impala基于CHD,提供针对HDFS,hbase的实时查询，查询语句类似于hive
包括几个组件 Clients:提供Hue, ODBC clients, JDBC clients, and the Impala Shell与impala交互查询 Hive Metastore:保存数据的元数据，让impala知道数据的结构等信息 Cloudera Impala：协调查询在每个datanode上，分发并行查询任务，并将查询返回客户端 HBase and HDFS：存储数据 环境 hadoop-2.0.0-cdh4.1.2 hive-0.9.0-cdh4.1.2 impala利用yum安装 增加yum库 [cloudera-impala] name=Impala baseurl=http://archive.cloudera.com/impala/redhat/5/x86_64/impala/1/ gpgkey = http://archive.cloudera.com/impala/redhat/5/x86_64/impala/RPM-GPG-KEY-cloudera gpgcheck = 1 加至/etc/yum.repos.d目录下 注意cdh与hive及impala需要版本匹配，具体去impala官网去查一下 需要内存比较大,需要64位机器（推荐有点忘了是否支持32位），支持的linux版本也有要求 http://www.cloudera.com/content/cloudera-content/cloudera-docs/Impala/latest/PDF/Installing-and-Using-Impala.pdf 安装CHD4 http://archive.cloudera.com/cdh4/cdh/4/ cdh与hive都可以在这找到 三台机器 master安装 namenode,secondnamenode,ResourceManager, impala-state-store,impala-shell,hive slave1安装 datanode,nodemanager,impala-server, impala-shell slave2安装 datanode,nodemanager,impala-server, impala-shell hadoop配置 在master机器上配置 $HADOOP_HOME/etc/hadoop中的core-site.xml增加 <property> <name>io.native.lib.available</name> <value>true</value> </property> <property> <name>fs.default.name</name> <value>hdfs://master:9000</value> <description>The name of the default file system.Either theliteral string "local" or a host:port for NDFS.</description> <final>true</final> </property> $HADOOP_HOME/etc/hadoop中的hdfs-site.xml增加 <property> <name>dfs.namenode.name.dir</name> <value>file:/home/hadoop/cloudera/hadoop/dfs/name</value> <description>Determines where on the local filesystem the DFS namenode should store the name table.If this is a comma-delimited list ofdirectories,then name table is replicated in all of the directories,forredundancy.</description> <final>true</final> </property> <property> <name>dfs.datanode.data.dir</name> <value>file:/home/hadoop/cloudera/hadoop/dfs/data</value> <description>Determines where on the local filesystem an DFS datanode should store its blocks.If this is a comma-delimited list ofdirectories,then data will be stored in all named directories,typically ondifferent devices.Directories that do not exist are ignored. </description> <final>true</final> </property> <property> <name>dfs.http.address</name> <value>fca-vm-arch-proxy1:50070</value> </property> <property> <name>dfs.replication</name> <value>2</value> </property> <property> <name>dfs.secondary.http.address</name> <value>fca-vm-arch-proxy1:50090</value> </property> <property> <name>dfs.permission</name> <value>false</value> </property> $HADOOP_HOME/etc/hadoop中的mapred-site.xml增加 <property> <name>mapreduce.framework.name</name> <value>yarn</value> </property> <property> <name>mapreduce.job.tracker</name> <value>hdfs://fca-vm-arch-proxy1:9001</value> <final>true</final> </property> <property> <name>mapreduce.map.memory.mb</name> <value>1536</value> </property> <property> <name>mapreduce.map.java.opts</name> <value>-Xmx1024M</value> </property> <property> <name>mapreduce.reduce.memory.mb</name> <value>3072</value> </property> <property> <name>mapreduce.reduce.java.opts</name> <value>-Xmx2560M</value> </property> <property> <name>mapreduce.task.io.sort.mb</name> <value>512</value> </property> <property> <name>mapreduce.task.io.sort.factor</name> <value>100</value> </property> <property> <name>mapreduce.reduce.shuffle.parallelcopies</name> <value>50</value> </property> $HADOOP_HOME/etc/hadoop/hadoop-env.sh增加 export JAVA_HOME=/jdk1.6.0_22 系统环境变量 $HOME/.bash_profile增加 export JAVA_HOME=/jdk1.6.0_22 export JAVA_BIN=${JAVA_HOME}/bin export CLASSPATH=.:$JAVA_HOME/lib/dt.jar:$JAVA_HOME/lib/tools.jar export HADOOP_HOME=/home/hadoop/cloudera/hadoop-2.0.0-cdh4.1.2 export HADOOP_MAPRED_HOME=${HADOOP_HOME} export HADOOP_COMMON_HOME=${HADOOP_HOME} export HADOOP_HDFS_HOME=${HADOOP_HOME} export HADOOP_YARN_HOME=${HADOOP_HOME} export PATH=$PATH:${JAVA_HOME}/bin:${HADOOP_HOME}/bin:${HADOOP_HOME}/sbin:${HIVE_HOME}/bin export JAVA_HOME JAVA_BIN PATH CLASSPATH JAVA_OPTS export HADOOP_LIB=${HADOOP_HOME}/lib export HADOOP_CONF_DIR=${HADOOP_HOME}/etc/hadoop source $HOME/.bash_profile使变量生效 yarn配置 $HADOOP_HOME/etc/hadoop/yarn-site.xml中增加 <property> <name>yarn.resourcemanager.address</name> <value>fca-vm-arch-proxy1:9002</value> </property> <property> <name>yarn.resourcemanager.scheduler.address</name> <value>fca-vm-arch-proxy1:9003</value> </property> <property> <name>yarn.resourcemanager.resource-tracker.address</name> <value>fca-vm-arch-proxy1:9004</value> </property> <property> <name>yarn.nodemanager.aux-services</name> <value>mapreduce.shuffle</value> </property> <property> <name>yarn.nodemanager.aux-services.mapreduce.shuffle.class</name> <value>org.apache.hadoop.mapred.ShuffleHandler</value> </property> $HADOOP_HOME/etc/hadoop/slaves增加 slave1 slave2 将master上的CHD目录及.bash_profile拷贝到slave1,slave2，并配置环境变量，并且配置好ssh无密码登陆，网上很多不详细说了 启动hdfs和yarn 以上步骤都执行完成后，用hadoop用户登录到master机器依次执行： hdfs namenode -format start-dfs.sh start-yarn.sh 通过jps命令查看： master成功启动了NameNode、ResourceManager、SecondaryNameNode进程； slave1,slave2成功启动了DataNode、NodeManager进程。 hive安装 hive只需要在master上安装因为impala-state-store需要hive读取元数据，hive又依赖于关系统型数据库（mysql)所以安装mysql 下载hive http://archive.cloudera.com/cdh4/cdh/4/ 解压缩hive $HOME/.bash_profile增加 export HIVE_HOME=/home/hadoop/hive-0.9.0-cdh4.1.2 export PATH=$PATH:${JAVA_HOME}/bin:${HADOOP_HOME}/bin:${HADOOP_HOME}/sbin:${HIVE_HOME}/bin export HIVE_CONF_DIR=$HIVE_HOME/conf export HIVE_LIB=$HIVE_HOME/lib source $HOME/.bash_profile使环境变量生效 在hive/lib目录下加入mysql-connector-java-5.1.8.jar $HIVE_HOME/conf/hive.site.xml增加 <property> <name>hive.metastore.uris</name> <value>thrift://master:9083</value> <description>Thrift uri for the remote metastore. Used by metastore client to connect to remote metastore.</description> </property> <property> <name>hive.metastore.local</name> <value>false</value> </property> <property> <name>javax.jdo.option.ConnectionURL</name> <value>jdbc:mysql://master:3306/hive?createDatabaseIfNoExist=true</value> <description>JDBC connect string for a JDBC metastore</description> </property> <property> <name>javax.jdo.option.ConnectionDriverName</name> <value>com.mysql.jdbc.Driver</value> <description>Driver class name for a JDBC metastore</description> </property> <property> <name>javax.jdo.option.ConnectionUserName</name> <value>root</value> <description>username to use against metastore database</description> </property> <property> <name>javax.jdo.option.ConnectionPassword</name> <value>password</value> <description>password to use against metastore database</description> </property> <property> <name>hive.security.authorization.enabled</name> <value>false</value> <description>enable or disable the hive client authorization</description> </property> <property> <name>hive.security.authorization.createtable.owner.grants</name> <value>ALL</value> <description>the privileges automatically granted to the owner whenever a table gets created. An example like "select,drop" will grant select and drop privilege to the owner of the table</description> </property> <property> <name>hive.querylog.location</name> <value>${user.home}/hive-logs/querylog</value> </property> 由于hive metstore我们是安装在远程节点上的所以hive.metastore.local是false hive.metastore.uris设置远程连接metstore 验证成功状态 完成以上步骤之后，验证hive安装是否成功 在master命令行执行hive，并输入”show tables;”，出现以下提示，说明hive安装成功： >hive hive> show databases; ok Time taken: 18.952 seconds impala安装 master上安装 impala-state-store sudo yum install impala-state-store master上安装 impala-shell sudo yum install impala-shell 配置impala 修改/etc/default/impala IMPALA_STATE_STORE_HOST=192.168.200.114 IMPALA_STATE_STORE_PORT=24000 IMPALA_BACKEND_PORT=22000 IMPALA_LOG_DIR=/var/log/impala IMPALA_STATE_STORE_ARGS=" -log_dir=${IMPALA_LOG_DIR} -state_store_port=${IMPALA_STATE_STORE_PORT}" IMPALA_SERVER_ARGS=" \ -log_dir=${IMPALA_LOG_DIR} \ -state_store_port=${IMPALA_STATE_STORE_PORT} \ -use_statestore \ -state_store_host=${IMPALA_STATE_STORE_HOST} \ -be_port=${IMPALA_BACKEND_PORT}" ENABLE_CORE_DUMPS=false LIBHDFS_OPTS=-Djava.library.path=/usr/lib/impala/lib MYSQL_CONNECTOR_JAR=/home/hadoop/cloudera/hive/hive-0.9.0-cdh4.1.2/lib/mysql-connector-java-5.1.8.jar IMPALA_BIN=/usr/lib/impala/sbin IMPALA_HOME=/usr/lib/impala HIVE_HOME=/home/hadoop/cloudera/hive/hive-0.9.0-cdh4.1.2 # HBASE_HOME=/usr/lib/hbase IMPALA_CONF_DIR=/usr/lib/impala/conf HADOOP_CONF_DIR=/usr/lib/impala/conf HIVE_CONF_DIR=/usr/lib/impala/conf # HBASE_CONF_DIR=/etc/impala/conf 拷贝hadoop的core-site.xml,hdfs-site.xml，hive的hive-site.xml到/usr/lib/impala/conf中 core-site.xml增加 <property> <name>dfs.client.read.shortcircuit</name> <value>true</value> </property> <property> <name>dfs.client.read.shortcircuit.skip.checksum</name> <value>false</value> </property> hdfs-site.xml增加,hadoop的hdfs-site.xml也增加 <property> <name>dfs.datanode.hdfs-blocks-metadata.enabled</name> <value>true</value> </property> <property> <name>dfs.datanode.data.dir.perm</name> <value>750</value> </property> <property> <name>dfs.block.local-path-access.user</name> <value>hadoop</value> </property> <property> <name>dfs.client.read.shortcircuit</name> <value>true</value> </property> <property> <name>dfs.client.file-block-storage-locations.timeout</name> <value>3000</value> </property> <property> <name>dfs.client.use.legacy.blockreader.local</name> <value>true</value> </property> 拷贝mysql-connector-java-5.1.8.jar到/usr/lib/impala/lib下 拷贝mysql-connector-java-5.1.8.jar到/var/lib/impala下 拷贝/usr/lib/impala/lib/*.so*到$HADOOP_HOME/lib/native/ slave1,slave2上安装 sudo yum install impala sudo yum install impala-server sudo yum install impala-shell master上的hive-site.xml,core-site.xml,hdfs-site.xml拷贝到slaver1,slaver2上，jar的拷贝与master一致 启动hive metastore 在master上执行hive --service metastore 启动impala statestore 在master上执行statestored -log_dir=/var/log/impala -state_store_port=24000 在slave1,slave2上启动impalad sudo /etc/init.d/impala-server start impala查看/var/log/impala/statestored.INFO是否成功 statestored.ERROR查看错误 注意先要在master启动hive metastore,impala statesored,再在slave1,slave2启动impalad-server 测试是否成功 master上执行 impala-shell [Not connected] >connect slave1; [slave1:21000] > use hive; Query: use hive [slave1:21000] >show tables; ok 没有错误说成功 如果在slave1上插入数据需要在slave2上refresh 表名才能同步数据，而不是网上说的refresh,后面必须加表名。 如果不是shell执行操作，应该可以同步数据，没测试过。 注意事项impala在插入数据时可以会出错
hdfsOpenFile(hdfs://fmaster:9000/user/hive/warehouse/test/.2038125373027453036......
是权限问题，因为我们是用sudo (root用户)启动的impala但是test表hadoop用户有增删改查的权限，但是root没有
解决方法
hdfs dfs -chmod -R 777 /user/hive/warehouse/test