All Products
Search

This Product

    DataWorks:O&M Assistant

    Document Center

    DataWorks:O&M Assistant

    Last Updated:Feb 08, 2026

    Install third-party packages for tasks running on exclusive resource groups for scheduling using O&M Assistant. Use one-click installation for common packages or Shell commands for custom packages.

    Usage notes

    • Resource groups: Exclusive resource groups for scheduling only.

    • Permissions required: You need one of the following policies: AliyunDataWorksFullAccess or ModifyResourceGroup.

      For more information, see Product and console access control: RAM Policy.

    • Packages installed on an exclusive resource group for scheduling affect all tasks running on that group. Version conflicts or compatibility issues may impact all tasks. Carefully select and test packages before installation.

    Quotas and limits

    • Command management: You cannot modify existing installation commands. Create new commands or delete existing ones.

    • Applicable node types: Python packages installed via O&M Assistant work only for Python nodes (such as PyODPS 2 or PyODPS 3) running on the resource group.

      For MaxCompute Python UDFs, see UDF example: Use third-party packages in Python UDFs.

    • Command support: Custom commands do not support pip commands.

    Access O&M Assistant

    1. Log on to the DataWorks console. In the top navigation bar, select the desired region. In the left-side navigation pane, click Resource Group to go to the Resource Groups page.

    2. On the Exclusive Resource Groups tab, find the resource group for Data Scheduling, click the image.png icon next to the Resource Group, and select O&M Assistant.

    Install a third-party package

    Create an installation command

    1. On the O&M Assistant page, click Create Command and select an installation method.

      Quick installation

      Use this method for common packages preset in DataWorks.

      image

      Parameters:

      Parameter

      Description

      Command Name

      The custom identifier for the command.

      Command Type

      Select Quick Installation.

      DataWorks generates the Shell installation command automatically.

      Package to Install

      Select the package and version to install. DataWorks supports common packages for Python 2, Python 3, and Yum, including:

      • Aliyun-python-sdk-core: Core library for Alibaba Cloud Python SDK

      • NumPy: High-performance multi-dimensional arrays and numerical computing

      • Pandas: Data structures and analysis tools for structured data.

      See the interface for the complete list.

      Generated Shell Script

      DataWorks auto-generates the installation command.

      Example: Selecting aliyun-python-sdk-core generates pip install aliyun-python-sdk-core.

      Timeout

      Maximum execution time in seconds. The command stops if it times out.

      Manual installation

      Use custom Shell commands when built-in packages don't meet your needs.

      image

      Parameter

      Description

      Command Name

      The custom identifier for the command.

      Command Type

      Select Manual Installation.

      Command Content

      Enter the command to execute. Example: yum install -y git.

      Important

      • Custom commands cannot use pip commands.

      • Test commands before executing in production.

      • Reference packages using absolute paths.

      Installation Directories

      Directory paths the command can access. Separate multiple paths with semicolons (;). DataWorks adds these to a whitelist for access.

      • Install to /home/ or non-/home/ directories. For /home/ installations, use: /home/admin/usertools/tools/.

      • Default: /home/admin/usertools/tools/.

      Timeout

      Maximum execution time in seconds. The command stops if it times out.

    2. Click OK to save the command configuration.

    Execute the installation command

    On the O&M Assistant page, find your command and click Run command in the Actions column. The system generates a command execution record.

    View execution results

    In the command record, view Execution Status and Command Content. Click View Result in the Actions column to analyze logs if the command fails. After successful installation, the package is available in tasks running on the current resource group.

    image

    View the environment configuration

    On the O&M Assistant page, click View Detailed Environment Configuration in the upper-left corner to see installed packages, versions, and status for the current resource group.

    Use installed packages

    New version of Data Studio:

    1. Go to Data Studio: Go to the DataWorks Workspaces page, switch to your target region, find your workspace, and click Shortcuts > DataStudio.

    2. Configure the resource group: In your task node, click Scheduling in the right pane.

      • Resource Group: Select Exclusive Resource Group for Scheduling.

        If the target resource group is not displayed, go to the Resource Group page and click Associate Workspace.

    3. Debug the node: In the Run Configuration pane, configure Computing Resource, Resource Group, and Script Parameters, then click Running Duration in the toolbar.

    4. Publish the node: Click Publish in the toolbar to deploy to production.

    Legacy version of DataStudio:

    1. Go to DataStudio: Log on to the DataWorks console, switch to your region, click Data Development and O&M > Data Development, select your workspace, and click Go to Data Development.

    2. Configure the resource group: In your task node, click Properties in the right pane.

      • Resource Group: Select Exclusive Resource Group for Scheduling.

        If the target resource group is not displayed, go to the Resource Group page and click Associate Workspace.

    3. Debug the node: Click Run with Parameters (image), configure Resource Group Name, then click Run.

    4. Publish the node: Click Save and Submit to deploy to production.

    References

    For detailed examples of using installed packages: