Edgar Hernández israel-hdez

Introduction

In the Open Data Hub project (ODH) KServe is provided, however not all KServe features are currently supported, because OpenShift integration is not complete yet.

The previous doesn't mean that OpenShift is not compatible with KServe (or vice-versa). You should be able to use KServe in OpenShift by following upstream instructions: https://kserve.github.io/website/latest/get_started/, but you will require to install the whole stack by yourself. In ODH one of the goals is to provide a AI platform that is well integrated with OpenShift and providing a package that is easy to install. For KServe, such integration is not finished yet.

Introduction

This guide describes the steps needed to generate a build and deploy/install the needed components that add authorization capabilities to the ODH KServe stack.

The authorization capabilities are implemented in two pull requests:

opendatahub-io/opendatahub-operator#784
- This pull request enables the ODH operator to install an Authorino instance, and to configure the Service Mesh to work with Authorino as an authorization provider. Additionally, when KServe is enabled, some additional configurations are installed whih are specific for KServe.
opendatahub-io/odh-model-controller#130

	FROM busybox

	# Based on ovms quick start guide
	ARG MODEL_TAR_URL=https://www.kaggle.com/api/v1/models/tensorflow/faster-rcnn-resnet-v1/tensorFlow2/faster-rcnn-resnet50-v1-640x640/1/download

	RUN mkdir -p /models/1; \
	chmod -R 775 /models; \
	wget ${MODEL_TAR_URL} -O - \| tar -zx -C /models/1

	# For this reproducer, CRC (aka OpenShift Local) was used.
	# Note: This script is NOT prepared nor tested to be run at the CLI.
	# Instead, commands were copied & pasted. This would be the recommended way to try.
	#
	# 1. Install ODH 2.11.
	# 1.1 Install dependencies - OSSM operator
	cat <<EOF \| oc apply -f -
	apiVersion: operators.coreos.com/v1alpha1
	kind: Subscription
	metadata:

	#!/usr/bin/env bash

	# Deploy to new namespace
	oc new-project kserve-test

	# Create a ServingRuntime
	curl -s https://raw.githubusercontent.com/opendatahub-io/kserve/master/config/runtimes/kserve-sklearnserver.yaml \| \
	sed 's/ClusterServingRuntime/ServingRuntime/' \| \
	sed "s\|kserve-sklearnserver:replace\|docker.io/kserve/sklearnserver:latest\|" \| \
	oc apply -f -

	kind: BuildConfig
	apiVersion: build.openshift.io/v1
	metadata:
	name: kserve-controller-bc
	spec:
	source:
	git:
	uri: https://github.com/opendatahub-io/kserve.git
	ref: master
	contextDir: "."

	FROM quay.io/opendatahub/openvino_model_server:2022.3-release

	USER root
	RUN mkdir /models && chown ovms:ovms /models

	USER ovms
	RUN curl --create-dirs \
	https://storage.openvinotoolkit.org/repositories/open_model_zoo/2022.1/models_bin/2/face-detection-retail-0004/FP32/face-detection-retail-0004.xml \
	https://storage.openvinotoolkit.org/repositories/open_model_zoo/2022.1/models_bin/2/face-detection-retail-0004/FP32/face-detection-retail-0004.bin \
	-o /models/1/face-detection-retail-0004.xml \

	#!/bin/bash

	set -e

	: ${ISTIO_DIR:="$HOME/apps/istio-1.9.0"}
	ISTIOCTL=$ISTIO_DIR/bin/istioctl

	: ${BOOKINFO_NS:=bookinfo}

	: ${CTX1:=kukulcan}

	apiVersion: apps/v1
	kind: Deployment
	metadata:
	name: workload-a
	labels:
	app: workload-a
	version: v1
	spec:
	replicas: 1
	selector: