最近集群准备升级到cdh5.2.0，并使用tez，在测试集群cdh5.2.0已经稳定运行了很长时间，因此开始折腾hive on tez了，期间遇到不少问题，这里记录下。

hive on tez的部署比较简单，可以参考wiki.主要注意几个地方

1.编译的时候

mvn clean package -Dtar -DskipTests=true -Dmaven.javadoc.skip=true

2.需要将tez相关的包upload到hdfs中，并设置tez-site.xml

  <property><name>tez.lib.uris</name><value>${fs.defaultFS}/tez,${fs.defaultFS}/tez/lib</value></property>

设置mapred-site.xml

  <property><name>mapreduce.framework.name</name><value>yarn-tez</value></property>

3.注意更新hadoop-env.sh中classpath的设置

export TEZ_HOME=/home/vipshop/platform/tez
for jar in `ls $TEZ_HOME |grep jar`; doexport HADOOP_CLASSPATH=$HADOOP_CLASSPATH:$TEZ_HOME/$jar
done
for jar in `ls $TEZ_HOME/lib`; doexport HADOOP_CLASSPATH=$HADOOP_CLASSPATH:$TEZ_HOME/lib/$jar
done

否则会报如下错误（加载不到对应的tez相关类，导致Cluster 初始化时失败）:

java.io.IOException: Cannot initialize Cluster. Please check your configuration for mapreduce.framework.name and the correspond server addresses.at org.apache.hadoop.mapreduce.Cluster.initialize(Cluster.java:120)at org.apache.hadoop.mapreduce.Cluster.<init>(Cluster.java:82)at org.apache.hadoop.mapreduce.Cluster.<init>(Cluster.java:75)at org.apache.hadoop.mapreduce.Job$9.run(Job.java:1265)at org.apache.hadoop.mapreduce.Job$9.run(Job.java:1261)at java.security.AccessController.doPrivileged(Native Method)at javax.security.auth.Subject.doAs(Subject.java:396)at org.apache.hadoop.security.UserGroupInformation.doAs(UserGroupInformation.java:1614)at org.apache.hadoop.mapreduce.Job.connect(Job.java:1260)at org.apache.hadoop.mapreduce.Job.submit(Job.java:1289)at org.apache.hadoop.mapreduce.Job.waitForCompletion(Job.java:1313)at org.apache.hadoop.mapreduce.SleepJob.run(SleepJob.java:261)at org.apache.hadoop.util.ToolRunner.run(ToolRunner.java:70)at org.apache.hadoop.mapreduce.SleepJob.main(SleepJob.java:194)at sun.reflect.NativeMethodAccessorImpl.invoke0(Native Method)at sun.reflect.NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.java:39)at sun.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:25)at java.lang.reflect.Method.invoke(Method.java:597)at org.apache.hadoop.util.ProgramDriver$ProgramDescription.invoke(ProgramDriver.java:72)at org.apache.hadoop.util.ProgramDriver.run(ProgramDriver.java:145)at org.apache.hadoop.test.MapredTestDriver.run(MapredTestDriver.java:118)at org.apache.hadoop.test.MapredTestDriver.main(MapredTestDriver.java:126)at sun.reflect.NativeMethodAccessorImpl.invoke0(Native Method)at sun.reflect.NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.java:39)at sun.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:25)at java.lang.reflect.Method.invoke(Method.java:597)at org.apache.hadoop.util.RunJar.main(RunJar.java:212)

部署完毕后，使用hadoop jar提交tez job运行正常，测试hive on tez:

hive -hiveconf hive.execution.engine=tez -hiveconf hive.root.logger=DEBUG,console

出现如下报错：

Exception in thread "main" java.lang.NoSuchMethodError:
org.apache.tez.mapreduce.hadoop.MRHelpers.updateEnvironmentForMRAM(Lorg/apache/hadoop/conf/Configuration;Ljava/util/Map;)Vat org.apache.hadoop.hive.ql.exec.tez.TezSessionState.open(TezSessionState.java:182)at org.apache.hadoop.hive.ql.exec.tez.TezSessionState.open(TezSessionState.java:123)at org.apache.hadoop.hive.ql.session.SessionState.start(SessionState.java:355)at org.apache.hadoop.hive.cli.CliDriver.run(CliDriver.java:681)at org.apache.hadoop.hive.cli.CliDriver.main(CliDriver.java:625)at sun.reflect.NativeMethodAccessorImpl.invoke0(Native Method)at sun.reflect.NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.java:39)at sun.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:25)at java.lang.reflect.Method.invoke(Method.java:597)at org.apache.hadoop.util.RunJar.main(RunJar.java:212)

从堆栈上来看是由于session初始化异常导致，

org.apache.hadoop.hive.cli.CliDriver.main->org.apache.hadoop.hive.cli.CliDriver.run->
org.apache.hadoop.hive.ql.session.SessionState.start

在SessionState.start方法中:

 if (HiveConf.getVar(startSs.getConf(), HiveConf.ConfVars.HIVE_EXECUTION_ENGINE)  .equals( "tez") && (startSs.isHiveServerQuery == false)) { //如果设置hive.execution.engine为tez（默认为mr）try {if (startSs.tezSessionState == null) {startSs.tezSessionState = new TezSessionState(startSs.getSessionId()); //实例化一个TezSessionState对象}startSs.tezSessionState.open(startSs.conf ); //调用TezSessionState.open方法} catch (Exception e) {throw new RuntimeException(e);}} else {LOG.info( "No Tez session required at this point. hive.execution.engine=mr.");}

TezSessionState.open中，首先使用createTezDir创建临时文件目录

    // create the tez tmp dirtezScratchDir = createTezDir(sessionId);String dir = tezScratchDir.toString();// Localize resources to session scratch dirlocalizedResources = utils.localizeTempFilesFromConf(dir, conf); //DagUtils.localizeTempFilesFromConf方法List<LocalResource> handlerLr = utils.localizeTempFiles(dir, conf, additionalFiles); // DagUtils.localizeTempFiles方法if (handlerLr != null) {if (localizedResources == null) {localizedResources = handlerLr;} else {localizedResources.addAll(handlerLr);}additionalFilesNotFromConf = new HashSet<String>();for (String originalFile : additionalFiles) {additionalFilesNotFromConf.add(originalFile);}}// generate basic tez configTezConfiguration tezConfig = new TezConfiguration(conf); //然后实例化一个TezConfiguration对象tezConfig.set(TezConfiguration.TEZ_AM_STAGING_DIR, tezScratchDir.toUri().toString()); //设置tez的staging目录，设置项为tez.staging-dir，默认值为/tmp/tez/staging
//这里默认最终为"/tmp/hive-" + System. getProperty( "user.name")/_tez_session_dir/sessionIdappJarLr = createJarLocalResource(utils.getExecJarPathLocal()); //localize hive-exec.jar// configuration for the application masterMap<String, LocalResource> commonLocalResources = new HashMap<String, LocalResource>();commonLocalResources.put( utils.getBaseName( appJarLr), appJarLr );if (localizedResources != null) {for (LocalResource lr : localizedResources) {commonLocalResources.put( utils.getBaseName(lr), lr);}}// Create environment for AM.Map<String, String> amEnv = new HashMap<String, String>();MRHelpers.updateEnvironmentForMRAM(conf, amEnv); //调用MRHelpers类的updateEnvironmentForMRAM方法

对于org.apache.tez.mapreduce.hadoop.MRHelpers类来说，在0.5.0中，这个updateEnvironmentForMRAM方法是不存在的，对应存在updateEnvBasedOnMRTaskEnv（配置Mappers和Reducers的环境变量）和updateEnvBasedOnMRAMEnv（配置am的环境变量）

public static void updateEnvBasedOnMRAMEnv(Configuration conf, Map<String, String> environment) {TezYARNUtils.appendToEnvFromInputString(environment, conf.get(MRJobConfig.MR_AM_ADMIN_USER_ENV),File.pathSeparator);TezYARNUtils.appendToEnvFromInputString(environment, conf.get(MRJobConfig.MR_AM_ENV),File.pathSeparator);
}

而在0.4.1-incubating中是有updateEnvironmentForMRAM这个方法的：

public static void updateEnvironmentForMRAM(Configuration conf, Map<String, String> environment) {TezYARNUtils.setEnvFromInputString(environment, conf.get(MRJobConfig.MR_AM_ADMIN_USER_ENV),File.pathSeparator);TezYARNUtils.setEnvFromInputString(environment, conf.get(MRJobConfig.MR_AM_ENV),File.pathSeparator);
}

对应的hive中：
hive0.13中：

    // Create environment for AM.Map<String, String> amEnv = new HashMap<String, String>();MRHelpers.updateEnvironmentForMRAM(conf, amEnv);

hive0.14中：

    // Create environment for AM.      Map<String, String> amEnv = new HashMap<String, String>();      MRHelpers.updateEnvBasedOnMRAMEnv(conf, amEnv);

可以看到0.4.x到0.5.x版本的tez api变动比较大，0.5.x的tez已经和hive0.13.x不能兼容了，要想使用tez-0.5.x版本，必须使用hive0.14.x版本。        
在github下载hive0.14的源码，编译并测试运行hive on tez:
https://codeload.github.com/apache/hive/zip/branch-0.14

mvn clean package -DskipTests -Phadoop-2 -Pdist