[[["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)。"],[[["\u003cp\u003eDataproc images bundle Google Cloud connectors and Apache Spark & Hadoop components, deploying them as a single package on a Dataproc cluster.\u003c/p\u003e\n"],["\u003cp\u003eImage versions are structured as \u003ccode\u003emajor.minor.subminor-os_distribution\u003c/code\u003e, allowing selection of specific software sets when creating clusters, with the OS distribution being mandatory for Rocky Linux and Ubuntu images.\u003c/p\u003e\n"],["\u003cp\u003eSpecifying \u003ccode\u003emajor.minor\u003c/code\u003e image versions is recommended for production environments to maintain compatibility, automatically selecting the latest weekly subminor and OS distribution release.\u003c/p\u003e\n"],["\u003cp\u003eNew major, minor, and subminor image versions are released periodically to incorporate updates, fixes, and new releases for components like Spark, Hadoop, and Google Cloud connectors.\u003c/p\u003e\n"],["\u003cp\u003eClusters using supported minor image versions are eligible for support for 24 months after their general availability, requiring recreation using the latest subminor version to receive fixes.\u003c/p\u003e\n"]]],[],null,["Dataproc uses images to tie together useful Google Cloud\nconnectors and Apache Spark \\& Apache Hadoop components into one package that\ncan be deployed on a Dataproc cluster.\nThese images contain the base operating system (Debian or Ubuntu) for the\ncluster, along with [core and optional components](/dataproc/docs/concepts/components/overview)\nneeded to run jobs, such as Spark, Hadoop, and Hive.\nThese images are periodically upgraded to include new improvements and\nfeatures. Dataproc versioning lets you select sets of software\nversions when you create clusters.\n\nHow versioning works\n\nWhen an image is created, it is given an\n[image version](/dataproc/docs/concepts/versioning/dataproc-versions) number\nin the following format:\n\n`version_major.version_minor.version_sub_minor-os_distribution`\n\nThe following OS distributions are maintained:\n\n| OS Distribution Code | OS Distribution |\n|----------------------|------------------|\n| debian12 | Debian 12 |\n| debian10 | Debian 10 |\n| debian11 | Debian 11 |\n| rocky8 | Rocky Linux 8 |\n| rocky9 | Rocky Linux 9 |\n| ubuntu18 | Ubuntu 18.04 LTS |\n| ubuntu20 | Ubuntu 20.04 LTS |\n| ubuntu22 | Ubuntu 22.04 LTS |\n\nSee [old image versions](#old_image_versions) for previously supported OS\ndistributions.\n\nThe recommended practice is to specify the `major.minor` image\nversion for production environments or when compatibility with specific component\nversions is important. The subminor and OS distributions are automatically\nset to the latest weekly release.\n\nSelect versions\n\nWhen you create a new Dataproc cluster, the latest available\n**Debian** image version are used by default. You can select a\nDebian, Rocky Linux or Ubuntu image version when creating a cluster (see the\n[Dataproc image version list](/dataproc/docs/concepts/versioning/dataproc-versions)).\nWhen specifying Debian-based images, you can omit the OS Distribution\nCode suffix, for example by specifying `2.0` to select the `2.0-debian10` image.\nThe OS suffix **must** be used to select a Rocky Linux or\nUbuntu-based image, for example by specifying `2.0-ubuntu18`. \n\ngcloud command\n\nWhen using the `gcloud dataproc clusters create` command, you can\nuse the `--image-version` argument to specify an image version for\nthe new cluster.\n\n**Debian image example:** \n\n```\ngcloud dataproc clusters create CLUSTER_NAME \\\n --image-version=2.0 \\\n --region=REGION\n```\n\n**Ubuntu image example:** \n\n```\ngcloud dataproc clusters create CLUSTER_NAME \\\n --image-version=2.0-ubuntu18 \\\n --region=REGION\n```\n\nBest practice is to omit the subminor version so that the latest subminor\nversion is used. However, if necessary, the subminor version can be specified,\nfor example, `2.0.20`.\n\nYou can check your current version with the Google Cloud CLI. \n\n```\ngcloud dataproc clusters describe CLUSTER_NAME \\\n --region=REGION\n```\n\nREST API\n\nYou can specify the `SoftwareConfig`\n[imageVersion](/dataproc/docs/reference/rest/v1/ClusterConfig#SoftwareConfig)\nfield as part of a\n[cluster.create](/dataproc/docs/reference/rest/v1/projects.regions.clusters/create)\nAPI request.\n\n**Example** \n\n```\nPOST /v1/projects/project-id/regions/us-central1/clusters/\n{\n \"projectId\": \"project-id\",\n \"clusterName\": \"example-cluster\",\n \"config\": {\n \"configBucket\": \"\",\n \"gceClusterConfig\": {\n \"subnetworkUri\": \"default\",\n \"zoneUri\": \"us-central1-b\"\n },\n \"masterConfig\": {\n ...\n }\n },\n \"workerConfig\": {\n ...\n }\n },\n \"softwareConfig\": {\n \"imageVersion\": \"2.0\"\n }\n }\n}\n \n```\n\nConsole\n\nOpen the Dataproc\n[Create a cluster](https://console.cloud.google.com/dataproc/clustersAdd)\npage. The **Set up cluster** panel is selected.\nThe **Image type** and **Version** field in the\n**Versioning** section shows the image that will be used when\ncreating the cluster. The image release date is also shown. Initially, the\ndefault image, the latest available Debian version, is shown as selected.\nClick **Change** to display a lists of available images. You\ncan select a standard or\n[custom image](/dataproc/docs/guides/dataproc-images) to use for\nyour cluster.\n\nWhen new versions are created\n\nNew **major** versions are periodically created to incorporate\none or more of the following:\n\n- Major releases for:\n - Spark, Hadoop, and other Big Data components\n - Google Cloud connectors\n- Major changes or updates to Dataproc functionality\n\nNew **preview** versions (with a `-RC` suffix) are released prior to the release\nof a new major version:\n\n- Preview images are not intended for use in production workloads.\n- Preview image component versions might be upgraded to the latest available component version in the post-preview GA image version.\n\nNew **minor** versions are periodically created to incorporate\none or more of the following:\n\n- Minor releases and updates for:\n - Spark, Hadoop, and other Big Data components\n - Google Cloud connectors\n- Minor changes or updates to Dataproc functionality\n\nWhen a new minor version is created, its Debian image becomes the default for the major\nversion, and represents the latest release of the major version.\n\nNew **subminor** versions are periodically created to incorporate\none or more of the following:\n\n- Patches or fixes for a component in the image\n- Component subminor version upgrades\n\nImage version and Dataproc support\n\nMinor image versions are supported for 24 months after initial\nGA (General Availability) release. During this period, clusters using these\nimage versions are eligible for support (to receive fixes, recreate your cluster\nusing the latest supported subminor image version). After the support window\nhas closed, clusters using the image versions aren't eligible for support.\n\nOld image versions\n\nPreviously supported OS distributions\n\nThe following OS distributions were previously supported:\n\n| OS Distribution Code | OS Distribution | Last Patched (End of support) |\n|----------------------|-----------------|-------------------------------|\n| debian9 | Debian 9 | July 10, 2020 |\n| deb8 | Debian 8 | October 26, 2018 |\n\nImage versions without explicit OS distribution\n\nPrior to August 16, 2018, image versions were built with Debian 8, and omitted\nthe OS Distribution Code. They are specified in the following format:\n\n`version_major.version_minor.version_sub_minor`\n\nVersions 0.1 and 0.2\n\nImage versions released as alpha or beta releases prior to\nDataproc version `1.0` general availability\naren't subject to the\n[Dataproc support policy](#image-version-support).\n\nImportant notes about versioning\n\n- Image versions contain the following components:\n - [Core components](/dataproc/docs/concepts/components/overview) that are installed on all clusters, such as Spark, Hadoop, and Hive\n - [Optional components](/dataproc/docs/concepts/components/overview#available_optional_components) that you specify when you create a cluster\n- Your Dataproc clusters are not automatically updated when new image versions are released.\n - **Recommendations:**\n - Run clusters with the latest [subminor image version](/dataproc/docs/concepts/versioning/dataproc-version-clusters#debian_images). Image metadata includes a `previous-subminor` label, which is set to `true` if the cluster is not using the latest subminor image version.\n - To view image metadata:\n 1. Run the following [`gcloud compute images list --filter`](/sdk/gcloud/reference/compute/images/list#--filter) command to list the resource name of a Dataproc image. \n\n ```\n gcloud compute images list --project=PROJECT_NAME --filter=\"labels.goog-dataproc-version ~ ^IMAGE_VERSION (such as `2.2.16-debian12`)\"\n ```\n 2. Run the following [`gcloud compute images describe`](/sdk/gcloud/reference/compute/images/describe) to view image metadata. \n\n ```\n gcloud compute images describe --project=PROJECT_NAME IMAGE_NAME\"\n ```\n - Test and validate that your applications run successfully on clusters created with new image versions, particularly when using new major image version releases."]]
