For years, we have used a “canned” Virtual Machine to stand up a single-node instance on a local workstation. Since the direction of the platform is toward containerized deployments, the virtual machine edition is no longer needed and will be deprecated as of version 8.4.0.
Containerized deployments using Docker Desktop or Minikube are easier to start up locally and provide more flexibility and stability.
What you will need:
Follow these simple steps in a terminal window:
Note: This can be a WSL 2 Linux Bash window, a Windows Command Prompt, a Mac Terminal window, or a Linux Bash terminal. The commands should be the same. As long as the Helm commands are available.
- Add a repository:
helm repo add hpcc https://hpcc-systems.github.io/helm-chart/
- Start a default system:
helm install mycluster hpcc/hpcc --version=8.2.12
The --version argument is optional, but recommended. It ensures that you know which version you are installing. If omitted, the latest non-development version is installed. This example uses 8.2.12, but you should use the version you want.
- Use kubectl to get your interface IPs
kubectl get svc
Find the ECL Watch external IP from the response.
- Access ECL Watch from a web browser using the IP and port (e.g. localhost:8010). You can also use the ECL Watch IP in a configuration for either the ECL IDE or VS Code.
- When you are finished, use this command to terminate the cluster.
helm uninstall mycluster
This stops the cluster, deletes the pods, and with the default settings, it also deletes the storage used.
For more details, read the Containerized HPCC Systems Platform manual.
To get started using your cluster, try the Data Tutorial.