HDP上安装impala

Impala是Cloudera公司主导开发的新型查询系统，它提供SQL语义，能查询存储在Hadoop的HDFS和HBase中的PB级大数据。Impala提供更快的查询速度，性能上号称比Hive快3~10倍。Impala是开源的，但通常都是经过cloudera manager或者在CDH版本上安装，今天主要介绍的是在HDP版本上的安装。

版本

Impala对于Hadoop的版本要求很高，如今说明一下当前安装的版本信息

Impala 2.5

HDP 2.2.8.0 基于Hadoop2.6

安装步骤

1. 在/etc/yum.repo.d 中建立impala.repo

[cloudera-cdh5]
# Packages for Cloudera's Distribution for Hadoop, Version 5, on RedHat or CentOS 6 x86_64
name=Cloudera's Distribution for Hadoop, Version 5
baseurl=https://archive.cloudera.com/cdh5/redhat/6/x86_64/cdh/5/
gpgkey =https://archive.cloudera.com/cdh5/redhat/6/x86_64/cdh/RPM-GPG-KEY-cloudera    
gpgcheck = 1

2. 执行yum命令

yum install  impala-server impala-catalog impala-state-store impala-shell

3.更改impala连接库信息

ln -sf  /usr/hdp/2.2.8.0-3150/hadoop/hadoop-common.jar /usr/lib/impala/lib/hadoop-common.jar
ln -sf  /usr/hdp/2.2.8.0-3150/hadoop/hadoop-annotations.jar /usr/lib/impala/lib/hadoop-annotations.jar
ln -sf  /usr/hdp/2.2.8.0-3150/hadoop/hadoop-auth.jar /usr/lib/impala/lib/hadoop-auth.jar
ln -sf  /usr/hdp/2.2.8.0-3150/hadoop/hadoop-aws.jar /usr/lib/impala/lib/hadoop-aws.jar
ln -sf  /usr/hdp/2.2.8.0-3150/hadoop-hdfs/hadoop-hdfs.jar /usr/lib/impala/lib/hadoop-hdfs.jar
ln -sf  /usr/hdp/2.2.8.0-3150/hadoop-mapreduce/hadoop-mapreduce-client-common.jar /usr/lib/impala/lib/hadoop-mapreduce-client-common.jar
ln -sf  /usr/hdp/2.2.8.0-3150/hadoop-mapreduce/hadoop-mapreduce-client-core.jar /usr/lib/impala/lib/hadoop-mapreduce-client-core.jar
ln -sf  /usr/hdp/2.2.8.0-3150/hadoop-mapreduce/hadoop-mapreduce-client-jobclient.jar /usr/lib/impala/lib/hadoop-mapreduce-client-jobclient.jar
ln -sf  /usr/hdp/2.2.8.0-3150/hadoop-mapreduce/hadoop-mapreduce-client-shuffle.jar /usr/lib/impala/lib/hadoop-mapreduce-client-shuffle.jar
ln -sf  /usr/hdp/2.2.8.0-3150/hadoop-yarn/hadoop-yarn-api.jar /usr/lib/impala/lib/hadoop-yarn-api.jar
ln -sf  /usr/hdp/2.2.8.0-3150/hadoop-yarn/hadoop-yarn-client.jar /usr/lib/impala/lib/hadoop-yarn-client.jar
ln -sf  /usr/hdp/2.2.8.0-3150/hadoop-yarn/hadoop-yarn-common.jar /usr/lib/impala/lib/hadoop-yarn-common.jar
ln -sf  /usr/hdp/2.2.8.0-3150/hadoop-yarn/hadoop-yarn-server-applicationhistoryservice.jar /usr/lib/impala/lib/hadoop-yarn-server-applicationhistoryservice.jar
ln -sf  /usr/hdp/2.2.8.0-3150/hadoop-yarn/hadoop-yarn-server-common.jar /usr/lib/impala/lib/hadoop-yarn-server-common.jar
ln -sf  /usr/hdp/2.2.8.0-3150/hadoop-yarn/hadoop-yarn-server-nodemanager.jar /usr/lib/impala/lib/hadoop-yarn-server-nodemanager.jar
ln -sf  /usr/hdp/2.2.8.0-3150/hadoop-yarn/hadoop-yarn-server-resourcemanager.jar /usr/lib/impala/lib/hadoop-yarn-server-resourcemanager.jar
ln -sf  /usr/hdp/2.2.8.0-3150/hadoop-yarn/hadoop-yarn-server-web-proxy.jar /usr/lib/impala/lib/hadoop-yarn-server-web-proxy.jar
ln -sf /usr/hdp/2.2.8.0-3150/hadoop/lib/native/libhadoop.so /usr/lib/impala/lib/libhadoop.so
ln -sf /usr/hdp/2.2.8.0-3150/hadoop/lib/native/libhadoop.so.1.0.0 /usr/lib/impala/lib/libhadoop.so.1.0.0
ln -sf /usr/hdp/2.2.8.0-3150/usr/lib/libhdfs.so /usr/lib/impala/lib/libhdfs.so
ln -sf /usr/hdp/2.2.8.0-3150/usr/lib/libhdfs.so.0.0.0 /usr/lib/impala/lib/libhdfs.so.0.0.0
ln -sf /usr/hdp/2.2.8.0-3150/hive/lib/hive-ant.jar /usr/lib/impala/lib/hive-ant.jar 
ln -sf /usr/hdp/2.2.8.0-3150/hive/lib/hive-beeline.jar /usr/lib/impala/lib/hive-beeline.jar 
ln -sf /usr/hdp/2.2.8.0-3150/hive/lib/hive-common.jar /usr/lib/impala/lib/hive-common.jar 
ln -sf /usr/hdp/2.2.8.0-3150/hive/lib/hive-exec.jar /usr/lib/impala/lib/hive-exec.jar 
ln -sf /usr/hdp/2.2.8.0-3150/hive/lib/hive-hbase-handler.jar /usr/lib/impala/lib/hive-hbase-handler.jar 
ln -sf /usr/hdp/2.2.8.0-3150/hive/lib/hive-metastore.jar /usr/lib/impala/lib/hive-metastore.jar 
ln -sf /usr/hdp/2.2.8.0-3150/hive/lib/hive-serde.jar /usr/lib/impala/lib/hive-serde.jar 
ln -sf /usr/hdp/2.2.8.0-3150/hive/lib/hive-service.jar /usr/lib/impala/lib/hive-service.jar 
ln -sf /usr/hdp/2.2.8.0-3150/hive/lib/hive-shims-common.jar /usr/lib/impala/lib/hive-shims-common.jar 
ln -sf /usr/hdp/2.2.8.0-3150/hive/lib/hive-shims-scheduler.jar /usr/lib/impala/lib/hive-shims-scheduler.jar 
ln -sf /usr/hdp/2.2.8.0-3150/hive/lib/hive-shims.jar /usr/lib/impala/lib/hive-shims.jar 
ln -sf /usr/hdp/2.2.8.0-3150/zookeeper/zookeeper.jar /usr/lib/impala/lib/zookeeper.jar

4. 增长JAVA_HOME

vi /etc/default/bigtop-utils 

增长 EXPORT JAVA_HOME= /usr/jdk64/jdk1.7.0_67

5. 修改配置文件

  增长 /etc/hadoop/conf/core-site.xml

<property>
    <name>dfs.client.read.shortcircuit</name> <value>true</value>
</property>

<property>
    <name>dfs.client.read.shortcircuit.skip.checksum</name>
        <value>false</value>
</property>

<property> 
    <name>dfs.datanode.hdfs-blocks-metadata.enabled</name> 
    <value>true</value>
</property>

  增长 /etc/hadoop/conf/hdfs-site.xml

<property>
    <name>dfs.datanode.hdfs-blocks-metadata.enabled</name> 
    <value>true</value>
</property>
<property> 
    <name>dfs.block.local-path-access.user</name> 
    <value>impala</value>
</property>
<property>
    <name>dfs.client.file-block-storage-locations.timeout.millis</name>
    <value>60000</value>
</property>

  执行拷贝

cp /etc/hadoop/conf/*.xml /etc/impala/conf
cp /etc/hive/conf/hive-site.xml /etc/impala/conf

  不然报错

E0721 15:52:54.265014 11246 impala-server.cc:247] Unsupported default filesystem. The default filesystem must be a DistributedFileSystem but the configured default filesystem is LocalFileSystem. fs.defaultFS (file:///) might be set incorrectly.ERROR: block location tracking is not properly enabled because
  - dfs.datanode.hdfs-blocks-metadata.enabled is not enabled.
  - dfs.client.file-block-storage-locations.timeout.millis is too low. It should be at least 10 seconds.

6. 修改 /etc/default/impala

   修改IMPALA_CATALOG_SERVICE_HOST，IMPALA_STATE_STORE_HOST 两个变量为你启动catalog，state_store两个服务的host

7. 从网上下载 htrace-core-3.0.4.jar更名htrace-core-2.00.jar 替换 /usr/lib/impala/lib/htrace-core-2.00.jar

不然会报错

E0711 13:55:00.451584  6925 impala-server.cc:247] NoClassDefFoundError: org/htrace/Trace
CAUSED BY: ClassNotFoundException: org.htrace.Trace
E0711 13:55:00.451653  6925 impala-server.cc:249] Aborting Impala Server startup due to improper configuration

8 一个host启动 catalog, state_store

   全部host启动 impalad

 

提示：

   安装过程当中须要学会查看/var/log/impala的日志，能够查看当前的错误信息

 

咱们已经编写好能够用的ambari组件进行impala安装，组件地址见下面的连接

https://github.com/cas-bigdatalab/ambari-impala-service