1. Install Docker and Docker Compose;
2. Go to docker/playground, and use docker compose up -d to run compose services as daemon;

1. Connect using beeline

docker exec -it kyuubi /opt/kyuubi/bin/beeline -u 'jdbc:hive2://0.0.0.0:10009/tpcds/tiny';

2. Connect using DBeaver

Add a Kyuubi datasource with

• connection url jdbc:hive2://0.0.0.0:10009/tpcds/tiny
• username: anonymous
• password: <empty>

3. Use built-in dataset

Kyuubi supply some built-in dataset, after Kyuubi started, you can run the following command to load the different datasets:

• For loading TPC-DS tiny dataset to spark_catalog.tpcds_tiny, run docker exec -it kyuubi /opt/kyuubi/bin/beeline -u 'jdbc:hive2://0.0.0.0:10009/' -f /opt/load_data/load-dataset-tpcds-tiny.sql
• For loading TPC-H tiny dataset to spark_catalog.tpch_tiny, run docker exec -it kyuubi /opt/kyuubi/bin/beeline -u 'jdbc:hive2://0.0.0.0:10009/' -f /opt/load_data/load-dataset-tpch-tiny.sql

• MinIO: http://localhost:9001
• PostgreSQL localhost:5432 (username: postgres, password: postgres)
• Spark UI: http://localhost:4040 (available after Spark application launching by Kyuubi, port may be 4041, 4042... if you launch more than one Spark applications)
• Prometheus: http://localhost:9090
• Grafana: http://localhost:3000 (username: admin, password: admin)

1. Stop compose services by docker compose down;

1. Build images docker/playground/build-image.sh;
2. Optional to use buildx to build and publish cross-platform images BUILDX=1 docker/playground/build-image.sh;