docker安装hive笔记

前两篇文章介绍了docker的基本命令如何安装hadoop

那么大家会比较了解docker的基本语法的安装过程。那么咱们今天来一起安装一下hive。

安装

1、下载gitHub，地址：https://github.com/prasanthj/docker-hive-on-tez。如果背墙了，可以选择下载zip。进入目录之后就能看见如下内容：

@~/git/github/docker-hive-on-tez-master $ ls
Dockerfile        datagen.py        hive-log4j.properties    store_sales.sql
LICENSE            hive-0.14        hive-site.xml        store_sales.txt
README.md        hive-bootstrap.sh    postgresql.conf

2、安装：

docker build --no-cache=true -t local-hive-on-tez .

这是一个漫长的过程，喝一杯咖啡，该干嘛干嘛，几个小时之后回来......

3、进入系统

docker --tls run -i -t -P local-hive-on-tez /etc/hive-bootstrap.sh -bash
Starting postgresql server...
/
-- :: GMT LOG:  database system was interrupted; last known up at -- :: GMT
-- :: GMT LOG:  database system was not properly shut down; automatic recovery in progress
-- :: GMT LOG:  redo starts at /1782A58

4、查看hive

root@2c1282c522bf:/# hive -f /opt/files/store_sales.sql

Logging initialized using configuration in file:/usr/local/hive-dist/apache-hive-0.15.-SNAPSHOT-bin/conf/hive-log4j.properties
SLF4J: Class path contains multiple SLF4J bindings.
SLF4J: Found binding in [jar:file:/usr/local/hadoop-2.5./share/hadoop/common/lib/slf4j-log4j12-1.7..jar!/org/slf4j/impl/StaticLoggerBinder.class]

5、端口映射

如果你只想登录虚拟机玩玩，这个就足够了，可是还想看看这个hadoop运行的怎么样，需要在页面查看一下

@~/VirtualBox VMs/boot2docker-vm $ docker ps -a
CONTAINER ID        IMAGE                      COMMAND                CREATED             STATUS                      PORTS                                                                                                                                                                                                                                                                                                                                                              NAMES
2c1282c522bf        local-hive-on-tez:latest   "/etc/hive-bootstrap   39 hours ago        Up 6 hours                  0.0.0.0:49181->8032/tcp, 0.0.0.0:49182->50075/tcp, 0.0.0.0:49183->50010/tcp, 0.0.0.0:49184->50090/tcp, 0.0.0.0:49185->8031/tcp, 0.0.0.0:49186->8040/tcp, 0.0.0.0:49187->8088/tcp, 0.0.0.0:49188->22/tcp, 0.0.0.0:49189->50020/tcp, 0.0.0.0:49190->8030/tcp, 0.0.0.0:49191->49707/tcp, 0.0.0.0:49192->8033/tcp, 0.0.0.0:49193->8042/tcp, 0.0.0.0:->/tcp   kickass_galileo     

发现上面的一个端口是49194对应虚拟机的50070端口。

这还不够，需要知道自己的boot2docker的ip是什么？默认是192.168.59.103

那么需要在自己的浏览器输入：http://192.168.59.103:49194/ 就会看见hadoop的运行状态了。

备注：

为什么下载github？

因为需要获取他的Dockerfile，好让docker知道它依赖docker-tez，然后在虚拟机执行下载和安装hive，内容如下：

FROM prasanthj/docker-tez:tez-0.5.2  #这是说明依赖什么，下面是安装命令
MAINTAINER Prasanth Jayachandran

# to configure postgres as hive metastore backend
RUN apt-get update
RUN apt-get -yq install vim postgresql-9.3 libpostgresql-jdbc-java

# having ADD commands will invalidate the cache forcing hive build from trunk everytime
# copy config, sql, data files to /opt/files
RUN mkdir /opt/files
ADD hive-site.xml /opt/files/
ADD hive-log4j.properties /opt/files/
ADD store_sales.* /opt/files/
ADD datagen.py /opt/files/

# clone and compile hive
ENV HIVE_VERSION 0.15.-SNAPSHOT
RUN cd /usr/local && git clone https://github.com/apache/hive.git #在天朝这可能被墙，所以速度非常慢
RUN cd /usr/local/hive && /usr/local/maven/bin/mvn clean install -DskipTests -Phadoop-,dist
RUN mkdir /usr/local/hive-dist && tar -xf /usr/local/hive/packaging/target/apache-hive-${HIVE_VERSION}-bin.tar.gz -C /usr/local/hive-dist

# set hive environment
ENV HIVE_HOME /usr/local/hive-dist/apache-hive-${HIVE_VERSION}-bin
ENV HIVE_CONF $HIVE_HOME/conf
ENV PATH $HIVE_HOME/bin:$PATH
ADD hive-site.xml $HIVE_CONF/hive-site.xml
ADD hive-log4j.properties $HIVE_CONF/hive-log4j.properties

# zookeeper pulls jline 0.9. and hive pulls jline2. This workaround is from HIVE-
RUN rm $HADOOP_PREFIX/share/hadoop/yarn/lib/jline-0.9..jar

# add postgresql jdbc jar to classpath
RUN ln -s /usr/share/java/postgresql-jdbc4.jar $HIVE_HOME/lib/postgresql-jdbc4.jar

# set permissions for hive bootstrap file
ADD hive-bootstrap.sh /etc/hive-bootstrap.sh
RUN chown root:root /etc/hive-bootstrap.sh
RUN chmod  /etc/hive-bootstrap.sh

# to avoid psql asking password, set PGPASSWORD
ENV PGPASSWORD hive
# 下面是安装postgresql，这个在国外很流行
# To overcome the bug in AUFS that denies postgres permission to read /etc/ssl/private/ssl-cert-snakeoil.key file.
# https://github.com/Painted-Fox/docker-postgresql/issues/30
# https://github.com/docker/docker/issues/783
# To avoid this issue lets disable ssl in postgres.conf. If we really need ssl to encrypt postgres connections we have to fix permissions to /etc/ssl/private directory everytime until AUFS fixes the issue
ENV POSTGRESQL_MAIN /var/lib/postgresql/9.3/main/
ENV POSTGRESQL_CONFIG_FILE $POSTGRESQL_MAIN/postgresql.conf
ENV POSTGRESQL_BIN /usr/lib/postgresql/9.3/bin/postgres
ADD postgresql.conf $POSTGRESQL_MAIN
RUN chown postgres:postgres $POSTGRESQL_CONFIG_FILE

USER postgres
# create metastore db, hive user and assign privileges
RUN /etc/init.d/postgresql start &&\
     psql --command "CREATE DATABASE metastore;" &&\
     psql --command "CREATE USER hive WITH PASSWORD 'hive';" && \
     psql --command "ALTER USER hive WITH SUPERUSER;" && \
     psql --command "GRANT ALL PRIVILEGES ON DATABASE metastore TO hive;" && \
     cd $HIVE_HOME/scripts/metastore/upgrade/postgres/ &&\
        psql -h localhost -U hive -d metastore -f hive-schema-0.15..postgres.sql

# revert back to root user
USER root

参考：

https://github.com/prasanthj/docker-hive-on-tez

https://github.com/prasanthj/docker-hadoop

https://github.com/prasanthj/docker-tez