Updated on 2026-11-12 GMT+08:00

Managing Images

In Management Center, you can view system images and manage custom images. By adding dependency packages to a basic image, you can create a custom image for more flexible development and scheduling. System images are built-in images and cannot be modified or deleted.

Notes and Constraints

  • System images are built-in images and cannot be modified or deleted.
  • The permissions of different roles in the preset roles are as follows:
    • The current workspace administrator and users with the DAYU Administrator or Tenant Administrator permission can create and manage all custom images.
    • Developers can create and manage custom images, and query all custom images in the workspace.
    • O&M personnel and guests can only query all custom images in the workspace.
  • Users with the custom role can create custom images and manage the custom images they have created if they have the Create permission. Users with the Delete, Operation, Query, and Edit permissions can perform the corresponding operations on all custom images in the workspace. For details about how to create a custom role, see (Optional) Defining a Workspace Role.
  • If a user has both preset and custom roles, the user has all the permissions of these roles.
  • You can create up to 100 custom images in a workspace.
  • The name of a custom image must be unique in a workspace.

Viewing System Images

  1. Log in to the DataArts Studio console.

    For details, see Accessing the DataArts Studio Instance Console.

  2. On the DataArts Studio console, locate a workspace and click Management Center.
  3. In the navigation pane on the left, choose Images.
  4. Click the System Images tab to view system images.

    Figure 1 Viewing system images

Creating a Custom Image

You can use a public image and a Dockerfile to create a custom image that contains required dependencies, push the image to SoftWare Repository for Container (SWR), and then register it on the DataArts Studio console. You can then select the registered custom image when creating a notebook runtime environment and configuring notebook job scheduling. In this way, you can manage runtime environments in a standardized manner and reuse them.

  1. Download an image.

    1. Prepare a container environment.

      For details, see "Step 1: Install a Container Engine" in The Lifecycle of an Image.

    2. Select a suitable base image. DataArts Studio supports the base images listed in Table 1.
      Table 1 Notebook base images

      Image Name

      Image Version

      Download Command

      Description

      dgc-notebook-kernel-public-x86_64:26.9.0.20260910160323

      26.9.0.20260910160323

      docker pull swr.cn-north-7.myhuaweicloud.com/dgc-notebook/dgc-notebook-kernel-public-x86_64:26.9.0.20260910160323

      python_version: 3.10.6

    3. Log in to the container environment as user root and run the command for downloading the image selected in 1.b to obtain the image.

      Example:

      docker pull swr.cn-north-7.myhuaweicloud.com/dgc-notebook/dgc-notebook-kernel-public-x86_64:26.9.0.20260910160323

  2. Create a custom image.

    1. Access SWR.
      1. Log in to the SWR console.
      2. In the left navigation pane, choose Dashboard. Click Generate Login Command in the upper right corner. In the window that appears, click to copy the login command.
      3. Run the login command on the VM where the container engine is installed.
    2. Create an SWR organization. If an organization already exists, skip this step.
      1. In the left navigation pane of the SWR console, choose Organizations. Then, click Create Organization in the upper right corner.
      2. Enter the organization name and click OK.
    3. Log in to the container environment as user root and create and compile a Dockerfile.

      This example uses the Python base image. If a Dockerfile starts with the FROM instruction, it must specify the image downloaded in 1 as the base image. After the FROM instruction, you can add instructions to customize the image.

      1. Create a file named Dockerfile.
        mkdir myimage
        cd myimage
        touch Dockerfile
      2. Edit the Dockerfile.
        vim Dockerfile

        Add the following content to the file:

        # Dockerfile
        FROM swr.cn-north-7.myhuaweicloud.com/dgc-notebook/dgc-notebook-kernel-public-x86_64:26.9.0.20260910160323
        
        # MODIFICATIONS GO HERE
        RUN pip install --upgrade pandas==2.1.0
      3. Press Esc and enter :wq to save the changes and exit.
    4. After creating and compiling the Dockerfile, run the following command to build a custom image:

      Example:

      # build the docker image
      docker build -t Huawei Cloud SWR domain name/Organization/Image name:Version
    5. Push the custom image to SWR.

      Go to the SWR homepage, click Generate Login Command to copy the temporary login command, run the command, and push the image to SWR.

      Example:

      # login to SWR
      docker login -u cn-north-7@HST3UHL10SX8xxxxxxxx -p cb50728xxxxx swr.cn-north-7.myhuaweicloud.com
      # push the docker image
      docker push Huawei Cloud SWR domain name/Organization/Image name:Version

  3. Register a custom image.

    1. Log in to the DataArts Studio console.

      For details, see Accessing the DataArts Studio Instance Console.

    2. On the DataArts Studio console, locate a workspace and click Management Center.
    3. In the navigation pane on the left, choose Images. On the displayed page, click the Custom Images tab.
    4. Click Register SWR Image and set the parameters by referring to the following table.
      Table 2 Parameters for registering an image

      Parameter

      Description

      Image Source

      SWR address of the custom image

      Click the image selection box, select the image name and version, and click OK.

      • After you select an image, its organization, name, and version are displayed.
      • If no SWR image is available, create a custom image in SWR by referring to Creating a Custom Image.

      Image Type

      Type of the image to be registered

      CPU Architecture

      CPU architecture for the image to be registered. Currently, only the x86_64 architecture is supported.

      Description

      Description of the custom image. This parameter is optional.

    5. Click Submit to complete the registration.
    6. In the custom image list, you can view the name, type, organization, number of versions, and update time of the custom image.
    7. You can add or delete versions for the custom image as required.
      • Click Add Version in the Operation column to add more versions to the image. A maximum of 100 software versions are supported.
      • Click Delete in the Operation column, review the impact of the operation, enter DELETE as prompted, and then click OK.
        • Deleting an image only disassociates it from SWR. The original image in SWR will remain intact.
        • Deleted images cannot be recovered. Exercise caution when performing this operation.