共用方式為


Microsoft.MachineLearningServices 工作區/計算 2023-04-01-preview

Bicep 資源定義

工作區/計算資源類型可以使用目標作業來部署:

如需每個 API 版本中已變更屬性的清單,請參閱 變更記錄

資源格式

若要建立 Microsoft.MachineLearningServices/workspaces/computes 資源,請將下列 Bicep 新增至您的範本。

resource symbolicname 'Microsoft.MachineLearningServices/workspaces/computes@2023-04-01-preview' = {
  name: 'string'
  location: 'string'
  tags: {
    tagName1: 'tagValue1'
    tagName2: 'tagValue2'
  }
    {customized property}: 'string'
  sku: {
    capacity: int
    family: 'string'
    name: 'string'
    size: 'string'
    tier: 'string'
  }
  parent: resourceSymbolicName
  identity: {
    type: 'string'
    userAssignedIdentities: {
      {customized property}: {}
    }
  }
  properties: {
    computeLocation: 'string'
    description: 'string'
    disableLocalAuth: bool
    resourceId: 'string'
    computeType: 'string'
    // For remaining properties, see Compute objects
  }
}

計算物件

設定 computeType 屬性以指定物件的類型。

針對 AKS,請使用:

  computeType: 'AKS'
  properties: {
    agentCount: int
    agentVmSize: 'string'
    aksNetworkingConfiguration: {
      dnsServiceIP: 'string'
      dockerBridgeCidr: 'string'
      serviceCidr: 'string'
      subnetId: 'string'
    }
    clusterFqdn: 'string'
    clusterPurpose: 'string'
    loadBalancerSubnet: 'string'
    loadBalancerType: 'string'
    sslConfiguration: {
      cert: 'string'
      cname: 'string'
      key: 'string'
      leafDomainLabel: 'string'
      overwriteExistingDomain: bool
      status: 'string'
    }
  }

針對 AmlCompute,請使用:

  computeType: 'AmlCompute'
  properties: {
    enableNodePublicIp: bool
    isolatedNetwork: bool
    osType: 'string'
    propertyBag: any()
    remoteLoginPortPublicAccess: 'string'
    scaleSettings: {
      maxNodeCount: int
      minNodeCount: int
      nodeIdleTimeBeforeScaleDown: 'string'
    }
    subnet: {
      id: 'string'
    }
    userAccountCredentials: {
      adminUserName: 'string'
      adminUserPassword: 'string'
      adminUserSshPublicKey: 'string'
    }
    virtualMachineImage: {
      id: 'string'
    }
    vmPriority: 'string'
    vmSize: 'string'
  }

針對 ComputeInstance,請使用:

  computeType: 'ComputeInstance'
  properties: {
    applicationSharingPolicy: 'string'
    autologgerSettings: {
      mlflowAutologger: 'string'
    }
    computeInstanceAuthorizationType: 'personal'
    customServices: [
      {
        docker: {
          privileged: bool
          {customized property}: any()
        }
        endpoints: [
          {
            hostIp: 'string'
            name: 'string'
            protocol: 'string'
            published: int
            target: int
          }
        ]
        environmentVariables: {
          {customized property}: {
            type: 'local'
            value: 'string'
            {customized property}: any()
          }
        }
        image: {
          reference: 'string'
          type: 'string'
          {customized property}: any()
        }
        name: 'string'
        volumes: [
          {
            bind: {
              createHostPath: bool
              propagation: 'string'
              selinux: 'string'
            }
            consistency: 'string'
            readOnly: bool
            source: 'string'
            target: 'string'
            tmpfs: {
              size: int
            }
            type: 'string'
            volume: {
              nocopy: bool
            }
          }
        ]
        {customized property}: any()
      }
    ]
    enableNodePublicIp: bool
    idleTimeBeforeShutdown: 'string'
    personalComputeInstanceSettings: {
      assignedUser: {
        objectId: 'string'
        tenantId: 'string'
      }
    }
    schedules: {
      computeStartStop: [
        {
          action: 'string'
          cron: {
            expression: 'string'
            startTime: 'string'
            timeZone: 'string'
          }
          recurrence: {
            frequency: 'string'
            interval: int
            schedule: {
              hours: [
                int
              ]
              minutes: [
                int
              ]
              monthDays: [
                int
              ]
              weekDays: [
                'string'
              ]
            }
            startTime: 'string'
            timeZone: 'string'
          }
          schedule: {
            id: 'string'
            provisioningStatus: 'string'
            status: 'string'
          }
          status: 'string'
          triggerType: 'string'
        }
      ]
    }
    setupScripts: {
      scripts: {
        creationScript: {
          scriptArguments: 'string'
          scriptData: 'string'
          scriptSource: 'string'
          timeout: 'string'
        }
        startupScript: {
          scriptArguments: 'string'
          scriptData: 'string'
          scriptSource: 'string'
          timeout: 'string'
        }
      }
    }
    sshSettings: {
      adminPublicKey: 'string'
      sshPublicAccess: 'string'
    }
    subnet: {
      id: 'string'
    }
    vmSize: 'string'
  }

針對 Databricks,請使用:

  computeType: 'Databricks'
  properties: {
    databricksAccessToken: 'string'
    workspaceUrl: 'string'
  }

針對 DataFactory,請使用:

  computeType: 'DataFactory'

針對 DataLakeAnalytics,請使用:

  computeType: 'DataLakeAnalytics'
  properties: {
    dataLakeStoreAccountName: 'string'
  }

針對 HDInsight,請使用:

  computeType: 'HDInsight'
  properties: {
    address: 'string'
    administratorAccount: {
      password: 'string'
      privateKeyData: 'string'
      publicKeyData: 'string'
      username: 'string'
    }
    sshPort: int
  }

針對 Kubernetes,請使用:

  computeType: 'Kubernetes'
  properties: {
    defaultInstanceType: 'string'
    extensionInstanceReleaseTrain: 'string'
    extensionPrincipalId: 'string'
    instanceTypes: {
      {customized property}: {
        nodeSelector: {
          {customized property}: 'string'
        }
        resources: {
          limits: {
            {customized property}: 'string'
            {customized property}: 'string'
          }
          requests: {
            {customized property}: 'string'
            {customized property}: 'string'
          }
        }
      }
    }
    namespace: 'string'
    relayConnectionString: 'string'
    serviceBusConnectionString: 'string'
    vcName: 'string'
  }

針對 SynapseSpark,請使用:

  computeType: 'SynapseSpark'
  properties: {
    autoPauseProperties: {
      delayInMinutes: int
      enabled: bool
    }
    autoScaleProperties: {
      enabled: bool
      maxNodeCount: int
      minNodeCount: int
    }
    nodeCount: int
    nodeSize: 'string'
    nodeSizeFamily: 'string'
    poolName: 'string'
    resourceGroup: 'string'
    sparkVersion: 'string'
    subscriptionId: 'string'
    workspaceName: 'string'
  }

針對 VirtualMachine,請使用:

  computeType: 'VirtualMachine'
  properties: {
    address: 'string'
    administratorAccount: {
      password: 'string'
      privateKeyData: 'string'
      publicKeyData: 'string'
      username: 'string'
    }
    isNotebookInstanceCompute: bool
    notebookServerPort: int
    sshPort: int
    virtualMachineSize: 'string'
  }

屬性值

workspaces/computes

名稱 描述
NAME 資源名稱

瞭解如何在 Bicep 中設定子資源的名稱和類型。
字串 (必要)

字元限制:計算實例 3-24
3-32 用於 AML 計算
2-16 用於其他計算類型

合法字元:
英數字元和連字號。
location 指定資源的位置。 字串
tags 包含定義為索引鍵/值組的資源標記。 標記名稱和值的字典。 請參閱 範本中的標記
sku 工作區的 SKU。 Sku
父系 (parent) 在 Bicep 中,您可以指定子資源的父資源。 只有在父資源之外宣告子資源時,才需要新增這個屬性。

如需詳細資訊,請參閱 父資源外部的子資源
類型資源的符號名稱: 工作區
身分識別 資源的身分識別。 ManagedServiceIdentity
properties 計算屬性 計算

ManagedServiceIdentity

名稱 描述
類型 (允許 SystemAssigned 和 UserAssigned 類型的受控服務識別類型) 。 'None'
'SystemAssigned'
'SystemAssigned,UserAssigned'
'UserAssigned' (必要)
userAssignedIdentities 與資源相關聯的使用者指派身分識別集。 userAssignedIdentities 字典索引鍵的格式為 ARM 資源標識符:'/subscriptions/{subscriptionId}/resourceGroups/{resourceGroupName}/providers/Microsoft.ManagedIdentity/userAssignedIdentities/{identityName}。 字典值可以是要求中 () 的空白物件 {} 。 UserAssignedIdentities

UserAssignedIdentities

名稱 描述
{自定義屬性} UserAssignedIdentity

UserAssignedIdentity

此物件不包含部署期間要設定的任何屬性。 所有屬性都是 ReadOnly。

計算

名稱 描述
computeLocation 基礎計算的位置 字串
description Machine Learning 計算的描述。 字串
disableLocalAuth 退出宣告本機驗證,並確保客戶只能使用 MSI 和 AAD 進行驗證。 bool
resourceId 基礎計算的 ARM 資源識別碼 字串
computeType 設定物件類型 AKS
AmlCompute
ComputeInstance
Databricks
DataFactory
DataLakeAnalytics
HDInsight
Kubernetes
SynapseSpark
VirtualMachine (必要)

AKS

名稱 描述
computeType 計算的類型 必要的 『AKS』 ()
properties AKS 屬性 AKSSchemaProperties

AKSSchemaProperties

名稱 描述
agentCount 代理程式數目 int

約束:
最小值 = 0
agentVmSize 代理程式虛擬機大小 字串
aksNetworkingConfiguration vnet 的 AKS 網路設定 AksNetworkingConfiguration
clusterFqdn 叢集完整功能變數名稱 字串
clusterPurpose 叢集的預定用法 'DenseProd'
'DevTest'
'FastProd'
loadBalancerSubnet Load Balancer 子網 字串
loadBalancerType Load Balancer類型 'InternalLoadBalancer'
'PublicIp'
sslConfiguration SSL 設定 SslConfiguration

AksNetworkingConfiguration

名稱 描述
dnsServiceIP 指派給 Kubernetes DNS 服務的 IP 位址。 它必須位於 serviceCidr 中指定的 Kubernetes 服務地址範圍內。 字串

約束:
模式 = ^(?:(?:25[0-5]|2[0-4][0-9]|[01]?[0-9][0-9]?)\.){3}(?:25[0-5]|2[0-4][0-9]|[01]?[0-9][0-9]?)$
dockerBridgeCidr 指派給 Docker 網橋網路的 CIDR 表示法 IP 範圍。 它不得與任何子網 IP 範圍或 Kubernetes 服務地址範圍重疊。 字串

約束:
模式 = ^([0-9]{1,3}\.){3}[0-9]{1,3}(\/([0-9]|[1-2][0-9]|3[0-2]))?$
serviceCidr 據以指派服務叢集 IP 的 CIDR 標記法 IP 範圍。 此範圍不得與任何子網路 IP 範圍重疊。 字串

約束:
模式 = ^([0-9]{1,3}\.){3}[0-9]{1,3}(\/([0-9]|[1-2][0-9]|3[0-2]))?$
subnetId 計算節點所屬的虛擬網路子網資源標識碼 字串

SslConfiguration

名稱 描述
cert 憑證數據 字串

約束:
敏感性值。 以安全參數的形式傳入。
cname 憑證的 CNAME 字串
索引鍵 索引鍵數據 字串

約束:
敏感性值。 以安全參數的形式傳入。
leafDomainLabel 公用端點的分葉網域標籤 字串
overwriteExistingDomain 指出是否要覆寫現有的網域卷標。 bool
status 啟用或停用ssl以進行評分 'Auto'
'Disabled'
'Enabled'

AmlCompute

名稱 描述
computeType 計算的類型 'AmlCompute' (必要)
properties AmlCompute 的屬性 AmlComputeProperties

AmlComputeProperties

名稱 描述
enableNodePublicIp 啟用或停用節點公用IP位址布建。 可能的值為:可能的值為:true - 表示計算節點會布建公用IP。 false - 表示計算節點會有私人端點,且沒有公用IP。 bool
isolatedNetwork 網路已隔離或未隔離 bool
osType 計算 OS 類型 'Linux'
'Windows'
propertyBag 包含其他屬性的屬性包。 針對 Bicep,您可以使用 any () 函式。
remoteLoginPortPublicAccess 公用 SSH 埠的狀態。 可能的值為:已停用 - 指出叢集的所有節點上已關閉公用 ssh 埠。 已啟用 - 指出叢集的所有節點上都已開啟公用 SSH 埠。 NotSpecified - 指出如果已定義 VNet,則叢集的所有節點上都會關閉公用 SSH 埠,否則會開啟所有公用節點。 只有在建立叢集之後,才會啟用或停用叢集建立期間的預設。 'Disabled'
'Enabled'
'NotSpecified'
scaleSettings AML 計算的調整設定 ScaleSettings
子網路 計算節點所屬的虛擬網路子網資源標識碼。 ResourceId
userAccountCredentials 將在每個計算節點上建立之系統管理員用戶帳戶的認證。 UserAccountCredentials
virtualMachineImage AML 計算的虛擬機映射 - 僅限視窗 VirtualMachineImage
vmPriority 虛擬機優先順序 'Dedicated'
'LowPriority'
vmSize 虛擬機器大小 字串

ScaleSettings

名稱 描述
maxNodeCount 要使用的節點數目上限 int (必要)
minNodeCount 要使用的節點數目下限 int
nodeIdleTimeBeforeScaleDown 相應減少 amlCompute 之前的節點閑置時間。 此字串必須是 RFC 格式。 字串

ResourceId

名稱 描述
id 資源的標識碼 字串 (必要)

UserAccountCredentials

名稱 描述
adminUserName 可用來透過 SSH 連線至節點的系統管理員用戶帳戶名稱。 字串 (必要)
adminUserPassword 系統管理員用戶帳戶的密碼。 字串

約束:
敏感性值。 以安全參數的形式傳入。
adminUserSshPublicKey 系統管理員用戶帳戶的 SSH 公鑰。 字串

約束:
敏感性值。 以安全參數的形式傳入。

VirtualMachineImage

名稱 描述
id 虛擬機映像路徑 字串 (必要)

ComputeInstance

名稱 描述
computeType 計算的類型 必要) ('ComputeInstance'
properties ComputeInstance 的屬性 ComputeInstanceProperties

ComputeInstanceProperties

名稱 描述
applicationSharingPolicy 在父工作區的用戶之間共用此計算實例上應用程式的原則。 如果為 Personal,則只有建立者可以存取此計算實例上的應用程式。 共用時,任何工作區使用者都可以根據指派的角色存取此實例上的應用程式。 'Personal'
'Shared'
autologgerSettings 指定自動記錄工具的設定。 ComputeInstanceAutologgerSettings
computeInstanceAuthorizationType 計算實例授權類型。 可用的值是個人 (預設) 。 'personal'
customServices 新增至計算的自定義服務清單。 CustomService[]
enableNodePublicIp 啟用或停用節點公用IP位址布建。 可能的值為:可能的值為:true - 表示計算節點會布建公用IP。 false - 表示計算節點會有私人端點,且沒有公用IP。 bool
idleTimeBeforeShutdown 在使用者定義閑置期間之後停止計算實例。 時間是以ISO8601格式定義。 最小值為 15 分鐘,最大值為 3 天。 字串
personalComputeInstanceSettings 個人計算實例的設定。 PersonalComputeInstanceSettings
schedules 要套用至計算的排程清單。 ComputeSchedules
setupScripts 要執行以設定叢集的自定義腳本詳細數據。 SetupScripts
sshSettings 指定 SSH 存取的原則和設定。 ComputeInstanceSshSettings
子網路 計算節點所屬的虛擬網路子網資源標識碼。 ResourceId
vmSize 虛擬機器大小 字串

ComputeInstanceAutologgerSettings

名稱 描述
mlflowAutologger 指出是否為筆記本啟用mlflow自動記錄器。 'Disabled'
'Enabled'

CustomService

名稱 描述
docker 描述映像的 Docker 設定 Docker
端點 設定容器的端點 Endpoint[]
environmentVariables 容器的環境變數 CustomServiceEnvironmentVariables
image 描述影像規格 映像
NAME 自訂服務的名稱 字串
磁碟區 設定容器的磁碟區 VolumeDefinition[]
{自定義屬性} 針對 Bicep,您可以使用 any () 函式。

Docker

名稱 描述
特權 指出容器是否應以特殊許可權或非特殊許可權模式執行。 bool
{自定義屬性} 針對 Bicep,您可以使用 any () 函式。

端點

名稱 描述
hostIp 應用程式從容器公開的主機IP 字串
NAME 端點的名稱 字串
protocol 透過此端點進行通訊的通訊協定 'http'
'tcp'
'udp'
published 應用程式從容器公開的埠。 int
目標 容器內的應用程式埠。 int

CustomServiceEnvironmentVariables

名稱 描述
{自定義屬性} EnvironmentVariable

EnvironmentVariable

名稱 描述
類型 環境變數的類型。 可能的值為:local - 針對局部變數 'local'
環境變數的值 字串
{自定義屬性} 針對 Bicep,您可以使用 any () 函式。

映像

名稱 描述
參考 影像參考 URL 字串
type 影像的類型。 可能的值為:docker - 針對 docker 映射。 azureml - 適用於 AzureML 映射 'azureml'
'docker'
{自定義屬性} 針對 Bicep,您可以使用 any () 函式。

VolumeDefinition

名稱 描述
bind 掛接的系結選項 BindOptions
一致性 磁碟區的一致性 字串
readOnly 指出是否要將磁碟區掛接為 readOnly。 這個的預設值為 false。 bool
來源 掛接的來源。 針對系結掛接,這是主機路徑。 字串
目標 掛接的目標。 針對系結掛接,這是容器中的路徑。 字串
Tmpfs 掛接的 tmpfs 選項 TmpfsOptions
類型 磁碟區定義的類型。 可能的值:bind、volume、tmpfs、npipe 'bind'
'npipe'
'tmpfs'
'volume'
磁碟區 掛接的磁碟區選項 VolumeOptions

BindOptions

名稱 描述
createHostPath 指出是否要建立主機路徑。 bool
傳播 系結選項的類型 字串
selinux 提及 selinux 選項。 字串

TmpfsOptions

名稱 描述
{1}size{2} 提及 Tmpfs 大小 int

VolumeOptions

名稱 描述
nocopy 指出磁碟區是否為 nocopy bool

PersonalComputeInstanceSettings

名稱 描述
assignedUser 明確指派給個人計算實例的使用者。 AssignedUser

AssignedUser

名稱 描述
objectId 使用者的 AAD 物件識別碼。 需要字串 ()
tenantId 使用者的 AAD 租使用者識別碼。 需要字串 ()

ComputeSchedules

名稱 描述
computeStartStop 要套用的計算開始停止排程清單。 ComputeStartStopSchedule[]

ComputeStartStopSchedule

名稱 描述
動作 [必要]計算能力動作。 'Start'
'Stop'
Cron 如果 triggerType 為 Cron,則為必要項。 Cron
週期 如果 triggerType 為 Recurrence,則為必要項。 週期性
schedule [已被取代]未再使用。 ScheduleBase
status 排程是否已啟用或停用? 'Disabled'
'Enabled'
triggerType [必要]排程觸發程序類型。 'Cron'
'Recurrence'

Cron

名稱 描述
expression [必要]指定排程的cron運算式。
表達式應遵循 NCronTab 格式。
字串
startTime yyyy-MM-ddTHH:mm:ss 格式的開始時間。 字串
timeZone 指定排程執行所在的時區。
TimeZone應遵循 Windows 時區格式。 請參閱: https://docs.microsoft.com/windows-hardware/manufacture/desktop/default-time-zones />
字串

定期

名稱 描述
frequency [必要]觸發排程的頻率。 'Day'
'Hour'
'Minute'
'Month'
'Week'
interval [必要]指定與頻率搭配的排程間隔 int
schedule [必要]週期排程。 RecurrenceSchedule
startTime yyyy-MM-ddTHH:mm:ss 格式的開始時間。 字串
timeZone 指定排程執行所在的時區。
TimeZone應遵循 Windows 時區格式。 請參閱: https://docs.microsoft.com/windows-hardware/manufacture/desktop/default-time-zones />
字串

RecurrenceSchedule

名稱 描述
小時 [必要]排程的時數清單。 int[] (必要)
分鐘 [必要]排程的分鐘清單。 int[] (必要)
monthDays 排程的月份天數清單 int[]
weekDays 排程的天數清單。 包含任何項目的字串數組:
'Friday'
'Monday'
'Saturday'
'Sunday'
'Thursday'
'Tuesday'
'Wednesday'

ScheduleBase

名稱 描述
id 排程的系統指派標識碼。 字串
provisioningStatus 排程的目前部署狀態。 'Completed'
'Failed'
'Provisioning'
status 排程是否已啟用或停用? 'Disabled'
'Enabled'

SetupScripts

名稱 描述
指令碼 自訂的安裝腳本 ScriptsToExecute

ScriptsToExecute

名稱 描述
creationScript 在布建計算期間只執行一次的腳本。 ScriptReference
startupScript 每次機器啟動時執行的腳本。 ScriptReference

ScriptReference

名稱 描述
scriptArguments 傳遞至要執行的腳本的選擇性命令行自變數。 字串
scriptData 掛接磁碟區中的腳本位置。 字串
scriptSource 腳本的儲存來源:內嵌、工作區。 字串
timeout 傳遞至逾時命令的選擇性時間週期。 字串

ComputeInstanceSshSettings

名稱 描述
adminPublicKey 將 SSH rsa 公鑰檔案指定為字串。 使用 「ssh-keygen -t rsa -b 2048」 來產生 SSH 金鑰組。 字串
sshPublicAccess 公用 SSH 埠的狀態。 可能的值為:已停用 - 指出此實例上的公用 SSH 埠已關閉。 已啟用 - 指出公用 ssh 埠已開啟,並視情況根據 VNet/子網原則進行存取。 'Disabled'
'Enabled'

Databricks

名稱 描述
computeType 計算的類型 需要 『Databricks』 ()
properties Databricks 的屬性 DatabricksProperties

DatabricksProperties

名稱 描述
databricksAccessToken Databricks 存取令牌 字串
workspaceUrl 工作區 URL 字串

DataFactory

名稱 描述
computeType 計算的類型 必要) 'DataFactory' (

DataLakeAnalytics

名稱 描述
computeType 計算的類型 必要 ('DataLakeAnalytics')
properties DataLakeAnalyticsSchemaProperties

DataLakeAnalyticsSchemaProperties

名稱 描述
dataLakeStoreAccountName DataLake Store 帳戶名稱 字串

HDInsight

名稱 描述
computeType 計算的類型 'HDInsight' (必要)
properties HDInsight 計算屬性 HDInsightProperties

HDInsightProperties

名稱 描述
address 叢集主要節點的公用IP位址。 字串
administratorAccount 管理員 叢集主要節點的認證 VirtualMachineSshCredentials
sshPort 在叢集的主要節點上開啟 SSH 連線的埠。 int

VirtualMachineSshCredentials

名稱 描述
password 系統管理員帳戶的密碼 字串
privateKeyData 私鑰資料 字串

約束:
敏感性值。 以安全參數的形式傳入。
publicKeyData 公鑰資料 字串

約束:
敏感性值。 以安全參數的形式傳入。
使用者名稱 系統管理員帳戶的用戶名稱 字串

Kubernetes

名稱 描述
computeType 計算的類型 必要) 'Kubernetes' (
properties Kubernetes 的屬性 KubernetesProperties

KubernetesProperties

名稱 描述
defaultInstanceType 預設執行個體類型 字串
extensionInstanceReleaseTrain 擴充實例發行定型。 字串
extensionPrincipalId 延伸模組principal-id。 字串
instanceTypes 實例類型架構 KubernetesPropertiesInstanceTypes
命名空間 計算命名空間 字串
relayConnectionString 轉 連接字串。 字串

約束:
敏感性值。 以安全參數的形式傳入。
serviceBusConnectionString ServiceBus 連接字串。 字串

約束:
敏感性值。 以安全參數的形式傳入。
vcName VC 名稱。 字串

KubernetesPropertiesInstanceTypes

名稱 描述
{自定義屬性} InstanceTypeSchema

InstanceTypeSchema

名稱 描述
nodeSelector 節點選取器 InstanceTypeSchemaNodeSelector
resources 此實例類型的資源要求/限制 InstanceTypeSchemaResources

InstanceTypeSchemaNodeSelector

名稱 描述
{自定義屬性} 字串

InstanceTypeSchemaResources

名稱 描述
限制 此實例類型的資源限制 InstanceResourceSchema
requests 此實例類型的資源要求 InstanceResourceSchema

InstanceResourceSchema

名稱 描述
{自定義屬性} 字串
{自定義屬性} 字串

SynapseSpark

名稱 描述
computeType 計算的類型 'SynapseSpark' (必要)
properties SynapseSparkProperties

SynapseSparkProperties

名稱 描述
autoPauseProperties 自動暫停屬性。 AutoPauseProperties
autoScaleProperties 自動調整屬性。 AutoScaleProperties
nodeCount 目前指派給計算的計算節點數目。 int
nodeSize 節點大小。 字串
nodeSizeFamily 節點大小系列。 字串
poolName 集區名稱。 字串
resourceGroup 工作區所在的資源群組名稱。 字串
sparkVersion Spark 版本。 字串
subscriptionId Azure 訂用帳戶識別碼。 字串
workspaceName Azure Machine Learning 工作區的名稱。 字串

AutoPauseProperties

名稱 描述
delayInMinutes int
已啟用 bool

AutoScaleProperties

名稱 描述
已啟用 bool
maxNodeCount int
minNodeCount int

VirtualMachine

名稱 描述
computeType 計算的類型 'VirtualMachine' (必要)
properties VirtualMachineSchemaProperties

VirtualMachineSchemaProperties

名稱 描述
address 虛擬機的公用IP位址。 字串
administratorAccount 虛擬機的 管理員 認證 VirtualMachineSshCredentials
isNotebookInstanceCompute 指出此計算是否將用於執行筆記本。 bool
notebookServerPort 針對SSH連線開啟的筆記本伺服器埠。 int
sshPort 針對SSH連線開啟的埠。 int
virtualMachineSize 虛擬機器大小 字串

SKU

名稱 描述
處理能力 如果 SKU 支援相應放大/縮小,則應該包含容量整數。 如果資源無法相應放大/縮小,可能會省略此專案。 int
family 如果服務有不同的硬體世代,則針對相同的 SKU,則可以在這裡擷取。 字串
NAME SKU 的名稱。 例如 - P3。 通常是字母+數位碼 需要字串 ()
{1}size{2} SKU 大小。 當名稱欄位是階層和其他一些值的組合時,這會是獨立程序代碼。 字串
tier 如果服務有一個以上的層級,但在PUT上則不需要此欄位,才能由資源提供者實作。 'Basic'
'Free'
'Premium'
'Standard'

快速入門範本

下列快速入門範本會部署此資源類型。

範本 描述
建立機器學習服務 ADLA 計算

部署至 Azure
此範本會建立 Machine Learning 服務 ADLA Compute。
建立機器學習服務 Aks 計算

部署至 Azure
此範本會建立機器學習服務 Aks 計算。
建立機器學習服務 DSVM 計算

部署至 Azure
此範本會建立機器學習服務 DSVM 計算。
建立機器學習服務 HDInsight 叢集

部署至 Azure
此範本會建立機器學習服務 HDInsight 叢集
建立 Azure Machine Learning aks 計算

部署至 Azure
此範本會建立 Azure Machine Learning aks 計算。
建立 Azure Machine Learning 計算叢集

部署至 Azure
此範本會建立 Azure Machine Learning 計算叢集。
建立 Azure Machine Learning 計算實例

部署至 Azure
此範本會代表另一位具有範例內嵌安裝腳本的使用者建立 Azure Machine Learning 計算實例
在 Azure Machine Learning 工作區中建立 LinkedService

部署至 Azure
此範本會在現有的 Azure Machine Learning 工作區中建立 LinkedService。
使用私人IP位址建立 AKS 計算目標

部署至 Azure
此範本會在具有私人IP位址的指定 Azure Machine Learning 服務工作區中建立 AKS 計算目標。

ARM 範本資源定義

工作區/計算資源類型可以使用目標作業進行部署:

如需每個 API 版本中已變更屬性的清單,請參閱 變更記錄檔。

資源格式

若要建立 Microsoft.MachineLearningServices/workspaces/computes 資源,請將下列 JSON 新增至範本。

{
  "type": "Microsoft.MachineLearningServices/workspaces/computes",
  "apiVersion": "2023-04-01-preview",
  "name": "string",
  "location": "string",
  "tags": {
    "tagName1": "tagValue1",
    "tagName2": "tagValue2"
  },
    "{customized property}": "string"
  "sku": {
    "capacity": "int",
    "family": "string",
    "name": "string",
    "size": "string",
    "tier": "string"
  },
  "identity": {
    "type": "string",
    "userAssignedIdentities": {
      "{customized property}": {}
    }
  },
  "properties": {
    "computeLocation": "string",
    "description": "string",
    "disableLocalAuth": "bool",
    "resourceId": "string",
    "computeType": "string"
    // For remaining properties, see Compute objects
  }
}

計算物件

設定 computeType 屬性以指定物件的類型。

針對 AKS,請使用:

  "computeType": "AKS",
  "properties": {
    "agentCount": "int",
    "agentVmSize": "string",
    "aksNetworkingConfiguration": {
      "dnsServiceIP": "string",
      "dockerBridgeCidr": "string",
      "serviceCidr": "string",
      "subnetId": "string"
    },
    "clusterFqdn": "string",
    "clusterPurpose": "string",
    "loadBalancerSubnet": "string",
    "loadBalancerType": "string",
    "sslConfiguration": {
      "cert": "string",
      "cname": "string",
      "key": "string",
      "leafDomainLabel": "string",
      "overwriteExistingDomain": "bool",
      "status": "string"
    }
  }

針對 AmlCompute,請使用:

  "computeType": "AmlCompute",
  "properties": {
    "enableNodePublicIp": "bool",
    "isolatedNetwork": "bool",
    "osType": "string",
    "propertyBag": {},
    "remoteLoginPortPublicAccess": "string",
    "scaleSettings": {
      "maxNodeCount": "int",
      "minNodeCount": "int",
      "nodeIdleTimeBeforeScaleDown": "string"
    },
    "subnet": {
      "id": "string"
    },
    "userAccountCredentials": {
      "adminUserName": "string",
      "adminUserPassword": "string",
      "adminUserSshPublicKey": "string"
    },
    "virtualMachineImage": {
      "id": "string"
    },
    "vmPriority": "string",
    "vmSize": "string"
  }

針對 ComputeInstance,請使用:

  "computeType": "ComputeInstance",
  "properties": {
    "applicationSharingPolicy": "string",
    "autologgerSettings": {
      "mlflowAutologger": "string"
    },
    "computeInstanceAuthorizationType": "personal",
    "customServices": [
      {
        "docker": {
          "privileged": "bool",
          "{customized property}": {}
        },
        "endpoints": [
          {
            "hostIp": "string",
            "name": "string",
            "protocol": "string",
            "published": "int",
            "target": "int"
          }
        ],
        "environmentVariables": {
          "{customized property}": {
            "type": "local",
            "value": "string",
            "{customized property}": {}
          }
        },
        "image": {
          "reference": "string",
          "type": "string",
          "{customized property}": {}
        },
        "name": "string",
        "volumes": [
          {
            "bind": {
              "createHostPath": "bool",
              "propagation": "string",
              "selinux": "string"
            },
            "consistency": "string",
            "readOnly": "bool",
            "source": "string",
            "target": "string",
            "tmpfs": {
              "size": "int"
            },
            "type": "string",
            "volume": {
              "nocopy": "bool"
            }
          }
        ],
        "{customized property}": {}
      }
    ],
    "enableNodePublicIp": "bool",
    "idleTimeBeforeShutdown": "string",
    "personalComputeInstanceSettings": {
      "assignedUser": {
        "objectId": "string",
        "tenantId": "string"
      }
    },
    "schedules": {
      "computeStartStop": [
        {
          "action": "string",
          "cron": {
            "expression": "string",
            "startTime": "string",
            "timeZone": "string"
          },
          "recurrence": {
            "frequency": "string",
            "interval": "int",
            "schedule": {
              "hours": [ "int" ],
              "minutes": [ "int" ],
              "monthDays": [ "int" ],
              "weekDays": [ "string" ]
            },
            "startTime": "string",
            "timeZone": "string"
          },
          "schedule": {
            "id": "string",
            "provisioningStatus": "string",
            "status": "string"
          },
          "status": "string",
          "triggerType": "string"
        }
      ]
    },
    "setupScripts": {
      "scripts": {
        "creationScript": {
          "scriptArguments": "string",
          "scriptData": "string",
          "scriptSource": "string",
          "timeout": "string"
        },
        "startupScript": {
          "scriptArguments": "string",
          "scriptData": "string",
          "scriptSource": "string",
          "timeout": "string"
        }
      }
    },
    "sshSettings": {
      "adminPublicKey": "string",
      "sshPublicAccess": "string"
    },
    "subnet": {
      "id": "string"
    },
    "vmSize": "string"
  }

針對 Databricks,請使用:

  "computeType": "Databricks",
  "properties": {
    "databricksAccessToken": "string",
    "workspaceUrl": "string"
  }

針對 DataFactory,請使用:

  "computeType": "DataFactory"

針對 DataLakeAnalytics,請使用:

  "computeType": "DataLakeAnalytics",
  "properties": {
    "dataLakeStoreAccountName": "string"
  }

針對 HDInsight,請使用:

  "computeType": "HDInsight",
  "properties": {
    "address": "string",
    "administratorAccount": {
      "password": "string",
      "privateKeyData": "string",
      "publicKeyData": "string",
      "username": "string"
    },
    "sshPort": "int"
  }

針對 Kubernetes,請使用:

  "computeType": "Kubernetes",
  "properties": {
    "defaultInstanceType": "string",
    "extensionInstanceReleaseTrain": "string",
    "extensionPrincipalId": "string",
    "instanceTypes": {
      "{customized property}": {
        "nodeSelector": {
          "{customized property}": "string"
        },
        "resources": {
          "limits": {
            "{customized property}": "string",
            "{customized property}": "string"
          },
          "requests": {
            "{customized property}": "string",
            "{customized property}": "string"
          }
        }
      }
    },
    "namespace": "string",
    "relayConnectionString": "string",
    "serviceBusConnectionString": "string",
    "vcName": "string"
  }

針對 SynapseSpark,請使用:

  "computeType": "SynapseSpark",
  "properties": {
    "autoPauseProperties": {
      "delayInMinutes": "int",
      "enabled": "bool"
    },
    "autoScaleProperties": {
      "enabled": "bool",
      "maxNodeCount": "int",
      "minNodeCount": "int"
    },
    "nodeCount": "int",
    "nodeSize": "string",
    "nodeSizeFamily": "string",
    "poolName": "string",
    "resourceGroup": "string",
    "sparkVersion": "string",
    "subscriptionId": "string",
    "workspaceName": "string"
  }

針對 VirtualMachine,請使用:

  "computeType": "VirtualMachine",
  "properties": {
    "address": "string",
    "administratorAccount": {
      "password": "string",
      "privateKeyData": "string",
      "publicKeyData": "string",
      "username": "string"
    },
    "isNotebookInstanceCompute": "bool",
    "notebookServerPort": "int",
    "sshPort": "int",
    "virtualMachineSize": "string"
  }

屬性值

workspaces/computes

名稱 描述
類型 資源類型 'Microsoft.MachineLearningServices/workspaces/computes'
apiVersion 資源 API 版本 '2023-04-01-preview'
NAME 資源名稱

瞭解如何在 JSON ARM 範本中設定子資源的名稱和類型。
需要字串 ()

字元限制:計算實例 3-24
3-32 用於 AML 計算
2-16 用於其他計算類型

合法字元:
英數字元和連字號。
location 指定資源的位置。 字串
tags 包含定義為索引鍵/值組的資源標記。 標記名稱和值的字典。 請參閱 範本中的標記
sku 工作區的 SKU。 Sku
身分識別 資源的身分識別。 ManagedServiceIdentity
properties 計算屬性 計算

ManagedServiceIdentity

名稱 描述
類型 (允許 SystemAssigned 和 UserAssigned 類型的受控服務識別類型) 。 'None'
'SystemAssigned'
'SystemAssigned,UserAssigned'
'UserAssigned' (必要)
userAssignedIdentities 與資源相關聯的使用者指派身分識別集。 userAssignedIdentities 字典索引鍵的格式為 ARM 資源標識符:'/subscriptions/{subscriptionId}/resourceGroups/{resourceGroupName}/providers/Microsoft.ManagedIdentity/userAssignedIdentities/{identityName}。 字典值可以是要求中 () 的空白物件 {} 。 UserAssignedIdentities

UserAssignedIdentities

名稱 描述
{自定義屬性} UserAssignedIdentity

UserAssignedIdentity

此物件不包含部署期間要設定的任何屬性。 所有屬性都是 ReadOnly。

計算

名稱 描述
computeLocation 基礎計算的位置 字串
description Machine Learning 計算的描述。 字串
disableLocalAuth 退出宣告本機驗證,並確保客戶只能使用 MSI 和 AAD 進行驗證。 bool
resourceId 基礎計算的 ARM 資源識別碼 字串
computeType 設定物件類型 AKS
AmlCompute
ComputeInstance
Databricks
DataFactory
DataLakeAnalytics
HDInsight
Kubernetes
SynapseSpark
VirtualMachine (必要)

AKS

名稱 描述
computeType 計算的類型 必要的 『AKS』 ()
properties AKS 屬性 AKSSchemaProperties

AKSSchemaProperties

名稱 描述
agentCount 代理程式數目 int

約束:
最小值 = 0
agentVmSize 代理程式虛擬機大小 字串
aksNetworkingConfiguration vnet 的 AKS 網路設定 AksNetworkingConfiguration
clusterFqdn 叢集完整功能變數名稱 字串
clusterPurpose 叢集的預定用法 'DenseProd'
'DevTest'
'FastProd'
loadBalancerSubnet Load Balancer 子網 字串
loadBalancerType Load Balancer類型 'InternalLoadBalancer'
'PublicIp'
sslConfiguration SSL 設定 SslConfiguration

AksNetworkingConfiguration

名稱 描述
dnsServiceIP 指派給 Kubernetes DNS 服務的 IP 位址。 它必須位於 serviceCidr 中指定的 Kubernetes 服務地址範圍內。 字串

約束:
模式 = ^(?:(?:25[0-5]|2[0-4][0-9]|[01]?[0-9][0-9]?)\.){3}(?:25[0-5]|2[0-4][0-9]|[01]?[0-9][0-9]?)$
dockerBridgeCidr 指派給 Docker 網橋網路的 CIDR 表示法 IP 範圍。 它不得與任何子網 IP 範圍或 Kubernetes 服務地址範圍重疊。 字串

約束:
模式 = ^([0-9]{1,3}\.){3}[0-9]{1,3}(\/([0-9]|[1-2][0-9]|3[0-2]))?$
serviceCidr 據以指派服務叢集 IP 的 CIDR 標記法 IP 範圍。 此範圍不得與任何子網路 IP 範圍重疊。 字串

約束:
模式 = ^([0-9]{1,3}\.){3}[0-9]{1,3}(\/([0-9]|[1-2][0-9]|3[0-2]))?$
subnetId 計算節點所屬的虛擬網路子網資源標識碼 字串

SslConfiguration

名稱 描述
cert 憑證數據 字串

約束:
敏感性值。 以安全參數的形式傳入。
cname 憑證的 CNAME 字串
索引鍵 索引鍵數據 字串

約束:
敏感性值。 以安全參數的形式傳入。
leafDomainLabel 公用端點的分葉網域標籤 字串
overwriteExistingDomain 指出是否要覆寫現有的網域卷標。 bool
status 啟用或停用 SSL 以進行評分 'Auto'
'Disabled'
'Enabled'

AmlCompute

名稱 描述
computeType 計算的類型 'AmlCompute' (必要)
properties AmlCompute 的屬性 AmlComputeProperties

AmlComputeProperties

名稱 描述
enableNodePublicIp 啟用或停用節點公用IP位址布建。 可能的值為:可能的值為:true - 表示計算節點會布建公用IP。 false - 指出計算節點會有私人端點,且沒有公用 IP。 bool
isolatedNetwork 網路已隔離或未隔離 bool
osType 計算 OS 類型 'Linux'
'Windows'
propertyBag 包含其他屬性的屬性包。
remoteLoginPortPublicAccess 公用 SSH 埠的狀態。 可能的值為:Disabled - 指出叢集的所有節點上已關閉公用 SSH 埠。 已啟用 - 表示公用 SSH 埠已在叢集的所有節點上開啟。 NotSpecified - 指出已定義 VNet 時,叢集的所有節點上都會關閉公用 SSH 埠,否則會開啟所有公用節點。 它只能在叢集建立期間預設,建立之後才會啟用或停用。 'Disabled'
'Enabled'
'NotSpecified'
scaleSettings AML 計算的調整設定 ScaleSettings
子網路 計算節點所屬的虛擬網路子網資源標識碼。 ResourceId
userAccountCredentials 將會在每個計算節點上建立之系統管理員用戶帳戶的認證。 UserAccountCredentials
virtualMachineImage AML 計算的虛擬機映射 - 僅限 windows VirtualMachineImage
vmPriority 虛擬機優先順序 'Dedicated'
'LowPriority'
vmSize 虛擬機器大小 字串

ScaleSettings

名稱 描述
maxNodeCount 要使用的節點數目上限 需要 int ()
minNodeCount 要使用的節點數目下限 int
nodeIdleTimeBeforeScaleDown 相應減少 amlCompute 之前的節點空閒時間。 此字串必須是 RFC 格式。 字串

ResourceId

名稱 描述
id 資源的標識碼 需要字串 ()

UserAccountCredentials

名稱 描述
adminUserName 可用來透過 SSH 連線至節點的系統管理員用戶帳戶名稱。 需要字串 ()
adminUserPassword 系統管理員用戶帳戶的密碼。 字串

約束:
敏感性值。 以安全參數的形式傳入。
adminUserSshPublicKey 系統管理員用戶帳戶的 SSH 公鑰。 字串

約束:
敏感性值。 以安全參數的形式傳入。

VirtualMachineImage

名稱 描述
id 虛擬機映像路徑 需要字串 ()

ComputeInstance

名稱 描述
computeType 計算的類型 'ComputeInstance' (必要)
properties ComputeInstance 的屬性 ComputeInstanceProperties

ComputeInstanceProperties

名稱 描述
applicationSharingPolicy 在父工作區用戶之間共用此計算實例上應用程式的原則。 如果為 Personal,則只有建立者可以存取此計算實例上的應用程式。 共用時,任何工作區使用者都可以根據其指派的角色存取此實例上的應用程式。 'Personal'
'Shared'
autologgerSettings 指定自動記錄工具的設定。 ComputeInstanceAutologgerSettings
computeInstanceAuthorizationType 計算實例授權類型。 可用的值為個人 (預設) 。 'personal'
customServices 新增至計算的自定義服務清單。 CustomService[]
enableNodePublicIp 啟用或停用節點公用IP位址布建。 可能的值為:可能的值為:true - 表示計算節點會布建公用IP。 false - 指出計算節點會有私人端點,且沒有公用 IP。 bool
idleTimeBeforeShutdown 在使用者定義閑置期間之後停止計算實例。 時間是以ISO8601格式定義。 最小值為 15 分鐘,最大值為 3 天。 字串
personalComputeInstanceSettings 個人計算實例的設定。 PersonalComputeInstanceSettings
schedules 要套用至計算的排程清單。 ComputeSchedules
setupScripts 要執行以設定叢集的自定義腳本詳細數據。 SetupScripts
sshSettings 指定 SSH 存取的原則和設定。 ComputeInstanceSshSettings
子網路 計算節點所屬的虛擬網路子網資源標識碼。 ResourceId
vmSize 虛擬機器大小 字串

ComputeInstanceAutologgerSettings

名稱 描述
mlflowAutologger 指出是否為筆記本啟用mlflow自動記錄器。 'Disabled'
'Enabled'

CustomService

名稱 描述
docker 描述映像的 Docker 設定 Docker
端點 設定容器的端點 Endpoint[]
environmentVariables 容器的環境變數 CustomServiceEnvironmentVariables
image 描述影像規格 映像
NAME 自訂服務的名稱 字串
磁碟區 設定容器的磁碟區 VolumeDefinition[]
{自定義屬性}

Docker

名稱 描述
特權 指出容器是否應以特殊許可權或非特殊許可權模式執行。 bool
{自定義屬性}

端點

名稱 描述
hostIp 應用程式從容器公開的主機IP 字串
NAME 端點的名稱 字串
protocol 透過此端點進行通訊的通訊協定 'http'
'tcp'
'udp'
published 應用程式從容器公開的埠。 int
目標 容器內的應用程式埠。 int

CustomServiceEnvironmentVariables

名稱 描述
{自定義屬性} EnvironmentVariable

EnvironmentVariable

名稱 描述
類型 環境變數的類型。 可能的值為:local - 針對局部變數 'local'
環境變數的值 字串
{自定義屬性}

映像

名稱 描述
參考 影像參考 URL 字串
type 影像的類型。 可能的值為:docker - 針對 docker 映射。 azureml - 適用於 AzureML 映射 'azureml'
'docker'
{自定義屬性}

VolumeDefinition

名稱 描述
bind 掛接的系結選項 BindOptions
一致性 磁碟區的一致性 字串
readOnly 指出是否要將磁碟區掛接為 readOnly。 默認值為 false。 bool
來源 掛接的來源。 對於系結掛接,這是主機路徑。 字串
目標 掛接的目標。 對於系結掛接,這是容器中的路徑。 字串
Tmpfs 掛接的 tmpfs 選項 TmpfsOptions
類型 磁碟區定義的類型。 可能的值:bind、volume、tmpfs、npipe 'bind'
'npipe'
'tmpfs'
'volume'
磁碟區 掛接的磁碟區選項 VolumeOptions

BindOptions

名稱 描述
createHostPath 指出是否要建立主機路徑。 bool
傳播 系結選項的類型 字串
selinux 提及 selinux 選項。 字串

TmpfsOptions

名稱 描述
{1}size{2} 提及 Tmpfs 大小 int

VolumeOptions

名稱 描述
nocopy 指出磁碟區是否為 nocopy bool

PersonalComputeInstanceSettings

名稱 描述
assignedUser 明確指派給個人計算實例的使用者。 AssignedUser

AssignedUser

名稱 描述
objectId 使用者的 AAD 物件識別碼。 字串 (必要)
tenantId 使用者的 AAD 租使用者識別碼。 字串 (必要)

ComputeSchedules

名稱 描述
computeStartStop 要套用的計算開始停止排程清單。 ComputeStartStopSchedule[]

ComputeStartStopSchedule

名稱 描述
動作 [必要]計算能力動作。 'Start'
'Stop'
Cron 如果 triggerType 為 Cron,則為必要專案。 Cron
週期 如果 triggerType 是週期,則為必要專案。 週期性
schedule [已淘汰]不再使用。 ScheduleBase
status 排程是否已啟用或停用? 'Disabled'
'Enabled'
triggerType [必要]排程觸發程序類型。 'Cron'
'Recurrence'

Cron

名稱 描述
expression [必要]指定排程的cron運算式。
表達式應該遵循 NCronTab 格式。
字串
startTime yyyy-MM-ddTHH:mm:ss 格式的開始時間。 字串
timeZone 指定排程執行所在的時區。
TimeZone應遵循 Windows 時區格式。 請參閱: https://docs.microsoft.com/windows-hardware/manufacture/desktop/default-time-zones />
字串

定期

名稱 描述
frequency [必要]觸發排程的頻率。 'Day'
'Hour'
'Minute'
'Month'
'Week'
interval [必要]指定排程間隔與頻率 int
schedule [必要]週期排程。 RecurrenceSchedule
startTime yyyy-MM-ddTHH:mm:ss 格式的開始時間。 字串
timeZone 指定排程執行所在的時區。
TimeZone應遵循 Windows 時區格式。 請參閱: https://docs.microsoft.com/windows-hardware/manufacture/desktop/default-time-zones />
字串

RecurrenceSchedule

名稱 描述
小時 [必要]排程的時數清單。 int[] (必要)
分鐘 [必要]排程的分鐘數清單。 int[] (必要)
monthDays 排程的月份天數清單 int[]
weekDays 排程的天數清單。 包含任何項目的字串數組:
'Friday'
'Monday'
'Saturday'
'Sunday'
'Thursday'
'Tuesday'
'Wednesday'

ScheduleBase

名稱 描述
id 排程的系統指派標識碼。 字串
provisioningStatus 排程的目前部署狀態。 'Completed'
'Failed'
'Provisioning'
status 排程是否已啟用或停用? 'Disabled'
'Enabled'

SetupScripts

名稱 描述
指令碼 自訂的安裝腳本 ScriptsToExecute

ScriptsToExecute

名稱 描述
creationScript 在布建計算期間只執行一次的腳本。 ScriptReference
startupScript 每次機器啟動時執行的腳本。 ScriptReference

ScriptReference

名稱 描述
scriptArguments 傳遞至要執行的腳本的選擇性命令行自變數。 字串
scriptData 掛接磁碟區中腳本的位置。 字串
scriptSource 腳本的儲存來源:內嵌工作區。 字串
timeout 傳遞至 timeout 命令的選擇性時間週期。 字串

ComputeInstanceSshSettings

名稱 描述
adminPublicKey 將 SSH rsa 公鑰檔案指定為字串。 使用 「ssh-keygen -t rsa -b 2048」 來產生 SSH 金鑰組。 字串
sshPublicAccess 公用 SSH 埠的狀態。 可能的值為:Disabled - 指出此實例上的公用 ssh 埠已關閉。 已啟用 - 指出公用 SSH 埠已開啟,並視情況根據 VNet/子網原則存取。 'Disabled'
'Enabled'

Databricks

名稱 描述
computeType 計算的類型 需要 『Databricks』 ()
properties Databricks 的屬性 DatabricksProperties

DatabricksProperties

名稱 描述
databricksAccessToken Databricks 存取令牌 字串
workspaceUrl 工作區 URL 字串

DataFactory

名稱 描述
computeType 計算的類型 需要 『DataFactory』 ()

DataLakeAnalytics

名稱 描述
computeType 計算的類型 需要 'DataLakeAnalytics' ()
properties DataLakeAnalyticsSchemaProperties

DataLakeAnalyticsSchemaProperties

名稱 描述
dataLakeStoreAccountName DataLake Store 帳戶名稱 字串

HDInsight

名稱 描述
computeType 計算的類型 需要 『HDInsight』 ()
properties HDInsight 計算屬性 HDInsightProperties

HDInsightProperties

名稱 描述
address 叢集主要節點的公用IP位址。 字串
administratorAccount 叢集主要節點的 管理員 認證 VirtualMachineSshCredentials
sshPort 針對叢集主要節點上的SSH連線開啟埠。 int

VirtualMachineSshCredentials

名稱 描述
password 系統管理員帳戶的密碼 字串
privateKeyData 私鑰資料 字串

約束:
敏感性值。 以安全參數的形式傳入。
publicKeyData 公鑰資料 字串

約束:
敏感性值。 以安全參數的形式傳入。
使用者名稱 系統管理員帳戶的用戶名稱 字串

Kubernetes

名稱 描述
computeType 計算的類型 必要) 'Kubernetes' (
properties Kubernetes 的屬性 KubernetesProperties

KubernetesProperties

名稱 描述
defaultInstanceType 預設執行個體類型 字串
extensionInstanceReleaseTrain 擴充實例發行定型。 字串
extensionPrincipalId 延伸模組principal-id。 字串
instanceTypes 實例類型架構 KubernetesPropertiesInstanceTypes
命名空間 計算命名空間 字串
relayConnectionString 轉 連接字串。 字串

約束:
敏感性值。 以安全參數的形式傳入。
serviceBusConnectionString ServiceBus 連接字串。 字串

約束:
敏感性值。 以安全參數的形式傳入。
vcName VC 名稱。 字串

KubernetesPropertiesInstanceTypes

名稱 描述
{自定義屬性} InstanceTypeSchema

InstanceTypeSchema

名稱 描述
nodeSelector 節點選取器 InstanceTypeSchemaNodeSelector
resources 此實例類型的資源要求/限制 InstanceTypeSchemaResources

InstanceTypeSchemaNodeSelector

名稱 描述
{自定義屬性} 字串

InstanceTypeSchemaResources

名稱 描述
限制 此實例類型的資源限制 InstanceResourceSchema
requests 此實例類型的資源要求 InstanceResourceSchema

InstanceResourceSchema

名稱 描述
{自定義屬性} 字串
{自定義屬性} 字串

SynapseSpark

名稱 描述
computeType 計算的類型 需要 『SynapseSpark』 ()
properties SynapseSparkProperties

SynapseSparkProperties

名稱 描述
autoPauseProperties 自動暫停屬性。 AutoPauseProperties
autoScaleProperties 自動調整屬性。 AutoScaleProperties
nodeCount 目前指派給計算的計算節點數目。 int
nodeSize 節點大小。 字串
nodeSizeFamily 節點大小系列。 字串
poolName 集區名稱。 字串
resourceGroup 工作區所在的資源組名。 字串
sparkVersion Spark 版本。 字串
subscriptionId Azure 訂用帳戶識別碼。 字串
workspaceName Azure Machine Learning 工作區的名稱。 字串

AutoPauseProperties

名稱 描述
delayInMinutes int
已啟用 bool

AutoScaleProperties

名稱 描述
已啟用 bool
maxNodeCount int
minNodeCount int

VirtualMachine

名稱 描述
computeType 計算的類型 'VirtualMachine' (必要)
properties VirtualMachineSchemaProperties

VirtualMachineSchemaProperties

名稱 描述
address 虛擬機的公用IP位址。 字串
administratorAccount 虛擬機 管理員 認證 VirtualMachineSshCredentials
isNotebookInstanceCompute 指出此計算是否將用於執行筆記本。 bool
notebookServerPort 針對 ssh 連線開啟的筆記本伺服器埠。 int
sshPort 針對 ssh 連線開啟埠。 int
virtualMachineSize 虛擬機器大小 字串

SKU

名稱 描述
處理能力 如果 SKU 支援相應放大/縮小,則應該包含容量整數。 如果資源無法相應放大/縮小,可能會省略此專案。 int
family 如果服務有不同的硬體世代,則針對相同的 SKU,則可以在這裡擷取。 字串
NAME SKU 的名稱。 例如 - P3。 通常是字母+數位碼 需要字串 ()
{1}size{2} SKU 大小。 當名稱欄位是階層和其他一些值的組合時,這會是獨立程序代碼。 字串
tier 如果服務有一個以上的層級,但在PUT上則不需要此欄位,才能由資源提供者實作。 'Basic'
'Free'
'Premium'
'Standard'

快速入門範本

下列快速入門範本會部署此資源類型。

範本 描述
建立機器學習服務 ADLA 計算

部署至 Azure
此範本會建立 Machine Learning 服務 ADLA Compute。
建立機器學習服務 Aks 計算

部署至 Azure
此範本會建立機器學習服務 Aks 計算。
建立機器學習服務 DSVM 計算

部署至 Azure
此範本會建立機器學習服務 DSVM 計算。
建立機器學習服務 HDInsight 叢集

部署至 Azure
此範本會建立機器學習服務 HDInsight 叢集
建立 Azure Machine Learning aks 計算

部署至 Azure
此範本會建立 Azure Machine Learning aks 計算。
建立 Azure Machine Learning 計算叢集

部署至 Azure
此範本會建立 Azure Machine Learning 計算叢集。
建立 Azure Machine Learning 計算實例

部署至 Azure
此範本會代表另一位具有範例內嵌安裝腳本的使用者建立 Azure Machine Learning 計算實例
在 Azure Machine Learning 工作區中建立 LinkedService

部署至 Azure
此範本會在現有的 Azure Machine Learning 工作區中建立 LinkedService。
使用私人IP位址建立 AKS 計算目標

部署至 Azure
此範本會在具有私人IP位址的指定 Azure Machine Learning 服務工作區中建立 AKS 計算目標。

Terraform (AzAPI 提供者) 資源定義

工作區/計算資源類型可以使用目標作業進行部署:

  • 資源群組

如需每個 API 版本中已變更屬性的清單,請參閱 變更記錄檔。

資源格式

若要建立 Microsoft.MachineLearningServices/workspaces/computes 資源,請將下列 Terraform 新增至範本。

resource "azapi_resource" "symbolicname" {
  type = "Microsoft.MachineLearningServices/workspaces/computes@2023-04-01-preview"
  name = "string"
  location = "string"
  parent_id = "string"
  tags = {
    tagName1 = "tagValue1"
    tagName2 = "tagValue2"
  }
    {customized property} = "string"
  identity {
    type = "string"
    identity_ids = []
  }
  body = jsonencode({
    properties = {
      computeLocation = "string"
      description = "string"
      disableLocalAuth = bool
      resourceId = "string"
      computeType = "string"
      // For remaining properties, see Compute objects
    }
    sku = {
      capacity = int
      family = "string"
      name = "string"
      size = "string"
      tier = "string"
    }
  })
}

計算物件

設定 computeType 屬性以指定物件的類型。

針對 AKS,請使用:

  computeType = "AKS"
  properties = {
    agentCount = int
    agentVmSize = "string"
    aksNetworkingConfiguration = {
      dnsServiceIP = "string"
      dockerBridgeCidr = "string"
      serviceCidr = "string"
      subnetId = "string"
    }
    clusterFqdn = "string"
    clusterPurpose = "string"
    loadBalancerSubnet = "string"
    loadBalancerType = "string"
    sslConfiguration = {
      cert = "string"
      cname = "string"
      key = "string"
      leafDomainLabel = "string"
      overwriteExistingDomain = bool
      status = "string"
    }
  }

針對 AmlCompute,請使用:

  computeType = "AmlCompute"
  properties = {
    enableNodePublicIp = bool
    isolatedNetwork = bool
    osType = "string"
    remoteLoginPortPublicAccess = "string"
    scaleSettings = {
      maxNodeCount = int
      minNodeCount = int
      nodeIdleTimeBeforeScaleDown = "string"
    }
    subnet = {
      id = "string"
    }
    userAccountCredentials = {
      adminUserName = "string"
      adminUserPassword = "string"
      adminUserSshPublicKey = "string"
    }
    virtualMachineImage = {
      id = "string"
    }
    vmPriority = "string"
    vmSize = "string"
  }

針對 ComputeInstance,請使用:

  computeType = "ComputeInstance"
  properties = {
    applicationSharingPolicy = "string"
    autologgerSettings = {
      mlflowAutologger = "string"
    }
    computeInstanceAuthorizationType = "personal"
    customServices = [
      {
        docker = {
          privileged = bool
        }
        endpoints = [
          {
            hostIp = "string"
            name = "string"
            protocol = "string"
            published = int
            target = int
          }
        ]
        environmentVariables = {
          {customized property} = {
            type = "local"
            value = "string"
          }
        }
        image = {
          reference = "string"
          type = "string"
        }
        name = "string"
        volumes = [
          {
            bind = {
              createHostPath = bool
              propagation = "string"
              selinux = "string"
            }
            consistency = "string"
            readOnly = bool
            source = "string"
            target = "string"
            tmpfs = {
              size = int
            }
            type = "string"
            volume = {
              nocopy = bool
            }
          }
        ]
      }
    ]
    enableNodePublicIp = bool
    idleTimeBeforeShutdown = "string"
    personalComputeInstanceSettings = {
      assignedUser = {
        objectId = "string"
        tenantId = "string"
      }
    }
    schedules = {
      computeStartStop = [
        {
          action = "string"
          cron = {
            expression = "string"
            startTime = "string"
            timeZone = "string"
          }
          recurrence = {
            frequency = "string"
            interval = int
            schedule = {
              hours = [
                int
              ]
              minutes = [
                int
              ]
              monthDays = [
                int
              ]
              weekDays = [
                "string"
              ]
            }
            startTime = "string"
            timeZone = "string"
          }
          schedule = {
            id = "string"
            provisioningStatus = "string"
            status = "string"
          }
          status = "string"
          triggerType = "string"
        }
      ]
    }
    setupScripts = {
      scripts = {
        creationScript = {
          scriptArguments = "string"
          scriptData = "string"
          scriptSource = "string"
          timeout = "string"
        }
        startupScript = {
          scriptArguments = "string"
          scriptData = "string"
          scriptSource = "string"
          timeout = "string"
        }
      }
    }
    sshSettings = {
      adminPublicKey = "string"
      sshPublicAccess = "string"
    }
    subnet = {
      id = "string"
    }
    vmSize = "string"
  }

針對 Databricks,請使用:

  computeType = "Databricks"
  properties = {
    databricksAccessToken = "string"
    workspaceUrl = "string"
  }

針對 DataFactory,請使用:

  computeType = "DataFactory"

針對 DataLakeAnalytics,請使用:

  computeType = "DataLakeAnalytics"
  properties = {
    dataLakeStoreAccountName = "string"
  }

針對 HDInsight,請使用:

  computeType = "HDInsight"
  properties = {
    address = "string"
    administratorAccount = {
      password = "string"
      privateKeyData = "string"
      publicKeyData = "string"
      username = "string"
    }
    sshPort = int
  }

針對 Kubernetes,請使用:

  computeType = "Kubernetes"
  properties = {
    defaultInstanceType = "string"
    extensionInstanceReleaseTrain = "string"
    extensionPrincipalId = "string"
    instanceTypes = {
      {customized property} = {
        nodeSelector = {
          {customized property} = "string"
        }
        resources = {
          limits = {
            {customized property} = "string"
            {customized property} = "string"
          }
          requests = {
            {customized property} = "string"
            {customized property} = "string"
          }
        }
      }
    }
    namespace = "string"
    relayConnectionString = "string"
    serviceBusConnectionString = "string"
    vcName = "string"
  }

針對 SynapseSpark,請使用:

  computeType = "SynapseSpark"
  properties = {
    autoPauseProperties = {
      delayInMinutes = int
      enabled = bool
    }
    autoScaleProperties = {
      enabled = bool
      maxNodeCount = int
      minNodeCount = int
    }
    nodeCount = int
    nodeSize = "string"
    nodeSizeFamily = "string"
    poolName = "string"
    resourceGroup = "string"
    sparkVersion = "string"
    subscriptionId = "string"
    workspaceName = "string"
  }

針對 VirtualMachine,請使用:

  computeType = "VirtualMachine"
  properties = {
    address = "string"
    administratorAccount = {
      password = "string"
      privateKeyData = "string"
      publicKeyData = "string"
      username = "string"
    }
    isNotebookInstanceCompute = bool
    notebookServerPort = int
    sshPort = int
    virtualMachineSize = "string"
  }

屬性值

workspaces/computes

名稱 描述
類型 資源類型 “Microsoft.MachineLearningServices/workspaces/computes@2023-04-01-preview”
NAME 資源名稱 需要字串 ()

字元限制:計算實例 3-24
3-32 用於 AML 計算
2-16 用於其他計算類型

合法字元:
英數字元和連字號。
location 指定資源的位置。 字串
parent_id 此資源父系之資源的標識碼。 類型資源的標識碼: 工作區
tags 包含定義為索引鍵/值組的資源標記。 標記名稱和值的字典。
sku 工作區的 SKU。 Sku
身分識別 資源的身分識別。 ManagedServiceIdentity
properties 計算屬性 計算

ManagedServiceIdentity

名稱 描述
類型 (允許 SystemAssigned 和 UserAssigned 類型的受控服務識別類型) 。 “SystemAssigned”
“SystemAssigned,UserAssigned”
“UserAssigned” (必要)
identity_ids 與資源相關聯的使用者指派身分識別集。 userAssignedIdentities 字典索引鍵的格式為 ARM 資源標識符:'/subscriptions/{subscriptionId}/resourceGroups/{resourceGroupName}/providers/Microsoft.ManagedIdentity/userAssignedIdentities/{identityName}。 字典值可以是要求中 ({}) 的空白物件。 使用者身分識別標識碼的陣列。

UserAssignedIdentities

名稱 描述
{自定義屬性} UserAssignedIdentity

UserAssignedIdentity

此物件不包含部署期間要設定的任何屬性。 所有屬性都是 ReadOnly。

計算

名稱 描述
computeLocation 基礎計算的位置 字串
description Machine Learning 計算的描述。 字串
disableLocalAuth 退出宣告本機驗證,並確保客戶只能單獨使用 MSI 和 AAD 進行驗證。 bool
resourceId 基礎計算的 ARM 資源識別碼 字串
computeType 設定物件類型 AKS
AmlCompute
ComputeInstance
Databricks
DataFactory
DataLakeAnalytics
HDInsight
Kubernetes
SynapseSpark
需要 VirtualMachine ()

AKS

名稱 描述
computeType 計算的類型 “AKS” (必要)
properties AKS 屬性 AKSSchemaProperties

AKSSchemaProperties

名稱 描述
agentCount 代理程式數目 int

約束:
最小值 = 0
agentVmSize 代理程式虛擬機大小 字串
aksNetworkingConfiguration vnet 的 AKS 網路設定 AksNetworkingConfiguration
clusterFqdn 叢集完整功能變數名稱 字串
clusterPurpose 叢集的預定用法 “DenseProd”
“DevTest”
“FastProd”
loadBalancerSubnet Load Balancer 子網 字串
loadBalancerType Load Balancer類型 “InternalLoadBalancer”
“PublicIp”
sslConfiguration SSL 設定 SslConfiguration

AksNetworkingConfiguration

名稱 描述
dnsServiceIP 指派給 Kubernetes DNS 服務的 IP 位址。 它必須位於 serviceCidr 中指定的 Kubernetes 服務地址範圍內。 字串

約束:
模式 = ^(?:(?:25[0-5]|2[0-4][0-9]|[01]?[0-9][0-9]?)\.){3}(?:25[0-5]|2[0-4][0-9]|[01]?[0-9][0-9]?)$
dockerBridgeCidr 指派給 Docker 網橋網路的 CIDR 表示法 IP 範圍。 它不得與任何子網 IP 範圍或 Kubernetes 服務地址範圍重疊。 字串

約束:
模式 = ^([0-9]{1,3}\.){3}[0-9]{1,3}(\/([0-9]|[1-2][0-9]|3[0-2]))?$
serviceCidr 據以指派服務叢集 IP 的 CIDR 標記法 IP 範圍。 此範圍不得與任何子網路 IP 範圍重疊。 字串

約束:
模式 = ^([0-9]{1,3}\.){3}[0-9]{1,3}(\/([0-9]|[1-2][0-9]|3[0-2]))?$
subnetId 計算節點所屬的虛擬網路子網資源標識碼 字串

SslConfiguration

名稱 描述
cert 憑證數據 字串

約束:
敏感性值。 以安全參數的形式傳入。
cname 憑證的 CNAME 字串
索引鍵 索引鍵數據 字串

約束:
敏感性值。 以安全參數的形式傳入。
leafDomainLabel 公用端點的分葉網域標籤 字串
overwriteExistingDomain 指出是否要覆寫現有的網域卷標。 bool
status 啟用或停用ssl以進行評分 “Auto”
“Disabled”
“Enabled”

AmlCompute

名稱 描述
computeType 計算的類型 必要 (“AmlCompute”)
properties AmlCompute 的屬性 AmlComputeProperties

AmlComputeProperties

名稱 描述
enableNodePublicIp 啟用或停用節點公用IP位址布建。 可能的值為:可能的值為:true - 表示計算節點會布建公用IP。 false - 表示計算節點會有私人端點,且沒有公用IP。 bool
isolatedNetwork 網路已隔離或未隔離 bool
osType 計算 OS 類型 "Linux"
"Windows"
propertyBag 包含其他屬性的屬性包。
remoteLoginPortPublicAccess 公用 SSH 埠的狀態。 可能的值為:已停用 - 指出叢集的所有節點上已關閉公用 ssh 埠。 已啟用 - 指出叢集的所有節點上都已開啟公用 SSH 埠。 NotSpecified - 指出如果已定義 VNet,則叢集的所有節點上都會關閉公用 SSH 埠,否則會開啟所有公用節點。 只有在建立叢集之後,才會啟用或停用叢集建立期間的預設。 “Disabled”
“Enabled”
“NotSpecified”
scaleSettings AML 計算的調整設定 ScaleSettings
子網路 計算節點所屬的虛擬網路子網資源標識碼。 ResourceId
userAccountCredentials 將在每個計算節點上建立之系統管理員用戶帳戶的認證。 UserAccountCredentials
virtualMachineImage AML 計算的虛擬機映射 - 僅限視窗 VirtualMachineImage
vmPriority 虛擬機優先順序 “專用”
“LowPriority”
vmSize 虛擬機器大小 字串

ScaleSettings

名稱 描述
maxNodeCount 要使用的節點數目上限 int (必要)
minNodeCount 要使用的節點數目下限 int
nodeIdleTimeBeforeScaleDown 相應減少 amlCompute 之前的節點閑置時間。 此字串必須是 RFC 格式。 字串

ResourceId

名稱 描述
id 資源的標識碼 字串 (必要)

UserAccountCredentials

名稱 描述
adminUserName 可用來透過 SSH 連線至節點的系統管理員用戶帳戶名稱。 字串 (必要)
adminUserPassword 系統管理員用戶帳戶的密碼。 字串

約束:
敏感性值。 以安全參數的形式傳入。
adminUserSshPublicKey 系統管理員用戶帳戶的 SSH 公鑰。 字串

約束:
敏感性值。 以安全參數的形式傳入。

VirtualMachineImage

名稱 描述
id 虛擬機映像路徑 字串 (必要)

ComputeInstance

名稱 描述
computeType 計算的類型 “ComputeInstance” (必要)
properties ComputeInstance 的屬性 ComputeInstanceProperties

ComputeInstanceProperties

名稱 描述
applicationSharingPolicy 在父工作區的用戶之間共用此計算實例上應用程式的原則。 如果為 Personal,則只有建立者可以存取此計算實例上的應用程式。 共用時,任何工作區使用者都可以根據指派的角色存取此實例上的應用程式。 “Personal”
“共用”
autologgerSettings 指定自動記錄工具的設定。 ComputeInstanceAutologgerSettings
computeInstanceAuthorizationType 計算實例授權類型。 可用的值為個人 (預設) 。 “personal”
customServices 新增至計算的自定義服務清單。 CustomService[]
enableNodePublicIp 啟用或停用節點公用IP位址布建。 可能的值為:可能的值為:true - 表示計算節點會布建公用IP。 false - 指出計算節點會有私人端點,且沒有公用 IP。 bool
idleTimeBeforeShutdown 在使用者定義閑置期間之後停止計算實例。 時間是以ISO8601格式定義。 最小值為 15 分鐘,最大值為 3 天。 字串
personalComputeInstanceSettings 個人計算實例的設定。 PersonalComputeInstanceSettings
schedules 要套用至計算的排程清單。 ComputeSchedules
setupScripts 要執行以設定叢集的自定義腳本詳細數據。 SetupScripts
sshSettings 指定 SSH 存取的原則和設定。 ComputeInstanceSshSettings
子網路 計算節點所屬的虛擬網路子網資源標識碼。 ResourceId
vmSize 虛擬機器大小 字串

ComputeInstanceAutologgerSettings

名稱 描述
mlflowAutologger 指出是否為筆記本啟用mlflow自動記錄器。 “Disabled”
“Enabled”

CustomService

名稱 描述
docker 描述映像的 Docker 設定 Docker
端點 設定容器的端點 Endpoint[]
environmentVariables 容器的環境變數 CustomServiceEnvironmentVariables
image 描述影像規格 映像
NAME 自訂服務的名稱 字串
磁碟區 設定容器的磁碟區 VolumeDefinition[]
{自定義屬性}

Docker

名稱 描述
特權 指出容器應該以具特殊許可權或非特殊許可權模式執行。 bool
{自定義屬性}

端點

名稱 描述
hostIp 應用程式從容器公開的主機IP 字串
NAME 端點的名稱 字串
protocol 透過此端點進行通訊的通訊協定 “http”
“tcp”
“udp”
published 應用程式從容器公開的埠。 int
目標 容器內的應用程式埠。 int

CustomServiceEnvironmentVariables

名稱 描述
{自定義屬性} EnvironmentVariable

EnvironmentVariable

名稱 描述
類型 環境變數的類型。 可能的值為:local - 針對局部變數 “local”
環境變數的值 字串
{自定義屬性}

映像

名稱 描述
參考 影像參考 URL 字串
type 影像的類型。 可能的值為:docker - 針對 docker 映射。 azureml - 適用於 AzureML 映射 “azureml”
“docker”
{自定義屬性}

VolumeDefinition

名稱 描述
bind 掛接的系結選項 BindOptions
一致性 磁碟區的一致性 字串
readOnly 指出是否要將磁碟區掛接為 readOnly。 默認值為 false。 bool
來源 掛接的來源。 對於系結掛接,這是主機路徑。 字串
目標 掛接的目標。 對於系結掛接,這是容器中的路徑。 字串
Tmpfs 掛接的 tmpfs 選項 TmpfsOptions
類型 磁碟區定義的類型。 可能的值:bind、volume、tmpfs、npipe “bind”
“npipe”
“tmpfs”
“volume”
磁碟區 掛接的磁碟區選項 VolumeOptions

BindOptions

名稱 描述
createHostPath 指出是否要建立主機路徑。 bool
傳播 系結選項的類型 字串
selinux 提及 selinux 選項。 字串

TmpfsOptions

名稱 描述
{1}size{2} 提及 Tmpfs 大小 int

VolumeOptions

名稱 描述
nocopy 指出磁碟區是否為 nocopy bool

PersonalComputeInstanceSettings

名稱 描述
assignedUser 明確指派給個人計算實例的使用者。 AssignedUser

AssignedUser

名稱 描述
objectId 使用者的 AAD 物件識別碼。 需要字串 ()
tenantId 使用者的 AAD 租使用者識別碼。 需要字串 ()

ComputeSchedules

名稱 描述
computeStartStop 要套用的計算開始停止排程清單。 ComputeStartStopSchedule[]

ComputeStartStopSchedule

名稱 描述
動作 [必要]計算能力動作。 "Start"
“Stop”
Cron 如果 triggerType 為 Cron,則為必要項。 Cron
週期 如果 triggerType 為 Recurrence,則為必要項。 週期性
schedule [已被取代]未再使用。 ScheduleBase
status 排程是否已啟用或停用? “Disabled”
“Enabled”
triggerType [必要]排程觸發程序類型。 “Cron”
“週期”

Cron

名稱 描述
expression [必要]指定排程的cron運算式。
表達式應遵循 NCronTab 格式。
字串
startTime yyyy-MM-ddTHH:mm:ss 格式的開始時間。 字串
timeZone 指定排程執行所在的時區。
TimeZone應遵循 Windows 時區格式。 請參閱: https://docs.microsoft.com/windows-hardware/manufacture/desktop/default-time-zones />
字串

定期

名稱 描述
frequency [必要]觸發排程的頻率。 “Day”
“Hour”
“Minute”
“Month”
“Week”
interval [必要]指定與頻率搭配的排程間隔 int
schedule [必要]週期排程。 RecurrenceSchedule
startTime yyyy-MM-ddTHH:mm:ss 格式的開始時間。 字串
timeZone 指定排程執行所在的時區。
TimeZone應遵循 Windows 時區格式。 請參閱: https://docs.microsoft.com/windows-hardware/manufacture/desktop/default-time-zones />
字串

RecurrenceSchedule

名稱 描述
小時 [必要]排程的時數清單。 int[] (必要)
分鐘 [必要]排程的分鐘清單。 int[] (必要)
monthDays 排程的月份天數清單 int[]
weekDays 排程的天數清單。 包含任何項目的字串數組:
"Friday"
"Monday"
"Saturday"
"Sunday"
"Thursday"
"Tuesday"
"Wednesday"

ScheduleBase

名稱 描述
id 排程的系統指派標識碼。 字串
provisioningStatus 排程的目前部署狀態。 “Completed”
“Failed”
「布建」
status 排程是否已啟用或停用? “Disabled”
“Enabled”

SetupScripts

名稱 描述
指令碼 自訂設定文稿 ScriptsToExecute

ScriptsToExecute

名稱 描述
creationScript 在布建計算期間只執行一次的腳本。 ScriptReference
startupScript 每次機器啟動時執行的腳本。 ScriptReference

ScriptReference

名稱 描述
scriptArguments 傳遞至要執行的腳本的選擇性命令行自變數。 字串
scriptData 掛接磁碟區中的腳本位置。 字串
scriptSource 腳本的儲存來源:內嵌、工作區。 字串
timeout 傳遞至逾時命令的選擇性時間週期。 字串

ComputeInstanceSshSettings

名稱 描述
adminPublicKey 將 SSH rsa 公鑰檔案指定為字串。 使用 「ssh-keygen -t rsa -b 2048」 來產生 SSH 金鑰組。 字串
sshPublicAccess 公用 SSH 埠的狀態。 可能的值為:已停用 - 指出此實例上的公用 SSH 埠已關閉。 已啟用 - 指出公用 ssh 埠已開啟,並視情況根據 VNet/子網原則進行存取。 “Disabled”
“Enabled”

Databricks

名稱 描述
computeType 計算的類型 ) 所需的 「Databricks」 (
properties Databricks 的屬性 DatabricksProperties

DatabricksProperties

名稱 描述
databricksAccessToken Databricks 存取令牌 字串
workspaceUrl 工作區 URL 字串

DataFactory

名稱 描述
computeType 計算的類型 必要) “DataFactory” (

DataLakeAnalytics

名稱 描述
computeType 計算的類型 “DataLakeAnalytics” (必要)
properties DataLakeAnalyticsSchemaProperties

DataLakeAnalyticsSchemaProperties

名稱 描述
dataLakeStoreAccountName DataLake Store 帳戶名稱 字串

HDInsight

名稱 描述
computeType 計算的類型 “HDInsight” (必要)
properties HDInsight 計算屬性 HDInsightProperties

HDInsightProperties

名稱 描述
address 叢集主要節點的公用IP位址。 字串
administratorAccount 管理員 叢集主要節點的認證 VirtualMachineSshCredentials
sshPort 針對叢集主要節點上的SSH連線開啟埠。 int

VirtualMachineSshCredentials

名稱 描述
password 系統管理員帳戶的密碼 字串
privateKeyData 私鑰資料 字串

約束:
敏感性值。 以安全參數的形式傳入。
publicKeyData 公鑰資料 字串

約束:
敏感性值。 以安全參數的形式傳入。
使用者名稱 系統管理員帳戶的用戶名稱 字串

Kubernetes

名稱 描述
computeType 計算的類型 “Kubernetes” (必要)
properties Kubernetes 的屬性 KubernetesProperties

KubernetesProperties

名稱 描述
defaultInstanceType 預設執行個體類型 字串
extensionInstanceReleaseTrain 擴充實例發行定型。 字串
extensionPrincipalId 延伸模組principal-id。 字串
instanceTypes 實例類型架構 KubernetesPropertiesInstanceTypes
命名空間 計算命名空間 字串
relayConnectionString 轉 連接字串。 字串

約束:
敏感性值。 以安全參數的形式傳入。
serviceBusConnectionString ServiceBus 連接字串。 字串

約束:
敏感性值。 以安全參數的形式傳入。
vcName VC 名稱。 字串

KubernetesPropertiesInstanceTypes

名稱 描述
{自定義屬性} InstanceTypeSchema

InstanceTypeSchema

名稱 描述
nodeSelector 節點選取器 InstanceTypeSchemaNodeSelector
resources 此實例類型的資源要求/限制 InstanceTypeSchemaResources

InstanceTypeSchemaNodeSelector

名稱 描述
{自定義屬性} 字串

InstanceTypeSchemaResources

名稱 描述
限制 此實例類型的資源限制 InstanceResourceSchema
requests 此實例類型的資源要求 InstanceResourceSchema

InstanceResourceSchema

名稱 描述
{自定義屬性} 字串
{自定義屬性} 字串

SynapseSpark

名稱 描述
computeType 計算的類型 “SynapseSpark” (必要)
properties SynapseSparkProperties

SynapseSparkProperties

名稱 描述
autoPauseProperties 自動暫停屬性。 AutoPauseProperties
autoScaleProperties 自動調整屬性。 AutoScaleProperties
nodeCount 目前指派給計算的計算節點數目。 int
nodeSize 節點大小。 字串
nodeSizeFamily 節點大小系列。 字串
poolName 集區名稱。 字串
resourceGroup 工作區所在的資源組名。 字串
sparkVersion Spark 版本。 字串
subscriptionId Azure 訂用帳戶識別碼。 字串
workspaceName Azure Machine Learning 工作區的名稱。 字串

AutoPauseProperties

名稱 描述
delayInMinutes int
已啟用 bool

AutoScaleProperties

名稱 描述
已啟用 bool
maxNodeCount int
minNodeCount int

VirtualMachine

名稱 描述
computeType 計算的類型 “VirtualMachine” (必要)
properties VirtualMachineSchemaProperties

VirtualMachineSchemaProperties

名稱 描述
address 虛擬機的公用IP位址。 字串
administratorAccount 管理員 虛擬機的認證 VirtualMachineSshCredentials
isNotebookInstanceCompute 指出此計算是否將用於執行筆記本。 bool
notebookServerPort 針對SSH連線開啟的筆記本伺服器埠。 int
sshPort 針對SSH連線開啟的埠。 int
virtualMachineSize 虛擬機器大小 字串

SKU

名稱 描述
處理能力 如果 SKU 支援相應放大/縮小,則應該包含容量整數。 如果資源無法相應放大/縮小,則可能會省略此專案。 int
family 如果服務有不同層代的硬體,則針對相同的SKU,則可以在這裡擷取。 字串
NAME SKU 的名稱。 例如 - P3。 通常是字母+數字代碼 字串 (必要)
{1}size{2} SKU 大小。 當名稱欄位是階層和其他一些值的組合時,這會是獨立程序代碼。 字串
tier 如果服務有多個層級,但 PUT 上不需要,則資源提供者必須實作此字段。 "Basic"
“Free”
“Premium”
“Standard”