Copy Druid jackson libs to Dataproc

druid-dataproc.sh

#!/bin/bash

DRUID="0.10.0-rc3-SNAPSHOT"
HADOOP="2.7.3"

# Create a Dataproc hadoop initialization
mkdir -p initialization/jars/

cat > initialization/initialization.sh << stop
#!/bin/bash

mkdir /tmp/initialization
gsutil -m rsync -r -d gs://hadoop-eu-atomx/${DRUID}-${HADOOP}/ /tmp/initialization/

rm /usr/lib/hadoop-mapreduce/jackson-*-2.*.jar
cp /tmp/initialization/jars/* /usr/lib/hadoop-mapreduce/

exit 0
stop

rm -rf initialization/jars/*.jar
mkdir -p initialization/jars
cp druid-${DRUID}/lib/jackson-*-2.*.jar initialization/jars/

gsutil -m rsync -d -r initialization/ gs://hadoop-eu-atomx/${DRUID}-${HADOOP}/

echo "Don't forget to add the script as initialization action when creating the Dataproc cluster."
echo "gs://hadoop-eu-atomx/${DRUID}-${HADOOP}/initialization.sh"