The built\-in **artifact download** step in continuous delivery supports downloading artifacts in a common format to pipeline tasks, making it convenient for subsequent steps in the same task to consume the artifact. This article introduces how to download artifacts. ## Background information Artifacts to be downloaded can come from: * Volcano Engine artifact repository. * Reference the product of the upstream **artifact upload** task. ## Prerequisites * Artifact repository * The Volcano Engine artifact repository has been activated.

tip

The artifact repository is currently in beta stage. To try it, please apply for trial.

* An artifact repository instance has been created. For specific operations, see [Create basic-edition instances](https://www.volcengine.com/docs/6420/1764549). * A Generic artifact format repository has been created, and the repository's release policy must not be read\-only. For specific operations, see [Create Generic repository](https://www.volcengine.com/docs/6420/1764555). * Continuous delivery * To use artifact download for the first time, cross\-service authorization must be completed. For specific operations, see [Cross-service authorization](https://www.volcengine.com/docs/6461/1148757). ## Add step 1. When creating or editing a pipeline, in the **Process orchestration** tab, select an existing task or add a new task. 2. In the task editing panel, click **Add step**, then select **Predefined steps \> Artifact \> Artifact download.** ## Parameter description The following describes the parameters for the artifact download step. |Parameter |Description | |---|---| |Step name |The default is artifact download; custom step names are supported. | |Artifact source |Select the artifact source; artifact repositories and upstream products are supported.

* **Artifact repository**: The artifact source is the Volcano Engine artifact repository. You need to further select the artifact repository's region, instance name, repository URL, and enter the artifact download link. For specific parameter descriptions, see the table below.

* **Upstream product**: Reference the product of the upstream **artifact upload** task. You need to further select the target upstream product. | |Download to container path |The directory path in the container where the artifact will be downloaded.

* Only directory paths are supported; file paths are not supported. If you enter a file path, the system will automatically add `/` at the end, converting it to a directory path. For example: the artifact name is `a.gzip`, and the container path is `./test`. The system will automatically complete the container path as `./test/`, then perform the download operation. The artifact's final path in the container is `./test/a.gzip`.

* The default is `.`, which means downloading to the container's current directory. | |Skip if failed |Optional configuration.

* Turn on: If the current step fails, it will be skipped directly. The remaining steps of the current task can continue to execute, and the pipeline can also continue to run.

* Turn off: After the current step fails, the remaining steps of the current task will not be executed, and the pipeline will fail. | When the artifact source is artifact repositories, the parameters are as follows: |Parameter |Description | |---|---| |Region |Select the region where the artifact repository is located. | |Instances |Select the artifact repository instance where the artifact is located. | |Repository URL |Select the target repository where the artifact is located. Only repositories that are not read\-only and use the Generic artifact format are supported. | |Artifact download link |Enter the download path for the artifact. The format is `https://xxx.com/repository//`.

It is recommended to copy the artifact's **path** from the artifact details page in the artifact repository console to avoid errors. | ## Advanced settings * **LimitRange**: Supports setting upper limits for CPU and memory. * **Disk capacity limit**: Supports setting the upper limit for disk capacity, up to 100 GiB. * **Timeout time**: Sets the timeout for the entire task. The task will automatically fail after timing out. A task can include multiple serial steps. The total execution time of all steps must be less than the task's timeout time. * [Webhook notification](https://www.volcengine.com/docs/6461/1251516): Supports pushing task\-level messages via Webhook and Feishu Bot, helping you stay informed about task status in a timely manner. * **Automatic retry on failure**: After enabling, enter the maximum number of retries supported. When task execution fails, it will retry up to the configured maximum number of times.

warning

For pipeline tasks in the private build resource pool, only Agent versions greater than or equal to v2.34.0 support Automatic retry on failure. For versions below v2.34.0, re\-adding the Agent will upgrade it to the latest version. For the method to add an Agent, see Create private resource pool in the what to do next section.

## Log After this step is complete, the pipeline shows the following details. * Execution log: Displays the real\-time execution log for artifact download. * Task configuration: Displays the configuration parameters for this task, including the configuration parameters for all steps.

Code Pipeline