> ## Documentation Index
> Fetch the complete documentation index at: https://docs.scitix.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Push and Pull Container Images

> Use Docker to log in to the registry and push or pull container images.

After pushing a local image to a project, you can pull the image in other environments with access permissions.

## Pre-Operation Confirmation

* A Registry project already exists in the target region, and the current account has read and write permissions for the project.
* Docker is installed in the environment where you run the commands.
* When pushing an image, the image to be pushed has been prepared locally.

## Get push and pull commands

* Get reference commands for login, push, and pull

  On the **Registry** page, click **Push & Pull**.
* Obtain the push command for a project

  Go to the target project details page. In the left navigation pane, click **Registry**, and then click **Get Push Command**.
* Get the pull command for an image

  1. Go to the details page of the target project and click **Registry** in the left navigation pane.
  2. Locate the target image registry and click its name.
  3. In the **Actions** column of the artifact list, click **Copy Pull Command**.

## Log In to the Registry

1. Log in to the [GPU Cloud console](https://console.scitix.ai/gpu-cloud/home).
2. Go to **STORAGE > Registry** and select the target region.
3. Click **Push & Pull**.
4. On the **Docker** tab, obtain the registry address and login instructions for the region.
5. Log in to the registry in the environment where you perform image operations:

   ```bash theme={null}
   docker login <REGISTRY_ENDPOINT> --username <USERNAME>
   ```

   * Replace `<REGISTRY_ENDPOINT>` with the registry address for the current region, for example, `registry-ap-southeast.scitix.ai`.
   * Replace `<USERNAME>` with the Registry Logon Username in the format `<username>@<tenant_name>`.
   * Enter the corresponding credentials as prompted. The default password is the platform logon password.

## Push Image

Log in to the Registry, add the target registry address, project path, and tag to the local image, and push it to the target project.

```bash theme={null}
docker tag <LOCAL_IMAGE>:<TAG> <REGISTRY_ENDPOINT>/<PROJECT>/<IMAGE>:<TAG>
docker push <REGISTRY_ENDPOINT>/<PROJECT>/<IMAGE>:<TAG>
```

* `<LOCAL_IMAGE>`: Local image.
* `<REGISTRY_ENDPOINT>`: Registry endpoint.
* `<PROJECT>`: Target project.
* `<IMAGE>`: Target image name.
* `<TAG>`: Image tag.

Example:

```bash theme={null}
docker tag myimage:latest registry-ap-southeast.scitix.ai/project-test/myimage:latest
docker push registry-ap-southeast.scitix.ai/project-test/myimage:latest
```

After the push is complete, go to the relevant project under the **Projects** tab in the target region, then confirm that the image can be found in **Registry**.

## Pull an Image

Log in to the repository in the environment where you need to use the image, then run:

```bash theme={null}
docker pull <REGISTRY_ENDPOINT>/<PROJECT>/<IMAGE>:<TAG>
```

* `<REGISTRY_ENDPOINT>`: Registry address.
* `<PROJECT>`: Target project.
* `<IMAGE>`: Target image name.
* `<TAG>`: Image tag.

Example:

```bash theme={null}
docker pull registry-ap-southeast.scitix.ai/project-test/myimage:latest
```

Replace the example address with the actual repository address. If the pull fails, first verify the region, project path, image name and tag, and whether the credentials used in the current environment can access the project.

## Frequently Asked Questions

> **What should I do if DENIED: ... exceed the configured upper limit is displayed when pushing an image?**
>
> This error indicates that the storage quota of the target registry project is insufficient or has reached its limit. The registry refuses to accept additional image data.
>
> These issues are usually not caused by Docker login status, image tag format, or network connectivity. You can address them in the following order:
>
> 1. Verify that the project name in the image address is correct to avoid pushing the image to a full or unintended project.
> 2. Delete image versions, tags, or other artifacts that are no longer needed. If an artifact retention policy is configured for the project, we recommend running it in dry-run mode and verifying the impact before performing the cleanup. For details, see [Configure Registry Artifact Retention Policies](https://docs.scitix.ai/gpu-cloud/registry/configure-artifact-retention.md).
> 3. If historical images still need to be retained, contact the project administrator to adjust the project storage quota, or push them to a project with sufficient quota.
> 4. After releasing or adjusting the quota, run `docker push` again.


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.