Skip to content

Latest commit

 

History

History
38 lines (31 loc) · 1.33 KB

README.md

File metadata and controls

38 lines (31 loc) · 1.33 KB

How to run the examples

Make sure you have docker and sbt installed. Tested using docker 20.10.0, sbt 1.4.1

Clone this repository:

git clone https://github.com/vertica/spark-connector.git

From the project's root directory:

cd docker
docker compose up -d

This will create a docker image for a client container and docker containers for a sandbox client environment and single-node clusters for both Vertica and HDFS.

Next, change directory to the example you want to run. For example, if you want to run the "demo" example:

cd examples/demo

Now create a lib directory:

mkdir lib

Put the assembly jar for the connector in the lib folder you just created. If you don't have the jar file, you can download it here: https://github.com/vertica/spark-connector/releases

Run the following commands to update the HDFS configuration files and restart HDFS:

docker exec docker_hdfs_1 cp /hadoop/conf/core-site.xml /opt/hadoop/etc/hadoop/core-site.xml
docker exec docker_hdfs_1 cp /hadoop/conf/hdfs-site.xml /opt/hadoop/etc/hadoop/hdfs-site.xml
docker exec docker_hdfs_1 /opt/hadoop/sbin/stop-dfs.sh
docker exec docker_hdfs_1 /opt/hadoop/sbin/start-dfs.sh

Run docker exec -it docker_client_1 /bin/bash to enter the sandbox client environment.

Now just run sbt run from the /spark-connector/examples/demo directory.