Grant permissions to the operating account and PAI service account to access Designer and dependent services such as MaxCompute, OSS, Flink, and general training resources.
Service dependencies
Designer depends on OSS and may also depend on MaxCompute, Flink, or general training resources. Required permissions vary by scenario.
MaxCompute: Designer includes hundreds of proprietary Alibaba algorithms built on MaxCompute.
Flink: Designer provides algorithm components that run on Flink, such as the PyAlink Script component.
OSS: Intermediate data and models generated during training are stored in OSS. Activate OSS and grant permissions before using Designer.
General training resources: Designer provides deep learning algorithms trained using general training resources. The custom Python Script component also depends on general training resources. Activate this service and grant permissions before using these features.
Log on to the PAI console and click to view cloud product dependencies and authorization details for each feature module.
Grant the operating account required permissions for Designer, MaxCompute, general training resources, Flink, and OSS. Also grant PAI permission to access OSS. To use all Designer features, grant permissions as described in the following sections:
Grant operating account permissions
Designer provides a visual development environment for the machine learning lifecycle and includes built-in algorithms. When using Designer for model development, activate the following cloud products and grant permissions in advance.
PAI sub-product: Designer
Account type
Scenarios
Procedure
Root account
Root account can perform all operations in Designer without additional authorization.
Not applicable
Resource Access Management (RAM) user
(Recommended)
PAI provides different member roles. Add a RAM user as a workspace member and assign a role as needed. For more information about role permissions, see Appendix: Roles and permissions.
Other dependent PAI sub-products: General computing resources
Designer requires computing resources for AI development. Use PAI general computing resources.
Use a root account to purchase general computing resources. If using a RAM user, grant the
AliyunPAIFullAccesspermission. For more information about permissions, see Permission management.To authorize a RAM user to submit training tasks to general computing resources associated with a workspace, add the RAM user to the workspace and assign the Developer, Algorithm O&M Engineer, or Administrator role.
Other dependent cloud products: OSS
Deep learning algorithm components depend on OSS data sources. Activate OSS and grant permissions before use.
Sub-scenario
Description
Procedure
Activate OSS
Use a root account to activate OSS. No additional authorization is required for root accounts. To use a RAM user to activate OSS, grant the
AliyunOSSFullAccesspermission to the RAM user.Activation: Activate OSS
Grant permissions to a RAM user: RAM Policy
Common operations: Create a bucket in the console
Use OSS
When using OSS:
Authorization: OSS provides detailed RAM control policies. Grant operation permissions to RAM users as needed.
Common operations: Create a bucket to upload files to OSS.
When granting an operating account permissions to access OSS, use a custom policy to define access permissions for a RAM user to OSS data in the PAI console. Perform the following steps:
Log on to the RAM console.
In the navigation pane on the left, choose .
On the Policies page, click Create Policy.
On the Create Policy page, click the Script Editor tab.
Enter the policy document and click OK.
OSS provides a complete permission control system. For more information about OSS authorization policies, see RAM Policy.
ImportantDefine the access policy carefully based on the permissions required by the RAM user.
Using OSS in the PAI console involves common operations such as listing accessible buckets and reading or writing data. Use the following custom policy to configure permissions for RAM users performing operations in the PAI console.
{ "Version": "1", "Statement": [ { "Action": [ "oss:GetObject", "oss:ListObjects", "oss:DeleteObject", "oss:ListParts", "oss:PutObject", "oss:AbortMultipartUpload", "oss:GetBucketCors", "oss:GetBucketCors", "oss:DeleteBucketCors" ], "Resource": [ "acs:oss:*:*:<yourBucketName>", "acs:oss:*:*:<yourBucketName>/*" ], "Effect": "Allow" }, { "Action": [ "oss:ListBuckets" ], "Resource": "*", "Effect": "Allow" } ] }Replace <yourBucketName> in the policy with the bucket name for which permissions are granted.
Enter a Name and Note for the access policy and click OK.
Grant PAI service account permissions
Authorization is usually completed when activating PAI. If authorization was not completed during activation, grant permissions by following these steps:
Log on to the PAI console.
In the navigation pane on the left, click . In the Designer section, find OSS.
In the Actions column, check the authorization status of OSS.
If permissions are not granted, click Authorize in the Actions column and follow the on-screen instructions to complete authorization.
After authorization is complete, click View Authorization Information in the Actions column to view details.