[Spark]Spark-sql与hive连接配置

一.在Mysql中配置hive数据库

1. 创建hive数据库，刷新root用户权限

create database hive;
grant all on *.* to root@'%' identified by'';
flush privileges;

1. 修改hive目录下/bin/hive
   vim /usr/local/src/apache-hive-1.2.2-bin/bin/hive

修改前

# add Spark assembly jar to the classpath
if [[ -n "$SPARK_HOME" ]]
then
  sparkAssemblyPath=`ls ${SPARK_HOME}/lib/spark-assembly-*.jar`
  CLASSPATH="${CLASSPATH}:${sparkAssemblyPath}"
修改后

# add Spark assembly jar to the classpath
if [[ -n "$SPARK_HOME" ]]
then
  sparkAssemblyPath=`ls ${SPARK_HOME}/jars/*.jar`
  CLASSPATH="${CLASSPATH}:${sparkAssemblyPath}"
fi

二.hadoop目录下存在老版的jline,替换掉

cd /usr/local/src
cp apache-hive-1.2.2-bin/lib/jline-2.12.jar hadoop-2.6.1/share/hado
op/yarn/lib/

三. 运行spark-shell

在spark目录下./spark-shell 运行后

scala> import org.apache.spark.sql.hive.HiveContext
import org.apache.spark.sql.hive.HiveContext

scala> val priors = spark.sql("select * from badou.orders")

报错：

// :: ERROR metastore.RetryingHMSHandler: AlreadyExistsException(message:Database default already exists)

// :: WARN component.AbstractLifeCycle: FAILED org.spark_project.jetty.server.Server@7b1e5e55: java.net.BindException: Address already in use
java.net.BindException: Address already in use

解决办法

步骤一 将hive-site.xml拷贝到spark/conf里

cp /usr/local/src/apache-hive-1.2.-bin/conf/hive-site.xml /usr/local/src/spark-2.0.-bin-hadoop2./conf/
scp /usr/local/src/apache-hive-1.2.-bin/conf/hive-site.xml root@slave1:/usr/local/src/spark-2.0.-bin-hadoop2./conf/
scp /usr/local/src/apache-hive-1.2.-bin/conf/hive-site.xml root@slave2:/usr/local/src/spark-2.0.-bin-hadoop2./conf/

步骤二 将mysql驱动拷贝到spark/jar里

cp /usr/local/src/apache-hive-1.2.-bin/lib/mysql-connector-java-5.1.-bin.jar /usr/local/src/spark-2.0.-bin-hadoop2./jars/
scp /usr/local/src/apache-hive-1.2.-bin/lib/mysql-connector-java-5.1.-bin.jar root@slave1:/usr/local/src/spark-2.0.-bin-hadoop2./jars/
scp /usr/local/src/apache-hive-1.2.-bin/lib/mysql-connector-java-5.1.-bin.jar root@slave2:/usr/local/src/spark-2.0.-bin-hadoop2./jars/