OpenShift (OCP) User Guide

OpenShift provides an efficient and manageable way to orchestrate deep learning workloads at scale. This document provides instructions on setting up a generic OpenShift-based solution for an on-premise platform based on CoreOS Host configuration.

Note

• This guide is applicable for Intel® Gaudi® software version 1.14.0 and below.

• OpenShift (OCP) installed on the host is based on RHEL8.6. It is no longer supported on CentOS8.3 due to end-of-life.

• RHEL8.6 subscription is required. Follow the OCP based host setup instructions in Create an OpenShift Cluster.

The following components are needed to deploy a generic Kubernetes solution:

• Docker image needed to build and load the Intel Gaudi driver. For more details, refer to Installation Guide.

• Intel Gaudi software packages required to install and load FW and Intel Gaudi driver. For more details, refer to Installation Guide.

• Intel Gaudi device plugin for Kubernetes. For more details, refer to Kubernetes User Guide.

The above components except Docker image are available in the Intel Gaudi vault.

The following sections detail the steps to set up OpenShift and provide implementation examples:

• Prepare Docker Image on OCP-based Host
• Build and Run Docker Container
• Load Intel Gaudi Driver Inside Running Docker Container
• Intel Gaudi Device Plugin for Kubernetes
• Usage Examples

For details on OpenShift, refer to OpenShift documentation.