Dockerfile

FROM bde2020/hadoop-base:2.0.0-hadoop2.7.4-java8

MAINTAINER Yiannis Mouchakis <gmouchakis@iit.demokritos.gr>
MAINTAINER Ivan Ermilov <ivan.s.ermilov@gmail.com>

# Allow buildtime config of HIVE_VERSION
ARG HIVE_VERSION
# Set HIVE_VERSION from arg if provided at build, env if provided at run, or default
# https://docs.docker.com/engine/reference/builder/#using-arg-variables
# https://docs.docker.com/engine/reference/builder/#environment-replacement
ENV HIVE_VERSION=${HIVE_VERSION:-2.3.2}

ENV HIVE_HOME /opt/hive
ENV PATH $HIVE_HOME/bin:$PATH
ENV HADOOP_HOME /opt/hadoop-$HADOOP_VERSION

WORKDIR /opt

#Install Hive and PostgreSQL JDBC
RUN apt-get update && apt-get install -y wget procps && \
	wget https://archive.apache.org/dist/hive/hive-$HIVE_VERSION/apache-hive-$HIVE_VERSION-bin.tar.gz && \
	tar -xzvf apache-hive-$HIVE_VERSION-bin.tar.gz && \
	mv apache-hive-$HIVE_VERSION-bin hive && \
	wget https://jdbc.postgresql.org/download/postgresql-9.4.1212.jar -O $HIVE_HOME/lib/postgresql-jdbc.jar && \
	rm apache-hive-$HIVE_VERSION-bin.tar.gz && \
	apt-get --purge remove -y wget && \
	apt-get clean && \
	rm -rf /var/lib/apt/lists/*


#Spark should be compiled with Hive to be able to use it
#hive-site.xml should be copied to $SPARK_HOME/conf folder

#Custom configuration goes here
ADD hive/conf/hive-site.xml $HIVE_HOME/conf
ADD hive/conf/beeline-log4j2.properties $HIVE_HOME/conf
ADD hive/conf/hive-env.sh $HIVE_HOME/conf
ADD hive/conf/hive-exec-log4j2.properties $HIVE_HOME/conf
ADD hive/conf/hive-log4j2.properties $HIVE_HOME/conf
ADD hive/conf/ivysettings.xml $HIVE_HOME/conf
ADD hive/conf/llap-daemon-log4j2.properties $HIVE_HOME/conf

COPY hivescript/startup.sh /usr/local/bin/
RUN chmod +x /usr/local/bin/startup.sh

COPY hivescript/entrypoint.sh /usr/local/bin/
RUN chmod +x /usr/local/bin/entrypoint.sh

EXPOSE 10000
EXPOSE 10002

ENTRYPOINT ["hivescript/entrypoint.sh"]
CMD startup.sh