[[["容易理解","easyToUnderstand","thumb-up"],["確實解決了我的問題","solvedMyProblem","thumb-up"],["其他","otherUp","thumb-up"]],[["難以理解","hardToUnderstand","thumb-down"],["資訊或程式碼範例有誤","incorrectInformationOrSampleCode","thumb-down"],["缺少我需要的資訊/範例","missingTheInformationSamplesINeed","thumb-down"],["翻譯問題","translationIssue","thumb-down"],["其他","otherDown","thumb-down"]],["上次更新時間:2025-09-04 (世界標準時間)。"],[],[],null,["# Run Arm workloads in GKE on AWS\n\nGKE on AWS lets you run Arm workloads built for Arm-based\n[AWS Graviton processors](https://aws.amazon.com/ec2/graviton/).\n\nLimitations\n-----------\n\n- Arm node pools running Kubernetes versions earlier than 1.24.8-gke.1300\n automatically add a taint during node pool creation to prevent Arm workloads\n from being scheduled on non-Arm nodes. Arm node pools in clusters at version\n 1.24.8-gke.1300 or higher no longer add this taint. If you're upgrading from a\n cluster earlier than 1.24.8-gke.1300, you must create this\n taint yourself or otherwise take this into account when upgrading.\n\n- Arm node pools on GKE on AWS don't support Cloud Service Mesh, Config Sync,\n or Policy Controller. You must run these products on an x86 node\n pool.\n\n- Clusters running Kubernetes version 1.24 need an x86 node pool to run the\n [Connect Agent](/anthos/fleet-management/docs/connect-agent).\n If your cluster runs Kubernetes version 1.25 or later, you don't need an\n x86 node pool.\n\nThis page explains how to create an Arm node pool, why\nmulti-architecture images are the recommended way to deploy Arm workloads, and\nhow to schedule Arm workloads.\n\n### Before you begin\n\nBefore you create node pools for your Arm workloads, you need the following\nresources:\n\n- An existing AWS cluster to create the node pool in. This cluster must run Kubernetes version 1.24 or later.\n- An [IAM instance profile](/kubernetes-engine/multi-cloud/docs/aws/how-to/create-aws-iam-roles#create_a_node_pool_iam_role) for the node pool VMs.\n- A [subnet](/kubernetes-engine/multi-cloud/docs/aws/how-to/create-aws-vpc#create_the_node_pool_subnets) where the node pool VMs will run.\n- If your cluster is running Kubernetes version 1.24, an x86 node pool to run\n the [Connect Agent](/anthos/fleet-management/docs/connect-agent).\n\n For details on how to create a node pool in GKE on AWS, see\n [Create a node pool](/kubernetes-engine/multi-cloud/docs/aws/how-to/create-node-pool).\n\n### Create an Arm node pool\n\nGKE on AWS supports node pools built on the Canonical Ubuntu\narm64 minimal node image and `containerd` runtime.\n\nTo create an Arm node pool and add it to an existing cluster, run the following\ncommand: \n\n gcloud container aws node-pools create \u003cvar translate=\"no\"\u003eNODE_POOL_NAME\u003c/var\u003e \\\n --cluster \u003cvar translate=\"no\"\u003eCLUSTER_NAME\u003c/var\u003e \\\n --instance-type \u003cvar translate=\"no\"\u003eINSTANCE_TYPE\u003c/var\u003e \\\n --root-volume-size \u003cvar translate=\"no\"\u003eROOT_VOLUME_SIZE\u003c/var\u003e \\\n --iam-instance-profile \u003cvar translate=\"no\"\u003eNODEPOOL_PROFILE\u003c/var\u003e \\\n --node-version \u003cvar translate=\"no\"\u003eNODE_VERSION\u003c/var\u003e \\\n --min-nodes \u003cvar translate=\"no\"\u003eMIN_NODES\u003c/var\u003e \\\n --max-nodes \u003cvar translate=\"no\"\u003eMAX_NODES\u003c/var\u003e \\\n --max-pods-per-node \u003cvar translate=\"no\"\u003eMAX_PODS_PER_NODE\u003c/var\u003e \\\n --location \u003cvar translate=\"no\"\u003eGOOGLE_CLOUD_LOCATION\u003c/var\u003e \\\n --subnet-id \u003cvar translate=\"no\"\u003eNODEPOOL_SUBNET\u003c/var\u003e \\\n --ssh-ec2-key-pair \u003cvar translate=\"no\"\u003eSSH_KEY_PAIR_NAME\u003c/var\u003e \\\n --config-encryption-kms-key-arn \u003cvar translate=\"no\"\u003eCONFIG_KMS_KEY_ARN\u003c/var\u003e\n\nReplace the following:\n\n- \u003cvar translate=\"no\"\u003eNODE_POOL_NAME\u003c/var\u003e: the name you choose for your node pool\n- \u003cvar translate=\"no\"\u003eCLUSTER_NAME\u003c/var\u003e: the name of the cluster to attach the node pool to\n- \u003cvar translate=\"no\"\u003eINSTANCE_TYPE\u003c/var\u003e: one of the following instance types:\n\n - `m6g`\n - `m6gd`\n - `t4g`\n - `r6g`\n - `r6gd`\n - `c6g`\n - `c6gd`\n - `c6gn`\n - `x2gd`\n - `c7g`\n - `im4gn`\n - `g5g`\n\n These instance types are powered by Arm-based AWS Graviton processors.\n You also need to specify the instance size that you want. For example,\n `m6g.medium`. For a complete list, see\n [Supported AWS instance types](/kubernetes-engine/multi-cloud/docs/aws/reference/supported-instance-types#supported_instance_types).\n- \u003cvar translate=\"no\"\u003eROOT_VOLUME_SIZE\u003c/var\u003e: the desired size for each node's root\n volume, in Gb\n\n- \u003cvar translate=\"no\"\u003eNODEPOOL_PROFILE\u003c/var\u003e: the IAM instance profile\n for node pool VMs\n\n- \u003cvar translate=\"no\"\u003eNODE_VERSION\u003c/var\u003e: the Kubernetes version to install on each\n node in the node pool, which must be version 1.24 or later. For example,\n `1.24.3-gke.200`.\n\n- \u003cvar translate=\"no\"\u003eMIN_NODES\u003c/var\u003e: the minimum number of nodes the node pool can\n contain\n\n- \u003cvar translate=\"no\"\u003eMAX_NODES\u003c/var\u003e: the maximum number of nodes the node pool can\n contain\n\n- \u003cvar translate=\"no\"\u003eMAX_PODS_PER_NODE\u003c/var\u003e: the maximum number of Pods that can be\n created on any single node in the pool\n\n- \u003cvar translate=\"no\"\u003eGOOGLE_CLOUD_LOCATION\u003c/var\u003e: the name of the Google Cloud\n\n location from which this node pool will be managed\n- \u003cvar translate=\"no\"\u003eNODEPOOL_SUBNET\u003c/var\u003e: the ID of the subnet the node pool will\n run on. If this subnet is outside of the VPC's primary CIDR block, you need\n to take additional steps. For more information, see\n [security groups](/kubernetes-engine/multi-cloud/docs/aws/reference/security-groups).\n\n- \u003cvar translate=\"no\"\u003eSSH_KEY_PAIR_NAME\u003c/var\u003e: the name of the AWS SSH key pair\n created for SSH access (optional)\n\n- \u003cvar translate=\"no\"\u003eCONFIG_KMS_KEY_ARN\u003c/var\u003e: the Amazon Resource Name (ARN) of the\n AWS KMS key that encrypts user data\n\nUnderstand multi-architecture images\n------------------------------------\n\nContainer images must be compatible with the architecture of the node where you\nintend to run the Arm workloads. To ensure that your container image is\nArm-compatible, we recommend that you use multi-architecture (\"multi-arch\")\nimages.\n\nA multi-arch image is an image that can support multiple architectures. It looks\nlike a single image with a single tag, but contains a set of images to run\non different machine architectures. Multi-arch images are\ncompatible with the Docker Image Manifest V2 Scheme 2 or OCI Image Index\nSpecifications.\n\nWhen you deploy a multi-arch image to a cluster, the container runtime\nautomatically chooses the image that is compatible with the architecture of the\nnode you're deploying to. Once you have a multi-arch image for a workload, you\ncan deploy this workload across multiple architectures. Scheduling a\nsingle-architecture image onto an incompatible node causes an error at load\ntime.\n\nTo learn more about how to use multi-arch images with Arm workloads, see\n[Build multi-architecture images for Arm workloads](/kubernetes-engine/docs/how-to/build-multi-arch-for-arm)\nin the Google Kubernetes Engine (GKE) documentation.\n\nWhat's next\n-----------\n\n- [Troubleshoot Arm workloads](/kubernetes-engine/multi-cloud/docs/aws/troubleshooting#troubleshoot-arm)."]]