Examples with TezConfiguration org.apache.tez.dag.api.TezConfiguration used on opensource projects

Example 1 with TezConfiguration

use of org.apache.tez.dag.api.TezConfiguration in project hive by apache.

the class DagUtils method createConfiguration.

/**
   * Creates and initializes a JobConf object that can be used to execute
   * the DAG. The configuration object will contain configurations from mapred-site
   * overlaid with key/value pairs from the hiveConf object. Finally it will also
   * contain some hive specific configurations that do not change from DAG to DAG.
   *
   * @param hiveConf Current hiveConf for the execution
   * @return JobConf base configuration for job execution
   * @throws IOException
   */
public JobConf createConfiguration(HiveConf hiveConf) throws IOException {
    hiveConf.setBoolean("mapred.mapper.new-api", false);
    JobConf conf = new JobConf(new TezConfiguration(hiveConf));
    conf.set("mapred.output.committer.class", NullOutputCommitter.class.getName());
    conf.setBoolean("mapred.committer.job.setup.cleanup.needed", false);
    conf.setBoolean("mapred.committer.job.task.cleanup.needed", false);
    conf.setClass("mapred.output.format.class", HiveOutputFormatImpl.class, OutputFormat.class);
    conf.set(MRJobConfig.OUTPUT_KEY_CLASS, HiveKey.class.getName());
    conf.set(MRJobConfig.OUTPUT_VALUE_CLASS, BytesWritable.class.getName());
    conf.set("mapred.partitioner.class", HiveConf.getVar(conf, HiveConf.ConfVars.HIVEPARTITIONER));
    conf.set("tez.runtime.partitioner.class", MRPartitioner.class.getName());
    // Removing job credential entry/ cannot be set on the tasks
    conf.unset("mapreduce.job.credentials.binary");
    hiveConf.stripHiddenConfigurations(conf);
    return conf;
}

Example 2 with TezConfiguration

use of org.apache.tez.dag.api.TezConfiguration in project hive by apache.

the class TezSessionState method openInternal.

protected void openInternal(final HiveConf conf, Collection<String> additionalFiles, boolean isAsync, LogHelper console, Path scratchDir) throws IOException, LoginException, IllegalArgumentException, URISyntaxException, TezException {
    this.conf = conf;
    // TODO Why is the queue name set again. It has already been setup via setQueueName. Do only one of the two.
    String confQueueName = conf.get(TezConfiguration.TEZ_QUEUE_NAME);
    if (queueName != null && !queueName.equals(confQueueName)) {
        LOG.warn("Resetting a queue name that was already set: was " + queueName + ", now " + confQueueName);
    }
    this.queueName = confQueueName;
    this.doAsEnabled = conf.getBoolVar(HiveConf.ConfVars.HIVE_SERVER2_ENABLE_DOAS);
    final boolean llapMode = "llap".equalsIgnoreCase(HiveConf.getVar(conf, HiveConf.ConfVars.HIVE_EXECUTION_MODE));
    // TODO This - at least for the session pool - will always be the hive user. How does doAs above this affect things ?
    UserGroupInformation ugi = Utils.getUGI();
    user = ugi.getShortUserName();
    LOG.info("User of session id " + sessionId + " is " + user);
    // create the tez tmp dir
    tezScratchDir = scratchDir == null ? createTezDir(sessionId) : scratchDir;
    additionalFilesNotFromConf.clear();
    if (additionalFiles != null) {
        additionalFilesNotFromConf.addAll(additionalFiles);
    }
    refreshLocalResourcesFromConf(conf);
    // unless already installed on all the cluster nodes, we'll have to
    // localize hive-exec.jar as well.
    appJarLr = createJarLocalResource(utils.getExecJarPathLocal());
    // configuration for the application master
    final Map<String, LocalResource> commonLocalResources = new HashMap<String, LocalResource>();
    commonLocalResources.put(utils.getBaseName(appJarLr), appJarLr);
    for (LocalResource lr : localizedResources) {
        commonLocalResources.put(utils.getBaseName(lr), lr);
    }
    if (llapMode) {
        // localize llap client jars
        addJarLRByClass(LlapTaskSchedulerService.class, commonLocalResources);
        addJarLRByClass(LlapProtocolClientImpl.class, commonLocalResources);
        addJarLRByClass(LlapProtocolClientProxy.class, commonLocalResources);
        addJarLRByClassName("org.apache.hadoop.registry.client.api.RegistryOperations", commonLocalResources);
    }
    // Create environment for AM.
    Map<String, String> amEnv = new HashMap<String, String>();
    MRHelpers.updateEnvBasedOnMRAMEnv(conf, amEnv);
    // and finally we're ready to create and start the session
    // generate basic tez config
    final TezConfiguration tezConfig = new TezConfiguration(conf);
    // set up the staging directory to use
    tezConfig.set(TezConfiguration.TEZ_AM_STAGING_DIR, tezScratchDir.toUri().toString());
    conf.stripHiddenConfigurations(tezConfig);
    ServicePluginsDescriptor servicePluginsDescriptor;
    Credentials llapCredentials = null;
    if (llapMode) {
        if (UserGroupInformation.isSecurityEnabled()) {
            llapCredentials = new Credentials();
            llapCredentials.addToken(LlapTokenIdentifier.KIND_NAME, getLlapToken(user, tezConfig));
        }
        // TODO Change this to not serialize the entire Configuration - minor.
        UserPayload servicePluginPayload = TezUtils.createUserPayloadFromConf(tezConfig);
        // we need plugins to handle llap and uber mode
        servicePluginsDescriptor = ServicePluginsDescriptor.create(true, new TaskSchedulerDescriptor[] { TaskSchedulerDescriptor.create(LLAP_SERVICE, LLAP_SCHEDULER).setUserPayload(servicePluginPayload) }, new ContainerLauncherDescriptor[] { ContainerLauncherDescriptor.create(LLAP_SERVICE, LLAP_LAUNCHER) }, new TaskCommunicatorDescriptor[] { TaskCommunicatorDescriptor.create(LLAP_SERVICE, LLAP_TASK_COMMUNICATOR).setUserPayload(servicePluginPayload) });
    } else {
        servicePluginsDescriptor = ServicePluginsDescriptor.create(true);
    }
    // container prewarming. tell the am how many containers we need
    if (HiveConf.getBoolVar(conf, ConfVars.HIVE_PREWARM_ENABLED)) {
        int n = HiveConf.getIntVar(conf, ConfVars.HIVE_PREWARM_NUM_CONTAINERS);
        n = Math.max(tezConfig.getInt(TezConfiguration.TEZ_AM_SESSION_MIN_HELD_CONTAINERS, TezConfiguration.TEZ_AM_SESSION_MIN_HELD_CONTAINERS_DEFAULT), n);
        tezConfig.setInt(TezConfiguration.TEZ_AM_SESSION_MIN_HELD_CONTAINERS, n);
    }
    setupSessionAcls(tezConfig, conf);
    final TezClient session = TezClient.newBuilder("HIVE-" + sessionId, tezConfig).setIsSession(true).setLocalResources(commonLocalResources).setCredentials(llapCredentials).setServicePluginDescriptor(servicePluginsDescriptor).build();
    LOG.info("Opening new Tez Session (id: " + sessionId + ", scratch dir: " + tezScratchDir + ")");
    TezJobMonitor.initShutdownHook();
    if (!isAsync) {
        startSessionAndContainers(session, conf, commonLocalResources, tezConfig, false);
        this.session = session;
    } else {
        FutureTask<TezClient> sessionFuture = new FutureTask<>(new Callable<TezClient>() {

            @Override
            public TezClient call() throws Exception {
                try {
                    return startSessionAndContainers(session, conf, commonLocalResources, tezConfig, true);
                } catch (Throwable t) {
                    LOG.error("Failed to start Tez session", t);
                    throw (t instanceof Exception) ? (Exception) t : new Exception(t);
                }
            }
        });
        new Thread(sessionFuture, "Tez session start thread").start();
        // We assume here nobody will try to get session before open() returns.
        this.console = console;
        this.sessionFuture = sessionFuture;
    }
}