Datalab Custom Build Deployment for the Google Cloud Platform

Question

For the project, we are trying to expand Google Cloud Datalab and deploy a modified version to the Google Cloud platform. As I understand it, the deployment process usually consists of the following steps:

• Create a Docker Image
• Click it in the Container Registry
• Use the container parameter using the Google Cloud Scanner to specify the correct Docker image, as described here .

Since the container registry is the default, i.e. gcr.io/cloud_datalab/datalab:<tag> , disabled for non-Datalab contributors, we clicked the Docker image on our own container registry, i.e. gcr.io/<project_id>/datalab:<tag> .

However, the Google Cloud deployer retrieves data directly from gcr.io/cloud_datalab/datalab:<tag> (with the tag specified by the container parameter), and does not seem to allow the registry specification of the source container. Deployment is not like open-source, so we cannot deploy our image to Google Cloud.

We looked at creating a custom deployment similar to the one below, but that never starts Datalab, so we suspect the start of the script is more complicated.

Question How can we deploy a Datalab image from our own container registry in Google Cloud?

Thank you very much in advance.

Answer 1

Deployment options can be guessed, but it’s easier to get the Google Cloud Datalab script deployed to ssh with a temporary node computation that is responsible for deploying and viewing the /datalab folder. It contains the runtime configuration file that is used with the flexible App Engine environment . Using this configuration file, the google preview app deploy command (which accepts the --image for Docker images) will correctly implement this in App Engine.