設定 HyperPod 叢集以進行模型部署

設定 kubectl 以連線至由 Amazon EKS 叢集協調的新建立 HyperPod 叢集。指定區域和 HyperPod 叢集名稱。

export HYPERPOD_CLUSTER_NAME=<hyperpod-cluster-name>
export REGION=<region>

# S3 bucket where tls certificates will be uploaded
BUCKET_NAME="<Enter name of your s3 bucket>" # This should be bucket name, not URI

export EKS_CLUSTER_NAME=$(aws --region $REGION sagemaker describe-cluster --cluster-name $HYPERPOD_CLUSTER_NAME \
  --query 'Orchestrator.Eks.ClusterArn' --output text | \
  cut -d'/' -f2)
aws eks update-kubeconfig --name $EKS_CLUSTER_NAME --region $REGION

設定預設 env 變數。

LB_CONTROLLER_POLICY_NAME="AWSLoadBalancerControllerIAMPolicy-$HYPERPOD_CLUSTER_NAME"
LB_CONTROLLER_ROLE_NAME="aws-load-balancer-controller-$HYPERPOD_CLUSTER_NAME"
S3_MOUNT_ACCESS_POLICY_NAME="S3MountpointAccessPolicy-$HYPERPOD_CLUSTER_NAME"
S3_CSI_ROLE_NAME="SM_HP_S3_CSI_ROLE-$HYPERPOD_CLUSTER_NAME"
KEDA_OPERATOR_POLICY_NAME="KedaOperatorPolicy-$HYPERPOD_CLUSTER_NAME"
KEDA_OPERATOR_ROLE_NAME="keda-operator-role-$HYPERPOD_CLUSTER_NAME"
PRESIGNED_URL_ACCESS_POLICY_NAME="PresignedUrlAccessPolicy-$HYPERPOD_CLUSTER_NAME"
HYPERPOD_INFERENCE_ACCESS_POLICY_NAME="HyperpodInferenceAccessPolicy-$HYPERPOD_CLUSTER_NAME"
HYPERPOD_INFERENCE_ROLE_NAME="HyperpodInferenceRole-$HYPERPOD_CLUSTER_NAME"
HYPERPOD_INFERENCE_SA_NAME="hyperpod-inference-service-account"
HYPERPOD_INFERENCE_SA_NAMESPACE="kube-system"
JUMPSTART_GATED_ROLE_NAME="JumpstartGatedRole-$HYPERPOD_CLUSTER_NAME"
FSX_CSI_ROLE_NAME="AmazonEKSFSxLustreCSIDriverFullAccess-$HYPERPOD_CLUSTER_NAME"

從叢集 ARN 擷取 Amazon EKS 叢集名稱、更新本機 kubeconfig，並透過列出跨命名空間的所有 Pod 來驗證連線。

kubectl get pods --all-namespaces

（選用） 安裝 NVIDIA 裝置外掛程式，以在叢集上啟用 GPU 支援。

#Install nvidia device plugin
kubectl create -f https://raw.githubusercontent.com/NVIDIA/k8s-device-plugin/v0.14.5/nvidia-device-plugin.yml
# Verify that GPUs are visible to k8s
kubectl get nodes -o=custom-columns=NAME:.metadata.name,GPU:.status.allocatable.nvidia.com/gpu

收集必要的 AWS 資源識別符和 ARNs，以設定 EKS、SageMaker 和 IAM 元件之間的服務整合。

%%bash -x

export ACCOUNT_ID=$(aws --region $REGION sts get-caller-identity --query 'Account' --output text)
export OIDC_ID=$(aws --region $REGION eks describe-cluster --name $EKS_CLUSTER_NAME --query "cluster.identity.oidc.issuer" --output text | cut -d '/' -f 5)
export EKS_CLUSTER_ROLE=$(aws eks --region $REGION describe-cluster --name $EKS_CLUSTER_NAME --query 'cluster.roleArn' --output text)

將 IAM OIDCidentity 提供者與您的 EKS 叢集建立關聯。

eksctl utils associate-iam-oidc-provider --region=$REGION --cluster=$EKS_CLUSTER_NAME --approve

建立 HyperPod 推論運算子 IAM 角色所需的信任政策和許可政策 JSON 文件。這些政策可在 EKS、SageMaker 和其他 AWS 服務之間啟用安全的跨服務通訊。

bash
# Create trust policy JSON
cat << EOF > trust-policy.json
{
    "Version": "2012-10-17",
    "Statement": [
        {
            "Effect": "Allow",
            "Principal": {
                "Service": [
                    "sagemaker.amazonaws.com"
                ]
            },
            "Action": "sts:AssumeRole"
        },
        {
            "Effect": "Allow",
            "Principal": {
                "Federated": "arn:aws:iam::${ACCOUNT_ID}:oidc-provider/oidc.eks.${REGION}.amazonaws.com/id/${OIDC_ID}"
            },
            "Action": "sts:AssumeRoleWithWebIdentity",
            "Condition": {
                "StringLike": {
                    "oidc.eks.${REGION}.amazonaws.com/id/${OIDC_ID}:aud": "sts.amazonaws.com",
                    "oidc.eks.${REGION}.amazonaws.com/id/${OIDC_ID}:sub": "system:serviceaccount:*:*"
                }
            }
        }
    ]
}
EOF

# Create permission policy JSON
cat << EOF > permission-policy.json
{
    "Version": "2012-10-17",
    "Statement": [
        {
            "Effect": "Allow",
            "Action": [
                "s3:ListBucket",
                "s3:PutObject",
                "s3:GetObject",
                "s3:DeleteObject"
            ],
            "Resource": [
                "arn:aws:s3:::$BUCKET_NAME"
                "arn:aws:s3:::$BUCKET_NAME/*"
            ]
        },
        {
            "Effect": "Allow",
            "Action": [
                "ecr:GetAuthorizationToken"
            ],
            "Resource": "*"
        },
        {
            "Effect": "Allow",
            "Action": [
                "ecr:BatchCheckLayerAvailability",
                "ecr:GetDownloadUrlForLayer",
                "ecr:GetRepositoryPolicy",
                "ecr:DescribeRepositories",
                "ecr:ListImages",
                "ecr:DescribeImages",
                "ecr:BatchGetImage",
                "ecr:GetLifecyclePolicy",
                "ecr:GetLifecyclePolicyPreview",
                "ecr:ListTagsForResource",
                "ecr:DescribeImageScanFindings"
            ],
            "Resource": "*"
        },
        {
            "Effect": "Allow",
            "Action": [
                "ec2:AssignPrivateIpAddresses",
                "ec2:AttachNetworkInterface",
                "ec2:CreateNetworkInterface",
                "ec2:DeleteNetworkInterface",
                "ec2:DescribeInstances",
                "ec2:DescribeTags",
                "ec2:DescribeNetworkInterfaces",
                "ec2:DescribeInstanceTypes",
                "ec2:DescribeSubnets",
                "ec2:DetachNetworkInterface",
                "ec2:DescribeDhcpOptions",
                "ec2:ModifyNetworkInterfaceAttribute",
                "ec2:UnassignPrivateIpAddresses",
                "ec2:CreateTags",
                "ec2:DescribeRouteTables",
                "ec2:DescribeSecurityGroups",
                "ec2:DescribeVolumes",
                "ec2:DescribeVolumesModifications",
                "ec2:DescribeVpcs"
            ],
            "Resource": "*"
        },
        {
            "Effect": "Allow",
            "Action": [
                "eks:Describe*",
                "eks:List*",
                "eks:AssociateAccessPolicy",
                "eks:AccessKubernetesApi",
                "eks-auth:AssumeRoleForPodIdentity"
            ],
            "Resource": "*"
        },
        {
            "Effect": "Allow",
            "Action": [
                "elasticloadbalancing:Create*",
                "elasticloadbalancing:Describe*"
            ],
            "Resource": "*"
        },
        {
            "Effect": "Allow",
            "Action": [
                "sagemaker:CreateModel",
                "sagemaker:DescribeModel",
                "sagemaker:DeleteModel",
                "sagemaker:ListModels",
                "sagemaker:CreateEndpointConfig",
                "sagemaker:DescribeEndpointConfig",
                "sagemaker:DeleteEndpointConfig",
                "sagemaker:CreateEndpoint",
                "sagemaker:DeleteEndpoint",
                "sagemaker:DescribeEndpoint",
                "sagemaker:UpdateEndpoint",
                "sagemaker:ListTags",
                "sagemaker:EnableClusterInference",
                "sagemaker:DescribeClusterInference",
                "sagemaker:DescribeHubContent"
            ],
            "Resource": "arn:aws:sagemaker:$REGION:*:*"
        },
        {
            "Effect": "Allow",
            "Action": [
                "fsx:DescribeFileSystems"
            ],
            "Resource": "*"
        },
        {
            "Effect": "Allow",
            "Action": [
                "acm:ImportCertificate",
                "acm:DeleteCertificate"
            ],
            "Resource": "arn:aws:acm:$REGION:$ACCOUNT_ID:certificate/*"
        },
        {
            "Sid": "AllowPassRoleToSageMaker",
            "Effect": "Allow",
            "Action": [
                "iam:PassRole"
            ],
            "Resource": "arn:aws:iam::$ACCOUNT_ID:role/$HYPERPOD_INFERENCE_ROLE_NAME",
            "Condition": {
                "StringEquals": {
                    "iam:PassedToService": "sagemaker.amazonaws.com"
                }
            }
        },
        {
            "Sid": "CloudWatchEMFPermissions",
            "Effect": "Allow",
            "Action": [
                "cloudwatch:PutMetricData",
                "logs:PutLogEvents",
                "logs:DescribeLogStreams",
                "logs:DescribeLogGroups",
                "logs:CreateLogStream",
                "logs:CreateLogGroup"
            ],
            "Resource": "*"
        }
    ]
}
EOF

為推論運算子建立執行角色。

aws iam create-policy --policy-name $HYPERPOD_INFERENCE_ACCESS_POLICY_NAME --policy-document file://permission-policy.json
export policy_arn="arn:aws:iam::${ACCOUNT_ID}:policy/$HYPERPOD_INFERENCE_ACCESS_POLICY_NAME"

# Create the IAM role
eksctl create iamserviceaccount --approve --role-only --name=$HYPERPOD_INFERENCE_SA_NAME --namespace=$HYPERPOD_INFERENCE_SA_NAMESPACE --cluster=$EKS_CLUSTER_NAME --attach-policy-arn=$policy_arn --role-name=$HYPERPOD_INFERENCE_ROLE_NAME --region=$REGION

aws iam create-role --role-name $HYPERPOD_INFERENCE_ROLE_NAME --assume-role-policy-document file://trust-policy.json
 
aws iam put-role-policy --role-name $HYPERPOD_INFERENCE_ROLE_NAME --policy-name InferenceOperatorInlinePolicy --policy-document file://permission-policy.json

下載並建立 AWS Load Balancer控制器所需的 IAM 政策，以在 EKS 叢集中管理 Application Load Balancer 和 Network Load Balancer。

%%bash -x 

export ALBController_IAM_POLICY_NAME=HyperPodInferenceALBControllerIAMPolicy

curl -o AWSLoadBalancerControllerIAMPolicy.json https://raw.githubusercontent.com/kubernetes-sigs/aws-load-balancer-controller/v2.13.0/docs/install/iam_policy.json
aws iam create-policy --policy-name $ALBController_IAM_POLICY_NAME --policy-document file://AWSLoadBalancerControllerIAMPolicy.json

建立將 Kubernetes 服務帳戶與 IAM 政策連結的 IAM 服務帳戶，讓 AWS Load Balancer控制器能夠透過 IRSA （服務帳戶的 IAM 角色） 取得必要的 AWS 許可。

%%bash -x 

export ALB_POLICY_ARN="arn:aws:iam::$ACCOUNT_ID:policy/$ALBController_IAM_POLICY_NAME"

# Create IAM service account with gathered values
eksctl create iamserviceaccount \
  --approve \
  --override-existing-serviceaccounts \
  --name=aws-load-balancer-controller \
  --namespace=kube-system \
  --cluster=$EKS_CLUSTER_NAME \
  --attach-policy-arn=$ALB_POLICY_ARN \
  --region=$REGION

# Print the values for verification
echo "Cluster Name: $EKS_CLUSTER_NAME"
echo "Region: $REGION"
echo "Policy ARN: $ALB_POLICY_ARN"

將標籤 (kubernetes.io.role/elb) 套用至 EKS 叢集中的所有子網路 （公有和私有）。

export VPC_ID=$(aws --region $REGION eks describe-cluster --name $EKS_CLUSTER_NAME --query 'cluster.resourcesVpcConfig.vpcId' --output text)

# Add Tags
aws ec2 describe-subnets \
  --filters "Name=vpc-id,Values=${VPC_ID}" "Name=map-public-ip-on-launch,Values=true" \
  --query 'Subnets[*].SubnetId' --output text | \
tr '\t' '\n' | \
xargs -I{} aws ec2 create-tags --resources {} --tags Key=kubernetes.io/role/elb,Value=1

# Verify Tags are added
aws ec2 describe-subnets \
  --filters "Name=vpc-id,Values=${VPC_ID}" "Name=map-public-ip-on-launch,Values=true" \
  --query 'Subnets[*].SubnetId' --output text | \
  tr '\t' '\n' |
xargs -n1 -I{} aws ec2 describe-tags --filters "Name=resource-id,Values={}" "Name=key,Values=kubernetes.io/role/elb" --query "Tags[0].Value" --output text

建立 KEDA 和 Cert Manager 的命名空間。

kubectl create namespace keda
kubectl create namespace cert-manager

建立 Amazon S3 VPC 端點。

aws ec2 create-vpc-endpoint \
  --vpc-id ${VPC_ID} \
  --vpc-endpoint-type Gateway \
  --service-name "com.amazonaws.${REGION}.s3" \
  --route-table-ids $(aws ec2 describe-route-tables --filters "Name=vpc-id,Values=${VPC_ID}" --query 'RouteTables[].Associations[].RouteTableId' --output text | tr ' ' '\n' | sort -u | tr '\n' ' ')

設定 S3 儲存體存取：

（選用） 設定 FSx 儲存體存取。為 Amazon FSx CSI 驅動程式建立 IAM 服務帳戶。FSx CSI 驅動程式將使用此服務帳戶來代表叢集與 Amazon FSx 服務互動。

%%bash -x 


eksctl create iamserviceaccount \
  --name fsx-csi-controller-sa \
  --namespace kube-system \
  --cluster $EKS_CLUSTER_NAME \
  --attach-policy-arn arn:aws:iam::aws:policy/AmazonFSxFullAccess \
  --approve \
  --role-name FSXLCSI-${EKS_CLUSTER_NAME}-${REGION} \
  --region $REGION

建立信任政策和許可政策。

# Create trust policy
cat <<EOF > /tmp/keda-trust-policy.json
{
    "Version": "2012-10-17",
    "Statement": [
        {
            "Effect": "Allow",
            "Principal": {
                "Federated": "arn:aws:iam::$ACCOUNT_ID:oidc-provider/oidc.eks.$REGION.amazonaws.com/id/$OIDC_ID"
            },
            "Action": "sts:AssumeRoleWithWebIdentity",
            "Condition": {
                "StringLike": {
                    "oidc.eks.$REGION.amazonaws.com/id/$OIDC_ID:sub": "system:serviceaccount:kube-system:keda-operator",
                    "oidc.eks.$REGION.amazonaws.com/id/$OIDC_ID:aud": "sts.amazonaws.com"
                }
            }
        }
    ]
}
EOF
 # Create permissions policy
cat <<EOF > /tmp/keda-policy.json
{
    "Version": "2012-10-17",
    "Statement": [
        {
            "Effect": "Allow",
            "Action": [
                "cloudwatch:GetMetricData",
                "cloudwatch:GetMetricStatistics",
                "cloudwatch:ListMetrics"
            ],
            "Resource": "*"
        },
        {
            "Effect": "Allow",
            "Action": [
                "aps:QueryMetrics",
                "aps:GetLabels",
                "aps:GetSeries",
                "aps:GetMetricMetadata"
            ],
            "Resource": "*"
        }
    ]
}
EOF
 # Create the role
aws iam create-role \
    --role-name keda-operator-role \
    --assume-role-policy-document file:///tmp/keda-trust-policy.json
 # Create the policy
KEDA_POLICY_ARN=$(aws iam create-policy \
    --policy-name KedaOperatorPolicy \
    --policy-document file:///tmp/keda-policy.json \
    --query 'Policy.Arn' \
    --output text)
 # Attach the policy to the role
aws iam attach-role-policy \
    --role-name keda-operator-role \
    --policy-arn $KEDA_POLICY_ARN

如果您使用的是門控模型，請建立 IAM 角色來存取門控模型。

安裝 HyperPod 推論運算子。此步驟會收集所需的 AWS 資源識別符，並使用適當的組態參數產生 Helm 安裝命令。

從 https://github.com/aws/sagemaker-hyperpod-cli/tree/main/helm_chart：// 存取 Helm Chart。

git clone https://github.com/aws/sagemaker-hyperpod-cli
cd sagemaker-hyperpod-cli
cd helm_chart/HyperPodHelmChart

%%bash -x

HYPERPOD_INFERENCE_ROLE_ARN=$(aws iam get-role --role-name=$HYPERPOD_INFERENCE_ROLE_NAME --query "Role.Arn" --output text)
echo $HYPERPOD_INFERENCE_ROLE_ARN
 
S3_CSI_ROLE_ARN=$(aws iam get-role --role-name=$S3_CSI_ROLE_NAME --query "Role.Arn" --output text)
echo $S3_CSI_ROLE_ARN

HYPERPOD_CLUSTER_ARN=$(aws sagemaker describe-cluster --cluster-name $HYPERPOD_CLUSTER_NAME --query "ClusterArn")

# Verify values
echo "Cluster Name: $EKS_CLUSTER_NAME"
echo "Execution Role: $HYPERPOD_INFERENCE_ROLE_ARN"
echo "Hyperpod ARN: $HYPERPOD_CLUSTER_ARN"
# Run the the HyperPod inference operator installation. 

helm install hyperpod-inference-operator charts/inference-operator
-n kube-system \
--set region=$REGION \
 --set eksClusterName=$EKS_CLUSTER_NAME \
--set hyperpodClusterArn=$HYPERPOD_CLUSTER_ARN \
--set executionRoleArn=$HYPERPOD_INFERENCE_ROLE_ARN \
--set s3.serviceAccountRoleArn=$S3_CSI_ROLE_ARN \
--set s3.node.serviceAccount.create=false \
--set keda.podIdentity.aws.irsa.roleArn="arn:aws:iam::$ACCOUNT_ID:role/keda-operator-role" \
--set tlsCertificateS3Bucket="s3://$BUCKET_NAME" \
--set alb.region=$REGION \
--set alb.clusterName=$EKS_CLUSTER_NAME \
--set alb.vpcId=$VPC_ID
 
# For JumpStart Gated Model usage, Add
# --set jumpstartGatedModelDownloadRoleArn=$UMPSTART_GATED_ROLE_ARN

設定 IAM 整合的服務帳戶註釋。此註釋可讓運算子的服務帳戶取得必要的 IAM 許可，以管理推論端點並與 AWS 服務互動。

%%bash -x 

EKS_CLUSTER_ROLE_NAME=$(echo $EKS_CLUSTER_ROLE | sed 's/.*\///')

# Annotate service account
kubectl annotate serviceaccount hyperpod-inference-operator-controller-manager \
  -n hyperpod-inference-system \
  eks.amazonaws.com/role-arn=arn:aws:iam::${ACCOUNT_ID}:role/${EKS_CLUSTER_ROLE_NAME} \
  --overwrite

建立模型部署組態檔案。這會建立 Kubernetes 資訊清單檔案，為 HyperPod 推論運算子定義 JumpStart 模型部署。組態指定如何從 Amazon SageMaker JumpStart 部署預先訓練的模型，做為 Amazon EKS 叢集上的推論端點。

cat <<EOF>> simple_model_install.yaml
---
apiVersion: inference.sagemaker.aws.amazon.com/v1alpha1
kind: JumpStartModel
metadata:
  name: testing-deployment-bert
  namespace: default
spec:
  model:
    modelId: "huggingface-eqa-bert-base-cased"
  sageMakerEndpoint:
    name: "hp-inf-ep-for-testing"
  server:
    instanceType: "ml.c5.2xlarge"
  environmentVariables:
    - name: SAMPLE_ENV_VAR
      value: "sample_value"
  maxDeployTimeInSeconds: 1800
EOF

部署模型並清除組態檔案。此步驟會建立 JumpStart 模型資源，並移除暫時組態檔案以維護乾淨的工作區。

%%bash -x 

kubectl create -f simple_model_install.yaml
rm -rfv simple_model_install.yaml

檢查模型是否已安裝並執行。此驗證可確保運算子可以成功取得管理推論端點的 AWS 許可。

%%bash

# Get the service account details
kubectl get serviceaccount -n hyperpod-inference-system

# Check if the service account has the AWS annotations
kubectl describe serviceaccount hyperpod-inference-operator-controller-manager -n hyperpod-inference-system

寫入模型輸入檔案。這會建立包含範例資料的 JSON 輸入檔案，以測試部署模型的問答功能。

%%writefile demo-input.json
{"question" :"what is the name of the planet?","context" : "earth"}

叫用 SageMaker 端點來執行負載測試，以驗證推論端點的效能和可靠性。

%%bash

#!/bin/bash

for i in {1..1000}
do
    echo "Invocation #$i"
    aws sagemaker-runtime invoke-endpoint \
        --endpoint-name testing-deployment-jumpstart-9 \
        --region {REGION} \
        --body fileb://demo-input.json \
        --content-type application/list-text \
        --accept application/json \
        "demoout_${i}.json"
    
    # Add a small delay to prevent throttling (optional)
    #sleep 0.5
    rm -f "demoout_${i}.json"
done