Skip to content

Latest commit

 

History

History
42 lines (25 loc) · 1.66 KB

README.md

File metadata and controls

42 lines (25 loc) · 1.66 KB

Ray demo

Here is how you set up and go through the Ray demo.

Requirements

Make sure you have the RHODS demo pack installed. Note that RHODS operator v2 is needed for deploying the CodeFlare stack:

  • If you're on RHODS 1.33.0, change the operator subscription channel to alpha.
  • Once operator 2.x is installed, create a new Data Science Cluster instance through the operator page.
  • In the form view of DataScienceCluster, open the Components drop-down view and ensure that the managementState of codeflare and ray has the value Managed. Click Create.
  • Ensure that the CodeFlare Operator is installed. The community operator version is recommended as of RHODS 1.33.

Deploying the demo requires Helm v3.

Deploying the demo

Ensure you're logged into the OpenShift cluster as cluster admin and the required operators are installed (see above).

Navigate to /manifests/ray-demo and edit values.yaml:

  • Set clusterdomainurl to the domain URL of your OpenShift cluster.
  • Set ocptoken to the token value of your OpenShift authentication token. You can find it in the OpenShift web console after clicking on your user name (top right corner) and Copy login command.

Deploy the Helm chart:

helm install ray-demo .

The deployment of all components may take up to 30 minutes.

Post-configuration

Prepare data sources

Open the Minio admin console via the minio-ui route URL in the minio namespace. Log in with minio / minio123. Navigate to Buckets and select Create Bucket. Enter ray-demo under bucket name and select Create Bucket.

Running the end-to-end workflow

Interactive

TODO