이 페이지에서는 Google Kubernetes Engine(GKE)에서 ARM 워크로드를 실행하는 방법을 설명합니다.
개요
Performance 또는 Scale-Out컴퓨팅 클래스를 사용하여 GKE Autopilot 클러스터에서 Arm 워크로드를 실행하거나 C4A 머신 시리즈(C4A) 또는 Tau T2A 머신 시리즈(T2A)를 사용하여 GKE Standard 클러스터에서 실행할 수 있습니다. x86 및 Arm 프로세서와 호환되는 단일 아키텍처 Arm 이미지 또는 멀티 아키텍처 이미지를 실행할 수 있습니다. Arm의 이점에 대한 자세한 내용은 컴퓨팅에 사용되는 Arm VM을 참조하세요.
ARM에 배포할 워크로드를 선택하고 이러한 워크로드를 배포할 수 있도록 준비하는 방법에 대한 자세한 내용은 다음 가이드를 참조하세요.
Arm에서 실행할 워크로드 선택: C4A 노드는 성능에 가장 민감한 Arm 기반 워크로드에 일관적으로 높은 성능을 제공하는 Arm 기반 컴퓨팅을 제공합니다. T2A 노드는 더 유연한 워크로드나 수평 확장에 의존하는 워크로드에 적합합니다. 각 머신 시리즈에 적합한 워크로드 유형에 대해 자세히 알아보려면 Compute Engine용 범용 머신 계열의 표를 참고하세요.
아키텍처에 배포: GKE에서는 멀티 아키텍처 이미지를 사용하여 Arm을 포함한 다양한 아키텍처를 사용하는 노드에 하나의 이미지 매니페스트를 배포할 수 있습니다.
[[["이해하기 쉬움","easyToUnderstand","thumb-up"],["문제가 해결됨","solvedMyProblem","thumb-up"],["기타","otherUp","thumb-up"]],[["이해하기 어려움","hardToUnderstand","thumb-down"],["잘못된 정보 또는 샘플 코드","incorrectInformationOrSampleCode","thumb-down"],["필요한 정보/샘플이 없음","missingTheInformationSamplesINeed","thumb-down"],["번역 문제","translationIssue","thumb-down"],["기타","otherDown","thumb-down"]],["최종 업데이트: 2025-09-01(UTC)"],[],[],null,["# Arm workloads on GKE\n\nAutopilot Standard\n\n*** ** * ** ***\n\nThis page explains how you can run Arm workloads on Google Kubernetes Engine (GKE).\n\nOverview\n--------\n\nYou can run Arm workloads in GKE Autopilot clusters\nusing the `Performance` or `Scale-Out` [compute\nclasses](/kubernetes-engine/docs/concepts/autopilot-compute-classes#when-to-use),\nor in GKE Standard\nclusters using the [C4A machine series\n(C4A)](/compute/docs/general-purpose-machines#c4a_series) or [Tau T2A machine\nseries (T2A)](/compute/docs/general-purpose-machines#t2a_machines). You can run\nsingle-architecture Arm images or multi-architecture (multi-arch) images\ncompatible with both x86 and Arm processors. To learn about the benefits of Arm,\nsee [Arm VMs on Compute](/compute/docs/instances/arm-on-compute).\n\nSee the following guides for more information about choosing workloads to deploy on Arm and preparing those\nworkloads for deployment:\n\n- **Choosing workloads to run on Arm** : C4A nodes provide Arm-based compute which achieves consistently high performance for your most performance-sensitive Arm-based workloads. T2A nodes are appropriate for more-flexible workloads, or workloads which rely on horizontal scale-out. To learn more about what types of workloads work well with each of these machine series, see the table in [General-purpose machine family for\n Compute Engine](/compute/docs/general-purpose-machines).\n- **Deploying across architectures** : With GKE, you can use multi-arch images to deploy one image manifest across nodes with different architectures, including Arm.\n - To ensure that your container image is Arm-compatible and can run on your targeted architectures, see [Build multi-architecture images for\n Arm workloads](/kubernetes-engine/docs/how-to/build-multi-arch-for-arm).\n - To follow a tutorial for using multi-arch images to deploy across architectures, see [Migrate x86 application on GKE to\n multi-arch with\n Arm](/kubernetes-engine/docs/tutorials/migrate-x86-to-multi-arch-arm).\n- **Preparing Arm workloads for deployment** : Once you have an Arm-compatible image, use [node\n affinity](https://kubernetes.io/docs/concepts/scheduling-eviction/assign-pod-node/#node-affinity) rules and [node selectors](https://kubernetes.io/docs/concepts/scheduling-eviction/assign-pod-node/#nodeselector) to make sure your workload is scheduled to nodes with a compatible architecture type.\n - **Autopilot clusters** : see [Deploy Autopilot\n workloads on Arm\n architecture](/kubernetes-engine/docs/how-to/autopilot-arm-workloads).\n - **Standard clusters** : see [Prepare an Arm workload for\n deployment](/kubernetes-engine/docs/how-to/prepare-arm-workloads-for-deployment).\n\nRequirements and limitations\n----------------------------\n\n- To create a cluster with C4A nodes that uses [Autopilot](/kubernetes-engine/docs/concepts/autopilot-overview)\n mode, [cluster\n autoscaling](/kubernetes-engine/docs/concepts/cluster-autoscaler),\n or [node\n auto-provisioning](/kubernetes-engine/docs/how-to/node-auto-provisioning),\n you must use the following versions or later:\n\n - 1.28.15-gke.1344000\n - 1.29.11-gke.1012000\n - 1.30.7-gke.1136000\n - 1.31.3-gke.1056000\n- To create a Standard cluster with C4A nodes, you must use one of the\n following versions or later:\n\n - 1.28.13-gke.1024000\n - 1.29.8-gke.1057000\n - 1.30.4-gke.1213000\n- Arm nodes are available in Google Cloud locations that support Arm\n architecture. For details, see [Available regions and\n zones](/compute/docs/regions-zones#available).\n\n- You can use [Local\n SSDs](/kubernetes-engine/docs/how-to/persistent-volumes/local-ssd) with C4A\n nodes with the following versions or later:\n\n - 1.29.15-gke.1325000\n - 1.30.12-gke.1033000\n - 1.31.8-gke.1045000\n - 1.32.1-gke.1357000\n- GKE doesn't support the following features with C4A nodes:\n\n - [Confidential GKE Nodes](/kubernetes-engine/docs/how-to/confidential-gke-nodes)\n - [Compact placement](/kubernetes-engine/docs/how-to/compact-placement)\n - [Simultaneous multi-threading (SMT)](/kubernetes-engine/docs/how-to/configure-smt)\n - [Persistent disks](/kubernetes-engine/docs/concepts/persistent-volumes) (use [Hyperdisk](/kubernetes-engine/docs/concepts/hyperdisk) instead, see [Supported disk types for\n C4A](/compute/docs/general-purpose-machines#supported_disk_types_for_c4a))\n - [Nested virtualization](/kubernetes-engine/docs/how-to/nested-virtualization)\n - [GPUs](/kubernetes-engine/docs/concepts/gpus)\n- GKE doesn't support the following features with T2A\n nodes:\n\n - [Confidential GKE Nodes](/kubernetes-engine/docs/how-to/confidential-gke-nodes)\n - [GPUs](/kubernetes-engine/docs/concepts/gpus)\n - [GKE Windows](/kubernetes-engine/docs/concepts/windows-server-gke)\n - [Local SSDs](/kubernetes-engine/docs/how-to/persistent-volumes/local-ssd)\n - [Policy Controller](/anthos-config-management/docs/concepts/policy-controller), [Config Sync](/anthos-config-management/docs/config-sync-overview), and [Config Controller](/anthos-config-management/docs/concepts/config-controller-overview)\n\nWhat's next\n-----------\n\n- [Create clusters and node pools with Arm nodes](/kubernetes-engine/docs/how-to/create-arm-clusters-nodes)\n- [Build multi-architecture images for Arm workloads](/kubernetes-engine/docs/how-to/build-multi-arch-for-arm)\n- [Prepare an Arm workload for deployment](/kubernetes-engine/docs/how-to/prepare-arm-workloads-for-deployment)\n- [Migrate x86 application on GKE to multi-arch with Arm](/kubernetes-engine/docs/tutorials/migrate-x86-to-multi-arch-arm)"]]
