Docker image for Zeppelin + DSE

build.sh

# Build Zeppelin 0.8 that is downloaded from Apache servers
docker build .
# you can build custom version by spefying Z_URL_PREFIX if you have Zeppelin downloaded already, and DSE_VERSION, and Z_VERSION like this
docker build --build-arg Z_URL_PREFIX=http://local-ip --build-arg DSE_VERSION=6.0.4 .

Dockerfile

ARG DSE_VERSION=6.7.0
FROM datastax/dse-server:${DSE_VERSION}

ARG Z_VERSION="0.8.0"
ENV Z_HOME="/zeppelin"

ARG Z_URL_PREFIX=http://archive.apache.org/dist/zeppelin/zeppelin-${Z_VERSION}
ARG Z_TARBALL=zeppelin-${Z_VERSION}-bin-all.tgz
ARG Z_DOWNLOAD_URL=${Z_URL_PREFIX}/${Z_TARBALL}

USER root

RUN echo "Download Zeppelin binary" && \
    wget -O /tmp/zeppelin-${Z_VERSION}-bin-all.tgz ${Z_DOWNLOAD_URL} && \
    cd / && tar -zxvf /tmp/zeppelin-${Z_VERSION}-bin-all.tgz && \
    rm -rf /tmp/zeppelin-${Z_VERSION}-bin-all.tgz && \
    mv /zeppelin-${Z_VERSION}-bin-all ${Z_HOME} && \
    chown -R dse ${Z_HOME}

EXPOSE 8080

USER dse

WORKDIR ${Z_HOME}
ENTRYPOINT [ "dse", "exec", "bin/zeppelin.sh" ]

zeppelin-configuration

After Zeppelin is started, go to "Interpreter", select "spark" interpreter and do following:
1. Find Analytics master IP by executing "dsetool status" on the DSE cluster node - it will be in the first line, right
2. Change "master" property to "dse://<Analytics-Master-IP>?"
3. Add property "spark.cassandra.connection.host" with the same IP
4. Save settings

After that you can create Spark notebook and jobs will be shown in the DSE Analytics Spark Master