Skip to content

MLflow Installation on OpenShift (OpenDataHub)

This guide describes how to install MLflow on OpenShift using the OpenDataHub MLflow Operator Helm chart.

Prerequisites

Before running the installation script, ensure you have the following tools installed and configured:

  • git: For cloning the operator repository.

  • helm: For managing the MLflow installation.

  • oc (OpenShift CLI): For interacting with the OpenShift cluster.

Installation Steps

  1. Prepare the environment: Log in to the jumphost, clone the repository, and navigate to the root directory: 

    git clone <repository-url>
cd MLOps-Workshop-Exercises

  2. Run the installation script: 

    chmod +x environment_preparations/mlflow_odh_installation.sh
./environment_preparations/mlflow_odh_installation.sh

  3. Configure the installation: The script will prompt you for the following configurations:

  4. OpenShift namespace: The project where MLflow will be deployed (default: mlflow).

  5. Helm release name: The name of the Helm release (default: mlflow).
  6. Backend store URI: The database URI (default: sqlite:////mlflow/mlflow.db).
  7. Persistent storage: Whether to enable persistent storage (default: true) and its size (default: 2Gi).
  8. Resource Limits: CPU and Memory requests/limits.

  9. Authentication Mode:

    • Disable auth: Direct browser access (recommended for labs).
    • Keep auth: Secure access via OpenShift OAuth proxy.

  10. Verify the installation: The script will automatically create the project, patch the chart, install MLflow via Helm, set up the OpenShift Route, and provide the final access URL.

Post-Installation

  • Access MLflow UI: Use the URL provided by the script at the end of the execution.
  • Add Workspaces: To allow other namespaces to use MLflow, label them: 
    oc label namespace <target-namespace> mlflow-enabled=true

Uninstall MLflow

To uninstall MLflow from your OpenShift cluster, run the following commands:

helm uninstall mlflow -n mlflow
oc delete route mlflow -n mlflow