Skip to content

Latest commit

 

History

History
66 lines (38 loc) · 2.06 KB

installation_guide.md

File metadata and controls

66 lines (38 loc) · 2.06 KB
Raw

Installation Guide

Process

1. Configure Environment

The below variables control authorization to the Anaconda Data Science Platform. These should be defined as AE5 secrets for the current user. Alternatively they can also be set within the anaconda-project.yml project files. See Variables below for specific details on each.

Variable Workflow Serving
AE5_HOSTNAME X
AE5_USERNAME X
AE5_PASSWORD X
ADSP_WORKER_MAX X
MLFLOW_TRACKING_URI X[1] X[1]
MLFLOW_REGISTRY_URI X[1] X[1]
MLFLOW_TRACKING_TOKEN X[2] X[2]

[1] It's recommended to set these AE5 installation wide as per Adding MLflow to Anaconda Enterprise 5.

[2] This should be provided by your admin as per Adding MLflow to Anaconda Enterprise 5.

2. Install Plugin

conda install -c https://conda.anaconda.org/ae5-admin mlflow-adsp

Variables

Order of resolution

  1. Anaconda Project Variables
  2. Anaconda Data Science User Secrets

If the below variables are defined within an Anaconda Project and as secrets within the Anaconda Data Science Platform, then the Secrets will have a higher precedence and be used.

  1. AE5_HOSTNAME

    Description

    • The FQDN (Fully Qualified Domain Name) for the Anaconda Data Science Platform instance.

  2. AE5_USERNAME

    Description

    • The username of the account to create the jobs with. This should be an account who owns the project or is a collaborator on the project.

  3. AE5_PASSWORD

    Description

    • The password for the user account used for start jobs.

  4. ADSP_WORKER_MAX

    Description

    • The default per-project number of background jobs the user will leverage during parallel execution.