[[["์ดํดํ๊ธฐ ์ฌ์","easyToUnderstand","thumb-up"],["๋ฌธ์ ๊ฐ ํด๊ฒฐ๋จ","solvedMyProblem","thumb-up"],["๊ธฐํ","otherUp","thumb-up"]],[["์ดํดํ๊ธฐ ์ด๋ ค์","hardToUnderstand","thumb-down"],["์๋ชป๋ ์ ๋ณด ๋๋ ์ํ ์ฝ๋","incorrectInformationOrSampleCode","thumb-down"],["ํ์ํ ์ ๋ณด/์ํ์ด ์์","missingTheInformationSamplesINeed","thumb-down"],["๋ฒ์ญ ๋ฌธ์ ","translationIssue","thumb-down"],["๊ธฐํ","otherDown","thumb-down"]],["์ต์ข ์ ๋ฐ์ดํธ: 2025-08-27(UTC)"],[[["\u003cp\u003eDataproc on GKE virtual clusters utilize node pools to run jobs, and if a specified node pool doesn't exist, Dataproc on GKE will create it with the designated settings.\u003c/p\u003e\n"],["\u003cp\u003eYou can define various settings for Dataproc on GKE node pools, including \u003ccode\u003eaccelerators\u003c/code\u003e, \u003ccode\u003emachineType\u003c/code\u003e, \u003ccode\u003eminNodeCount\u003c/code\u003e, \u003ccode\u003emaxNodeCount\u003c/code\u003e, \u003ccode\u003epreemptible\u003c/code\u003e, and \u003ccode\u003espot\u003c/code\u003e, among others.\u003c/p\u003e\n"],["\u003cp\u003eWhen a Dataproc on GKE cluster is deleted, the associated node pools are not automatically deleted, and they need to be deleted separately.\u003c/p\u003e\n"],["\u003cp\u003eNode pool zones must reside within the same region as the associated virtual cluster, and they can be specified when creating or updating the virtual cluster.\u003c/p\u003e\n"],["\u003cp\u003eNode pools are assigned roles such as \u003ccode\u003edefault\u003c/code\u003e, \u003ccode\u003espark-driver\u003c/code\u003e, and \u003ccode\u003espark-executor\u003c/code\u003e, with at least one node pool needing to be assigned the \u003ccode\u003edefault\u003c/code\u003e role.\u003c/p\u003e\n"]]],[],null,["When you\n[create](/dataproc/docs/guides/dpgke/quickstarts/dataproc-gke-quickstart-create-cluster) or\n[update](/dataproc/docs/guides/dpgke/dataproc-gke-recreate-cluster)\na Dataproc on GKE virtual cluster, you specify one or more node pools that\nthe virtual cluster will use to run jobs (this cluster is referred to as the\ncluster \"used by\" or \"associated\" with the specified node pools). If a specified node pool\ndoes not exist on your GKE cluster, Dataproc on GKE\nwill create the node pool on the GKE cluster with settings\nyou specify. If the node pool exists and was created by Dataproc,\nit will be validated to confirm that its settings match the specified settings.\n\nDataproc on GKE node pool settings\n\nYou can specify the following\n[settings](/dataproc/docs/reference/rest/v1/projects.regions.clusters#gkenodeconfig)\non node pools used by your Dataproc on GKE virtual clusters (these\nsettings are a subset of\n[GKE node pool settings](/kubernetes-engine/docs/reference/rest/v1/NodeConfig)):\n\n- `accelerators`\n- `acceleratorCount`\n- `acceleratorType`\n- `gpuPartitionSize`\\*\n- `localSsdCount`\n- `machineType`\n- `minCpuPlatform`\n- `minNodeCount`\n- `maxNodeCount`\n- `preemptible`\n- `spot`\\*\n\nNotes:\n\n- `gpuPartitionSize` can be set in the Dataproc API [`GkeNodePoolAcceleratorConfig`](/dataproc/docs/reference/rest/v1/GkeClusterConfig#gkenodepoolacceleratorconfig).\n- `spot` can be set in the Dataproc API [GkeNodeConfig](/dataproc/docs/reference/rest/v1/projects.regions.clusters#gkenodeconfig).\n\nNode pool deletion\n\nWhen a Dataproc on GKE cluster is deleted, the node pools used by the cluster\nare not deleted. See [Delete a node pool](/kubernetes-engine/docs/how-to/node-pools#deleting_a_node_pool)\nto delete node pools no longer in use by Dataproc on GKE clusters.\n\nNode pool location\n\nYou can specify the\n[zone](/dataproc/docs/reference/rest/v1/projects.regions.clusters#GkeNodePoolConfig.FIELDS.locations)\nlocation of node pools associated with your Dataproc on GKE virtual cluster\nwhen you create or update the virtual cluster. The node pool zones must be\nlocated in the region of the associated virtual cluster.\n\nRole to node pool mapping\n\nNode pool [roles](/dataproc/docs/reference/rest/v1/projects.regions.clusters#Role)\nare defined for Spark driver and executor work, with a default role\ndefined for all types of work by a node pool. Dataproc on GKE clusters must have\nat least one a node pool that is assigned the `default` role.\nAssigning other roles is optional.\n\n**Recommendation:** Create separate node pools for each role type, with node type\nand size based on role requirements.\n\ngcloud CLI virtual cluster creation example: \n\n```\ngcloud dataproc clusters gke create \"${DP_CLUSTER}\" \\\n --region=${REGION} \\\n --gke-cluster=${GKE_CLUSTER} \\\n --spark-engine-version=latest \\\n --staging-bucket=${BUCKET} \\\n --pools=\"name=${DP_POOLNAME},roles=default \\\n --setup-workload-identity\n --pools=\"name=${DP_CTRL_POOLNAME},roles=default,machineType=e2-standard-4\" \\\n --pools=\"name=${DP_DRIVER_POOLNAME},min=1,max=3,roles=spark-driver,machineType=n2-standard-4\" \\\n --pools=\"name=${DP_EXEC_POOLNAME},min=1,max=10,roles=spark-executor,machineType=n2-standard-8\"\n```"]]
