Prerequisites

  • RHEL 8.x jumpbox that is registred with Red Hat
  • Review and run the openshift-ai-workload.sh
  • If using demo.redhat.com use RHEL 8 Base instance to configure a jumpbox and run the fork and run rhel8-bootstrap https://github.com/tosin2013/rhel8-bootstrap/actions/workflows/configure-rhel8-bastion.yml to run GitHub Actions to configure the jumpbox. Igf you do not want to run this on your laptop.

SSH into RHEL jumpbox

Deploying GPUs for AI Workloads on OpenShift on AWS

Export the following AWS variables before running this script:
export aws_access_key_id="YOUR_ACCESS_KEY_ID"
export aws_secret_access_key="YOUR_SECRET_ACCESS_KEY"
export aws_region="YOUR_AWS_REGION"

curl -OL https://gist.githubusercontent.com/tosin2013/76e47de3f32de4486ab4699c21b2188e/raw/959ae5dd2117edf124e4531cfae5216c722a3358/openshift-ai-workload.sh
chmod +x openshift-ai-workload.sh
./openshift-ai-workload.sh m6i.4xlarge # without GPU

./openshift-ai-workload.sh g4dn.2xlarge # with GPU

Configure SSL for OpenShift Console

$ sudo su - 
$ curl -OL https://gist.githubusercontent.com/tosin2013/866522a1420ac22f477d2253121b4416/raw/35d6fa88675d63b6ecf58a827df32356ccf3ddde/configure-keys-on-openshift.sh
$ chmod +x configure-keys-on-openshift.sh
$ ./configure-keys-on-openshift.sh AWS_ACCESS_KEY AWS_SECRET_ACCESS_KEY podman

Configure jumpbox for deployment

curl -OL https://raw.githubusercontent.com/tosin2013/redhat-edge-ai-industrial-demo-infra/main/dev-box.sh
chmod +x dev-box.sh
./dev-box.sh

Deploy Gitea

curl -OL https://raw.githubusercontent.com/tosin2013/openshift-demos/master/quick-scripts/deploy-gitea.sh
chmod +x deploy-gitea.sh
./deploy-gitea.sh

Configure Gitea

Click on route in OpenShift Console 20240215124034 (1) Click on Register (2) Register a new account 20240215124218 (3) Click on New Migration click on Git 20240215124321 (4) Click on Migrate From Git

  • https://github.com/tosin2013/redhat-edge-ai-industrial-demo-infra.git 20240215124519 (5) Edit the following folders in the repo
  • Replace with your gita URL
    File path: clusters/overlays/aws/patch-applicationset-repo-revision.yaml 20240215130537 File path 2: clusters/overlays/aws/patch-application-repo-revision.yaml 20240215130742

    Apply configure changes to argocd

$ git clone https://$(oc get route -n gitea | grep gitea-with-admin | awk '{print $2}' | head -1)/user1/redhat-edge-ai-industrial-demo-infra.git
$ cd redhat-edge-ai-industrial-demo-infra
$ oc create -k clusters/overlays/aws

Ensure the ArgoCD application is in sync

20240213165739

Configure Enviornment

  1. Run tekton pipeline - The Tekton pipeline is used to build the container image for the application.
  2. Configure rhods - The Red Hat OpenShift Data Science environment is used to train the AI model.
  3. Run rhods Notebooks - Run notebooks to train the AI model.

Back to Homepage