || dagStatus.getState() == DAGStatus.State.SUCCEEDED || dagStatus.getState() == DAGStatus.State.FAILED || dagStatus.getState() == DAGStatus.State.KILLED || dagStatus.getState() == DAGStatus.State.ERROR) {
break;
}
try {
Thread.sleep(500);
} catch (InterruptedException e) {
// continue;
}
}
while (dagStatus.getState() != DAGStatus.State.SUCCEEDED && dagStatus.getState() != DAGStatus.State.FAILED && dagStatus.getState() != DAGStatus.State.KILLED && dagStatus.getState() != DAGStatus.State.ERROR) {
if (dagStatus.getState() == DAGStatus.State.RUNNING) {
ExampleDriver.printDAGStatus(dagClient, vNames);
}
try {
try {
Thread.sleep(1000);
} catch (InterruptedException e) {
// continue;
}
dagStatus = dagClient.getDAGStatus(statusGetOpts);
} catch (TezException e) {
LOG.error("Failed to get application progress. Exiting");
return -1;
}
}
ExampleDriver.printDAGStatus(dagClient, vNames, true, true);
LOG.info("DAG " + dagIndex + " completed. " + "FinalState=" + dagStatus.getState());
if (dagStatus.getState() != DAGStatus.State.SUCCEEDED) {
LOG.info("DAG " + dagIndex + " diagnostics: " + dagStatus.getDiagnostics());
}
}
} catch (Exception e) {
LOG.error("Error occurred when submitting/running DAGs", e);
throw e;
} finally {
if (!retainStagingDir) {
pathFs.delete(stagingDir, true);
}
LOG.info("Shutting down session");
tezSession.stop();
}
if (!useTezSession) {
ExampleDriver.printDAGStatus(dagClient, vNames);
LOG.info("Application completed. " + "FinalState=" + dagStatus.getState());
}
return dagStatus.getState() == DAGStatus.State.SUCCEEDED ? 0 : 1;
}
use of org.apache.tez.hadoop.shim.HadoopShim in project tez by apache.
the class TestHistoryEventHandler method createHandler.
private HistoryEventHandler createHandler(HistoryLogLevel logLevel) {
Configuration conf = new Configuration(baseConfig);
conf.setBoolean(TezConfiguration.DAG_RECOVERY_ENABLED, false);
conf.set(TezConfiguration.TEZ_HISTORY_LOGGING_SERVICE_CLASS, InMemoryHistoryLoggingService.class.getName());
if (logLevel != null) {
conf.setEnum(TezConfiguration.TEZ_HISTORY_LOGGING_LOGLEVEL, logLevel);
}
DAG dag = mock(DAG.class);
when(dag.getConf()).thenReturn(conf);
AppContext appContext = mock(AppContext.class);
when(appContext.getApplicationID()).thenReturn(appId);
when(appContext.getHadoopShim()).thenReturn(new HadoopShim() {
});
when(appContext.getAMConf()).thenReturn(conf);
when(appContext.getCurrentDAG()).thenReturn(dag);
HistoryEventHandler handler = new HistoryEventHandler(appContext);
handler.init(conf);
return handler;
}
use of org.apache.tez.hadoop.shim.HadoopShim in project tez by apache.
the class TestATSV15HistoryLoggingService method createService.
private ATSV15HistoryLoggingService createService(int numDagsPerGroup) throws IOException, YarnException {
ATSV15HistoryLoggingService service = new ATSV15HistoryLoggingService();
appContext = mock(AppContext.class);
when(appContext.getApplicationID()).thenReturn(appId);
when(appContext.getHadoopShim()).thenReturn(new HadoopShim() {
});
service.setAppContext(appContext);
Configuration conf = new Configuration(false);
if (numDagsPerGroup != -1) {
conf.setInt(TezConfiguration.TEZ_HISTORY_LOGGING_TIMELINE_NUM_DAGS_PER_GROUP, numDagsPerGroup);
}
service.init(conf);
// Set timeline service.
timelineClient = mock(TimelineClient.class);
entityLog = new HashMap<>();
// timelineClient.init(conf);
when(timelineClient.getDelegationToken(anyString())).thenReturn(null);
when(timelineClient.renewDelegationToken(Matchers.<Token<TimelineDelegationTokenIdentifier>>any())).thenReturn(0L);
when(timelineClient.putEntities(Matchers.<TimelineEntity>anyVararg())).thenAnswer(new Answer() {
@Override
public TimelinePutResponse answer(InvocationOnMock invocation) throws Throwable {
return putEntityHelper(DEFAULT_GROUP_ID, invocation.getArguments(), 0);
}
});
when(timelineClient.putEntities(any(ApplicationAttemptId.class), any(TimelineEntityGroupId.class), Matchers.<TimelineEntity>anyVararg())).thenAnswer(new Answer() {
@Override
public TimelinePutResponse answer(InvocationOnMock invocation) throws Throwable {
return putEntityHelper(invocation.getArgumentAt(1, TimelineEntityGroupId.class), invocation.getArguments(), 2);
}
});
service.timelineClient = timelineClient;
return service;
}
use of org.apache.tez.hadoop.shim.HadoopShim in project tez by apache.
the class TezChild method main.
public static void main(String[] args) throws IOException, InterruptedException, TezException {
final Configuration defaultConf = new Configuration();
Thread.setDefaultUncaughtExceptionHandler(new YarnUncaughtExceptionHandler());
final String pid = System.getenv().get("JVM_PID");
assert args.length == 5;
String host = args[0];
int port = Integer.parseInt(args[1]);
final String containerIdentifier = args[2];
final String tokenIdentifier = args[3];
final int attemptNumber = Integer.parseInt(args[4]);
final String[] localDirs = TezCommonUtils.getTrimmedStrings(System.getenv(Environment.LOCAL_DIRS.name()));
LOG.info("TezChild starting with PID=" + pid + ", containerIdentifier=" + containerIdentifier);
if (LOG.isDebugEnabled()) {
LOG.debug("Info from cmd line: AM-host: " + host + " AM-port: " + port + " containerIdentifier: " + containerIdentifier + " appAttemptNumber: " + attemptNumber + " tokenIdentifier: " + tokenIdentifier);
}
// Security framework already loaded the tokens into current ugi
DAGProtos.ConfigurationProto confProto = TezUtilsInternal.readUserSpecifiedTezConfiguration(System.getenv(Environment.PWD.name()));
TezUtilsInternal.addUserSpecifiedTezConfiguration(defaultConf, confProto.getConfKeyValuesList());
UserGroupInformation.setConfiguration(defaultConf);
Credentials credentials = UserGroupInformation.getCurrentUser().getCredentials();
HadoopShim hadoopShim = new HadoopShimsLoader(defaultConf).getHadoopShim();
// log the system properties
if (LOG.isInfoEnabled()) {
String systemPropsToLog = TezCommonUtils.getSystemPropertiesToLog(defaultConf);
if (systemPropsToLog != null) {
LOG.info(systemPropsToLog);
}
}
TezChild tezChild = newTezChild(defaultConf, host, port, containerIdentifier, tokenIdentifier, attemptNumber, localDirs, System.getenv(Environment.PWD.name()), System.getenv(), pid, new ExecutionContextImpl(System.getenv(Environment.NM_HOST.name())), credentials, Runtime.getRuntime().maxMemory(), System.getenv(ApplicationConstants.Environment.USER.toString()), null, true, hadoopShim);
tezChild.run();
}
use of org.apache.tez.hadoop.shim.HadoopShim in project tez by apache.
the class TestATSHistoryV15 method testGetGroupId.
@Test
public void testGetGroupId() throws Exception {
ApplicationId appId = ApplicationId.newInstance(1000l, 1);
TezDAGID dagid = TezDAGID.getInstance(appId, 1);
for (final HistoryEventType eventType : HistoryEventType.values()) {
HistoryEvent historyEvent = new HistoryEvent() {
@Override
public HistoryEventType getEventType() {
return eventType;
}
@Override
public boolean isRecoveryEvent() {
return false;
}
@Override
public boolean isHistoryEvent() {
return false;
}
@Override
public void toProtoStream(OutputStream outputStream) throws IOException {
}
@Override
public void fromProtoStream(InputStream inputStream) throws IOException {
}
};
DAGHistoryEvent event = new DAGHistoryEvent(dagid, historyEvent);
ATSV15HistoryLoggingService service = new ATSV15HistoryLoggingService();
AppContext appContext = mock(AppContext.class);
when(appContext.getApplicationID()).thenReturn(appId);
when(appContext.getHadoopShim()).thenReturn(new HadoopShim() {
});
service.setAppContext(appContext);
TimelineEntityGroupId grpId = service.getGroupId(event);
Assert.assertNotNull(grpId);
Assert.assertEquals(appId, grpId.getApplicationId());
switch(eventType) {
case AM_LAUNCHED:
case APP_LAUNCHED:
case AM_STARTED:
case CONTAINER_LAUNCHED:
case CONTAINER_STOPPED:
Assert.assertEquals(appId.toString(), grpId.getTimelineEntityGroupId());
break;
default:
Assert.assertEquals(dagid.toString(), grpId.getTimelineEntityGroupId());
}
service.close();
}
}
Aggregations