hive-tez

1 参考：
https://www.freesion.com/article/9435149734/
https://blog.csdn.net/Shea1992/article/details/101041244
https://www.jianshu.com/p/9fb9f32e1f0f
https://www.jianshu.com/p/45c95a51a8c2
https://blog.csdn.net/weixin_43941899/article/details/105787688
2 我的环境：
hadoop版本：3.0.0-cdh6.3.2
linux环境:centos7.6
jdk1.8.0_131
3 安装maven
3.1 下载maven
https://maven.apache.org/download.cgi

3.2上传解压
tar -zxvf apache-maven-3.6.3-bin.tar.gz -C /data/module

3.3 配置 MVN_HOMR
[root@cluster2-slave2 ~]# vim /etc/profile

export MVN_HOME=/data/module/apache-maven-3.6.3
export PATH=$PATH:$MVN_HOME/bin
[root@cluster2-slave2 ~]# source /etc/profile

3.4 验证maven
[root@cluster2-slave2 ~]# mvn -v

4 安装protobuf-2.5.0.tar.gz
4.1 下载
只能是2.5.0这个版本

因为后面安装tez0.91的时候加压后在pom.xml里可以看到，要求就是2.5.0的

hadoop使用protocol buffer进行通信，需要下载和安装 protobuf-2.5.0.tar.gz。

但是现在 protobuf-2.5.0.tar.gz已经无法在官网 https://code.google.com/p/protobuf/downloads/list中 下载了

我在百度网盘找到了下载链接这里附上下载链接

链接：https://pan.baidu.com/s/1hm7D2_wxIxMKbN9xnlYWuA
提取码：haz4
复制这段内容后打开百度网盘手机App，操作更方便哦
4.2 上传解压
tar -zxvf protobuf-2.5.0.tar.gz

4.3 configure校验
cd protobuf-2.5.0/

[root@cluster2-slave2 protobuf-2.5.0]# ./configure
第一次我校验失败：

checking whether to enable maintainer-specific portions of Makefiles… yes
checking build system type… x86_64-unknown-linux-gnu
checking host system type… x86_64-unknown-linux-gnu
checking target system type… x86_64-unknown-linux-gnu
checking for a BSD-compatible install… /usr/bin/install -c
checking whether build environment is sane… yes
checking for a thread-safe mkdir -p… /usr/bin/mkdir -p
checking for gawk… gawk
checking whether make sets $(MAKE)… yes
checking for gcc… gcc
checking whether the C compiler works… yes
checking for C compiler default output file name… a.out
checking for suffix of executables…
checking whether we are cross compiling… no
checking for suffix of object files… o
checking whether we are using the GNU C compiler… yes
checking whether gcc accepts -g… yes
checking for gcc option to accept ISO C89… none needed
checking for style of include used by make… GNU
checking dependency style of gcc… gcc3
checking for g++… no
checking for c++… no
checking for gpp… no
checking for aCC… no
checking for CC… no
checking for cxx… no
checking for cc++… no
checking for cl.exe… no
checking for FCC… no
checking for KCC… no
checking for RCC… no
checking for xlC_r… no
checking for xlC… no
checking whether we are using the GNU C++ compiler… no
checking whether g++ accepts -g… no
checking dependency style of g++… none
checking how to run the C++ preprocessor… /lib/cpp
configure: error: in /data/software/protobuf-2.5.0': configure: error: C++ preprocessor "/lib/cpp" fails sanity check Seeconfig.log’ for more details
错误信息：

configure: error: in `/data/software/protobuf-2.5.0’:
configure: error: C++ preprocessor “/lib/cpp” fails sanity check

问题的根源是缺少必要的C++库。如果是CentOS系统，运行，如下命令解决

yum install glibc-headers

yum install gcc-c++
结束后日志

…省略…

Installed:
gcc-c++.x86_64 0:4.8.5-39.el7

Dependency Installed:
libstdc+±devel.x86_64 0:4.8.5-39.el7

Dependency Updated:
libstdc++.x86_64 0:4.8.5-39.el7

Complete!
再次检查通过

[root@cluster2-slave2 protobuf-2.5.0]# ./configure
checking whether to enable maintainer-specific portions of Makefiles… yes
checking build system type… x86_64-unknown-linux-gnu
checking host system type… x86_64-unknown-linux-gnu
checking target system type… x86_64-unknown-linux-gnu
checking for a BSD-compatible install… /usr/bin/install -c
checking whether build environment is sane… yes
checking for a thread-safe mkdir -p… /usr/bin/mkdir -p

。。。。。。。。省略。。。。。。。。

checking whether the g++ linker (/usr/bin/ld -m elf_x86_64) supports shared libraries… yes
checking dynamic linker characteristics… (cached) GNU/Linux ld.so
checking how to hardcode library paths into programs… immediate
checking for python… /usr/bin/python
checking for the pthreads library -lpthreads… no
checking whether pthreads work without any flags… no
checking whether pthreads work with -Kthread… no
checking whether pthreads work with -kthread… no
checking for the pthreads library -llthread… no
checking whether pthreads work with -pthread… yes
checking for joinable pthread attribute… PTHREAD_CREATE_JOINABLE
checking if more special flags are required for pthreads… no
checking whether to check for GCC pthread/shared inconsistencies… yes
checking whether -pthread is sufficient with -shared… yes
configure: creating ./config.status
config.status: creating Makefile
config.status: creating scripts/gtest-config
config.status: creating build-aux/config.h
config.status: executing depfiles commands
config.status: executing libtool commands
[root@cluster2-slave2 protobuf-2.5.0]#
4.4 make
[root@cluster2-slave2 protobuf-2.5.0]# make
…省略…

libtool: link: ranlib .libs/libprotobuf-lite.a
libtool: link: ( cd “.libs” && rm -f “libprotobuf-lite.la” && ln -s “…/libprotobuf-lite.la” “libprotobuf-lite.la” )
make[3]: Leaving directory /data/software/protobuf-2.5.0/src' make[2]: Leaving directory/data/software/protobuf-2.5.0/src’
make[1]: Leaving directory `/data/software/protobuf-2.5.0’
4.5 make install
[root@cluster2-slave2 protobuf-2.5.0]# make install
…省略…

/usr/bin/mkdir -p ‘/usr/local/include/google/protobuf/io’
/usr/bin/install -c -m 644 google/protobuf/io/coded_stream.h google/protobuf/io/gzip_stream.h google/protobuf/io/printer.h google/protobuf/io/tokenizer.h google/protobuf/io/zero_copy_stream.h google/protobuf/io/zero_copy_stream_impl.h google/protobuf/io/zero_copy_stream_impl_lite.h ‘/usr/local/include/google/protobuf/io’
make[3]: Leaving directory /data/software/protobuf-2.5.0/src' make[2]: Leaving directory/data/software/protobuf-2.5.0/src’
make[1]: Leaving directory `/data/software/protobuf-2.5.0/src’
4.6 验证protobuf
[root@cluster2-slave2 protobuf-2.5.0]# protoc --version
libprotoc 2.5.0
可见安装成功

5 安装 Tez
5.1 下载
http://www.apache.org/dyn/closer.lua/tez/0.9.1/

下载源码自己根据自己的环境编译

5.2 上传解压
[root@cluster2-slave2 software]# tar -zxvf apache-tez-0.9.1-src.tar.gz -C …/module/

5.3 修改pom.xml
有四处修改：

（1）hadoop.version版本对应修改
确认自己的hadoop版本

修改如下

（2）repository.cloudera

(3) 新增pluginRepository.cloudera

(4) 注释不必要的东西减少下载和编译出错概率

完整pom.xml 如下

<?xml version="1.0" encoding="UTF-8"?>

4.0.0
org.apache.tez
tez
pom
0.9.1
tez

The Apache Software License, Version 2.0 http://www.apache.org/licenses/LICENSE-2.0.txt Apache Software Foundation http://www.apache.org ${distMgmtStagingId} ${distMgmtStagingName} ${distMgmtStagingUrl} ${distMgmtSnapshotsId} ${distMgmtSnapshotsName} ${distMgmtSnapshotsUrl} ${distMgmtSnapshotsId} ${distMgmtSnapshotsName} ${distMgmtSnapshotsUrl} cloudera https://repository.cloudera.com/artifactory/cloudera-repos/ Cloudera Repositories false maven2-repository.atlassian Atlassian Maven Repository https://maven.atlassian.com/repository/public default ${distMgmtSnapshotsId} ${distMgmtSnapshotsName} ${distMgmtSnapshotsUrl} default cloudera Cloudera Repositories https://repository.cloudera.com/artifactory/cloudera-repos/ hadoop-shim tez-api tez-common tez-runtime-library tez-runtime-internals tez-mapreduce tez-examples tez-tests tez-dag tez-plugins tez-tools hadoop-shim-impls tez-dist docs 5.4 maven编译 注意：我们不需要javadoc和test所以可以编译时跳过

所以用下面领命编译

在pom.xml同级目录下执行

mvn clean package -Dmaven.javadoc.skip=true -Dmaven.test.skip=true
开始编译

5.5 遇到问题
第一次编译要下载很多包有可能中途下载失败

比如下面就是一次包的依赖没有下载成功报错了

重试后所有依赖的jar下载成功

但是编译还遇到了一个问题

报ApplicationReport.newInstance() 89行异常

/data/module/apache-tez-0.9.1-src/tez-mapreduce/src/main/java/org/apache/tez/mapreduce/clienttRunningJob.java:[89,29] no suitable method found for newInstance(org.apache.hadoop.yarn.api.records.ApplicationId,org.apache.hadoop.yarn.api.records.ApplicationAttemptId,java.lang.String,java.lang.String,java.lang.String,java.lang.String,int,,org.apache.hadoop.yarn.api.records.YarnApplicationState,java.lang.String,java.lang.String,int,int,org.apache.hadoop.yarn.api.records.FinalApplicationStatus,,java.lang.String,float,java.lang.String,)

解决办法：更换ApplicationReport.newInstance()的另一个方法

更换源代码：

return ApplicationReport.newInstance(unknownAppId, unknownAttemptId, “N/A”,
“N/A”, “N/A”, “N/A”, 0, null, YarnApplicationState.NEW, “N/A”, “N/A”,
0, 0, FinalApplicationStatus.UNDEFINED, null, “N/A”, 0.0f, “TEZ_MRR”, null)
更换为新代码：

return ApplicationReport.newInstance(unknownAppId, unknownAttemptId, “N/A”,
“N/A”, “N/A”, “N/A”, 0, null, YarnApplicationState.NEW, “N/A”, “N/A”,
0, 0, 0, FinalApplicationStatus.UNDEFINED, null, “N/A”, 0.0f, “TEZ_MRR”, null);
vim tez-mapreduce/src/main/java/org/apache/tez/mapreduce/clienttRunningJob.java

然后继续编译

mvn clean package -Dmaven.javadoc.skip=true -Dmaven.test.skip=true
5.6 最后编译成功

编译后的文件在tez-dist/target下面

我的在：/data/module/apache-tez-0.9.1-src/tez-dist/target

5.7 整个tez到hdfs
hdfs上创建相应的tez目录

[root@cluster2-slave2 target]# hadoop fs -mkdir /user/tez
上传tez-0.9.1.tar.gz到 /user/tez目录

[root@cluster2-slave2 target]# hadoop fs -put tez-0.9.1.tar.gz /user/tez

5.8 整个tez到hive
（1）进入CDH lib目录

cd /opt/cloudera/parcels/CDH-6.3.2-1.cdh6.3.2.p0.1605554/lib
(2) 创建tez相关目录目录

[root@cluster2-slave2 lib]# mkdir -p tez/conf
（3）创建 tez-site.xml文件

[root@cluster2-slave2 lib]# cd tez/conf/
[root@cluster2-slave2 conf]# vim tez-site.xml


tez.lib.uris
${fs.defaultFS}/user/tez/tez-0.9.1.tar.gz


tez.use.cluster.hadoop-libs
false

（4）tez-0.9.1-minimal拷贝到tez中

下面的目录

/opt/cloudera/parcels/CDH-6.3.2-1.cdh6.3.2.p0.1605554/lib/tez

[root@cluster2-slave2 tez-0.9.1-minimal]# cp ./*.jar /opt/cloudera/parcels/CDH-6.3.2-1.cdh6.3.2.p0.1605554/lib/tez
[root@cluster2-slave2 tez-0.9.1-minimal]# cp -r lib /opt/cloudera/parcels/CDH-6.3.2-1.cdh6.3.2.p0.1605554/lib/tez

（5）分发tez这个文件到各个节点

[root@cluster2-slave2 lib]# scp -r tez root@cluster2-slave1:/opt/cloudera/parcels/CDH-6.3.2-1.cdh6.3.2.p0.1605554/lib
[root@cluster2-slave2 lib]# scp -r tez root@cluster2-master:/opt/cloudera/parcels/CDH-6.3.2-1.cdh6.3.2.p0.1605554/lib

6 配置hvie环境

6.1 配置环境
在cdh找到hive客户端配置

HADOOP_CLASSPATH=/opt/cloudera/parcels/CDH-6.3.2-1.cdh6.3.2.p0.1605554/lib/tez/conf:/opt/cloudera/parcels/CDH-6.3.2-1.cdh6.3.2.p0.1605554/lib/tez/:/opt/cloudera/parcels/CDH-6.3.2-1.cdh6.3.2.p0.1605554/lib/tez/lib/

然后保存并部署客户端配置，使生效

6.2 测试效果
报错

---

    VERTICES      MODE        STATUS  TOTAL  COMPLETED  RUNNING  PENDING  FAILED  KILLED  

---

Map 1 container FAILED -1 0 0 -1 0 0
Reducer 2 container KILLED 1 0 0 1 0 0

VERTICES: 00/02 [>>--------------------------] 0% ELAPSED TIME: 0.15 s

20/08/07 13:01:23 INFO SessionState: Map 1: -/- Reducer 2: 0/1
Status: Failed
20/08/07 13:01:23 ERROR SessionState: Status: Failed
Vertex failed, vertexName=Map 1, vertexId=vertex_1596710469388_0001_1_00, diagnostics=[Vertex vertex_1596710469388_0001_1_00 [Map 1] killed/failed due to:INIT_FAILURE, Fail to create InputInitializerManager, org.apache.tez.dag.api.TezReflectionException: Unable to instantiate class with 1 arguments: org.apache.hadoop.hive.ql.exec.tez.HiveSplitGenerator
at org.apache.tez.common.ReflectionUtils.getNewInstance(ReflectionUtils.java:71)
at org.apache.tez.common.ReflectionUtils.createClazzInstance(ReflectionUtils.java:89)
at org.apache.tez.dag.app.dag.RootInputInitializerManager$1.run(RootInputInitializerManager.java:152)
at org.apache.tez.dag.app.dag.RootInputInitializerManager$1.run(RootInputInitializerManager.java:148)
at java.security.AccessController.doPrivileged(Native Method)
at javax.security.auth.Subject.doAs(Subject.java:422)
at org.apache.hadoop.security.UserGroupInformation.doAs(UserGroupInformation.java:1875)
at org.apache.tez.dag.app.dag.RootInputInitializerManager.createInitializer(RootInputInitializerManager.java:148)
at org.apache.tez.dag.app.dag.RootInputInitializerManager.runInputInitializers(RootInputInitializerManager.java:121)
at org.apache.tez.dag.app.dag.impl.VertexImpl.setupInputInitializerManager(VertexImpl.java:4101)
at org.apache.tez.dag.app.dag.impl.VertexImpl.access$3100(VertexImpl.java:205)atorg.apache.tez.dag.app.dag.impl.VertexImpl$InitTransition.handleInitEvent(VertexImpl.java:2912)
at org.apache.tez.dag.app.dag.impl.VertexImpl$InitTransition.transition(VertexImpl.java:2859)atorg.apache.tez.dag.app.dag.impl.VertexImpl$InitTransition.transition(VertexImpl.java:2841)
at org.apache.hadoop.yarn.state.StateMachineFactory$MultipleInternalArc.doTransition(StateMachineFactory.java:385)
at org.apache.hadoop.yarn.state.StateMachineFactory.doTransition(StateMachineFactory.java:302)
at org.apache.hadoop.yarn.state.StateMachineFactory.access$500(StateMachineFactory.java:46)atorg.apache.hadoop.yarn.state.StateMachineFactory$InternalStateMachine.doTransition(StateMachineFactory.java:487)
at org.apache.tez.state.StateMachineTez.doTransition(StateMachineTez.java:59)
at org.apache.tez.dag.app.dag.impl.VertexImpl.handle(VertexImpl.java:1939)
at org.apache.tez.dag.app.dag.impl.VertexImpl.handle(VertexImpl.java:204)
at org.apache.tez.dag.app.DAGAppMaster$VertexEventDispatcher.handle(DAGAppMaster.java:2317)atorg.apache.tez.dag.app.DAGAppMaster$VertexEventDispatcher.handle(DAGAppMaster.java:2303)
at org.apache.tez.common.AsyncDispatcher.dispatch(AsyncDispatcher.java:180)
at org.apache.tez.common.AsyncDispatcher$1.run(AsyncDispatcher.java:115)atjava.lang.Thread.run(Thread.java:748)Causedby:java.lang.reflect.InvocationTargetExceptionatsun.reflect.NativeConstructorAccessorImpl.newInstance0(NativeMethod)atsun.reflect.NativeConstructorAccessorImpl.newInstance(NativeConstructorAccessorImpl.java:62)atsun.reflect.DelegatingConstructorAccessorImpl.newInstance(DelegatingConstructorAccessorImpl.java:45)atjava.lang.reflect.Constructor.newInstance(Constructor.java:423)atorg.apache.tez.common.ReflectionUtils.getNewInstance(ReflectionUtils.java:68)...25moreCausedby:java.lang.NoClassDefFoundError:com/esotericsoftware/kryo/Serializeratorg.apache.hadoop.hive.ql.exec.Utilities.getBaseWork(Utilities.java:404)atorg.apache.hadoop.hive.ql.exec.Utilities.getMapWork(Utilities.java:317)atorg.apache.hadoop.hive.ql.exec.tez.HiveSplitGenerator.<init>(HiveSplitGenerator.java:131)...30moreCausedby:java.lang.ClassNotFoundException:com.esotericsoftware.kryo.Serializeratjava.net.URLClassLoader.findClass(URLClassLoader.java:381)atjava.lang.ClassLoader.loadClass(ClassLoader.java:424)atsun.misc.Launcher$AppClassLoader.loadClass(Launcher.java:335)
at java.lang.ClassLoader.loadClass(ClassLoader.java:357)
… 33 more
]
6.3 异常处理
ClassNotFoundException: com.esotericsoftware.kryo.Serializer

解决办法 修改：

删除或者重命名hive/auxlib下的hive-exec-2.1.1-cdh6.3.2-core.jar和hive-exec-core.jar

所有节点我都做了

cd /opt/cloudera/parcels/CDH-6.3.2-1.cdh6.3.2.p0.1605554/lib/hive/auxlib
原本这样的

hive-exec-2.1.1-cdh6.3.2-core.jar -> …/…/…/jars/hive-exec-2.1.1-cdh6.3.2-core.jar
hive-exec-core.jar -> hive-exec-2.1.1-cdh6.3.2-core.jar
修改：重命名

mv hive-exec-2.1.1-cdh6.3.2-core.jar hive-exec-2.1.1-cdh6.3.2-core.jar.bck
mv hive-exec-core.jar hive-exec-core.jar.bck
然后重启hive使生效

继续测试

注意必须有set hive.tez.container.size=3020;否则报错

hive> set hive.tez.container.size=3020;
hive> set hive.execution.engine=tez;
执行查询

hive> select sum(salemoney) from crm_sal_shop_sale;
打印出

6.4 日志太多的处理
实际上可以看出这个过程日志太多了。看着不清爽，下面的操作减少日志输出量

这是因为/opt/cloudera/parcels/CDH-6.3.2-1.cdh6.3.2.p0.1605554/lib/tez/lib目录下有

slf4j-api-1.7.10.jar
slf4j-log4j12-1.7.10.jar
两个包

并且现在hive日志级别是 INFO

那么现在通过CDH控制台把hive日志级别设置成ERROR

重启hive使生效

结果查看配置

查看hive查询日志

可见清爽多了

也可以直接删除下面这两个包

slf4j-api-1.7.10.jar
slf4j-log4j12-1.7.10.jar

这个时候yarn上也显示 ApplicationType 是 Tez

6.5 hiveserver2也配置能启用tez

set hive.tez.container.size=3020;
set hive.execution.engine=tez;

select brandid,vipid,sku_id,time from pro30049.add_to_cart_dt_partition where vipid=1197032 and time>=1588953500000 and time<=1589039999000;

6.6 替换hive默认计算引擎
从上面可以看出，使用tez的时候需要临时设置

set hive.tez.container.size=3020;
set hive.execution.engine=tez;

这里我遇到的一个问题本应该

上面配置

如上图修改的时候就可以把配置分发到hive-site.xml 但是我的没有分发

测试下来还是需要添加临时设置

set hive.tez.container.size=3020;
set hive.execution.engine=tez;

后面尝试直接在一个节点cluster2-slave2去修改hive-site.xml文件

重启hive服务

然后测试

不用设置临时启用tez引擎了

最后

以上就是标致金毛为你收集整理的hive-tez的全部内容，希望文章能够帮你解决hive-tez所遇到的程序开发问题。

如果觉得靠谱客网站的内容还不错，欢迎将靠谱客网站推荐给程序员好友。