Checkmark=facebooklinklinkedinOngoing indicator +Right-facing arrowCheckmarkDown-pointing chevronLeft-pointing chevronRight-pointing chevrontwitter

Quickstart: Using private images

Now that you know how to use datasets, you may want to use a private Docker image from a registry of your choice. For example, there are registries like Docker store and

Run it on Nerdalize

To allow your private image to start on the Nerdalize cluster, you need to create what we call a ‘secret’. A secret can be one of two types: ‘opaque’ or ‘registry’. For now, we’ll focus on the registry secret type. This informs the Nerdalize cluster from which registry it should pull the private image and what credentials it should use.

For this example, we’ll use the registry. Keep in mind that the name of the secret needs to be a real registry name.

  1. Create a secret

    $ nerd secret create --name=<name> --password=<password> --type=registry
    New Secret:

    Let’s unpack this:

    • nerd secret create tells the CLI to create a secret in the cluster.
    • specifies from which registry your private image should be pulled.
    • --name=<name> --password=<password> --type=registry gives the right information to create the secret.
      • --name=<name> --password=<password> collects the credentials to be able to have access to your image.
      • --type=registry is an important parameter as it specifies which kind of secret should be created.
  2. List your secrets

    $ nerd secret list
  3. Start your workload

    $ nerd workload start <your_image> \
    --input-dataset <dataset id> --instances=5
    Workload created with ID: caa9ffb86d65b70f8903
  4. If needed, delete your secret

    $ nerd secret delete
    Secret successfully deleted

From here the instructions are the same as in the the simple start. Indeed, the Nerdalize cloud knows how to use secrets on its own, so you can use the exact same commands.