Skip to main content

Quickstart

Install the core of stardelt on a cluster, load a small file, and query it. It takes about 15 minutes, most of it waiting for images.

You need: a Kubernetes cluster with a node of at least 4 vCPU and 8 GB and a default StorageClass, plus kubectl, helm 3.18 or newer, and git. No cluster? Create one with the Hetzner guide, step 1.

No sign-in in this install

This install has no sign-in and no permissions, which keeps it simple. Reach it only through the port-forward below, and do not load real data.

1. Install the core​

Check that kubectl points at the right cluster, then install:

kubectl config current-context

git clone https://github.com/stardelt/stardelt-platform
cd stardelt-platform
helm dependency build chart

helm upgrade --install stardelt chart \
  --namespace stardelt --create-namespace \
  -f chart/values-slim.yaml --wait --timeout 15m

This installs object storage, the catalog, the SQL engine and Nova, the interface. When it returns, every pod should be Running or Completed:

kubectl -n stardelt get pods

2. Open Nova​

kubectl -n stardelt port-forward svc/nova 8080:8080

Leave that running and open http://localhost:8080.

3. Load a file​

Save this as cities.csv:

city,country,population
Berlin,DE,3700000
Paris,FR,2100000
Madrid,ES,3300000

In Nova, open Load & transfer. Choose the file, enter the schema demo and the table cities, and load it.

4. Query it​

Open SQL editor and run:

SELECT country, sum(population) AS people
FROM warehouse.demo.cities
GROUP BY country
ORDER BY people DESC;

You have just written an Iceberg table to object storage and read it back through Trino. Open Catalog to see the table and its columns, and Query log to see the query you ran.

5. Clean up​

helm uninstall stardelt -n stardelt
kubectl delete namespace stardelt     # also deletes the data

Next steps​