Documentation

Deploying Containers

  1. Enter https://Atlas IES IP address in the search box of the browser and press Enter to go to the Atlas IES page.
  2. Choose Management > Service Software > Service Instances, and click Add Service Instance.

  3. On the Add Service Instance page displayed, enter the service instance Name and Description.

  4. In the Container Information area, configure the container information as prompted. For details, see Table 1.

    Table 1 Item description

    Item

    Description

    Container Image File

    Click to upload the image file. The file must be in *.tar, *.tar.gz, and *.zip format and cannot exceed 512 MB (including the file decompressed).

    If you set Verify Image to Yes, the uploaded image file must contain the CMS digital signature and CRL.

    NOTE:

    If the container fails to be deployed, delete the image manually if it is not required.

    Verify Image

    Specifies whether to verify the uploaded image file.

    • If you select Yes, ensure that the uploaded image file contains the CMS digital signature and CRL. Otherwise, the verification fails.
    • If you select No, the CMS digital signature and CRL are not verified.
    NOTE:

    If the image integrity is not verified, security risks may exist. You are advised to verify the image integrity.

    Resource Restriction
    Indicates limits on the following resources:
    • CPU: specifies the maximum number of CPU cores that can be used by the container.
    • Memory: specifies the maximum memory capacity that can be used by the container.
    • AI Compute Power: specifies the maximum AI compute power that can be used by the container. (This option is displayed only when the Atlas 500 AI edge station (model 3000) is configured with Atlas 200 AI accelerator modules (EP).)
    NOTE:
    • Currently, multiple containers cannot share NPU resources on the Atlas IES. However, you can enable NPU resource sharing on FusionDirector. (To ensure data security, share NPU resources in single-tenant scenarios.) For details, see Deploying Containers.
    • The limits of the resources to be configured cannot exceed the container resource limits in the AtlasEdge container security configuration file podconfig.json. For details, see Modify the security configurations of the AtlasEdge container.
  5. In the Environment Variable area, set Variable and Value.

    • System environment variables can be configured in the container operating environment and can be modified after the service instance is deployed.
    • The values of environment variables are displayed in plaintext. Do not enter sensitive information. If sensitive information is involved, encrypt it to prevent information leakage.
    • The environment variable name is a string of 1 to 2048 characters and can contain only uppercase letters, lowercase letters, digits, and underscores (_). The name must start with an uppercase letter, a lowercase letter, or underscore (_). The environment variable value is a string of 1 to 2048 characters and can contain only uppercase letters, lowercase letters, digits, and underscores (_).
  6. In the Advanced Settings dialog box, set the container restart policy, container service network, and container user and group IDs.
    • The container restart policy can be set to Restart Upon Failure, Not Restart, or Always Restart.

      If the restart policy of a container is set to Not Restart, the container remains in the abnormal state when an exception occurs.

    • Port Mapping is used for container service network.

      • To delete a port mapping, click Delete in the Operations column.

      • To add a port mapping, click .
      • Configure Container Port, Host Port, and Host IP. The value range supported by the current container port is [1024,65535].

        Before configuring a host port, you are advised to log in to the Atlas 500 AI edge station (model 3000) background and run the usr/bin/netstat -tuln command to check whether the host port to be configured is occupied.

    • Set Container User and Group ID: After you click to change it to , the function of setting the user ID and group ID for the container is enabled by default. If this function is disabled, the user and group specified in the Dockerfile used to create the container image are used to run the container. If the Dockerfile is not specified, the root user and group are used to run the container. You are advised to run the container a non-root user and group to restrict the container permission.
      • User ID: ID of the user who runs the container. The value ranges from 1 to 65535. Only non-root users are supported.
      • User Group ID: ID of the user group who runs the container. The value ranges from 1 to 65535. Only non-root groups are supported.
      • When setting the user ID and group ID, ensure that the user ID and group ID exist in the container image. For details about the user ID and group ID in the container image, see "Container Usage Guide" in the MindX Toolbox User Guide.
      • By default, the MindXEdge user is used on the Atlas 500 AI edge station (model 3000). The user ID and user group ID are 1024 by default.
    Figure 1 Advanced settings
  7. Click OK. The system automatically starts the deployment.

    When the deployment is successful, the newly added service instance is displayed in the list. If Running Status is running, the service is successfully deployed and the service instance is automatically started. In addition, you can stop, start, and modify the service instance.

    • The deployment requires a long time. Do not close the page before the deployment result is displayed. Otherwise, the deployment will fail.
    • If the running status of a deployed service instance is exited, the service instance cannot be modified. However, all other operations (such as start and stop) can be performed on the service instance.