search

canonical / bundle-kubeflow

Charmed Kubeflow
Apache License 2.0
104 stars 50 forks source link

In Kubeflow deployment without MLflow we cannot deploy model with Seldon from Minio #560

Open Barteus opened 1 year ago

Barteus commented 1 year ago

The secret to accessing Minio for models is created only if the deployment has MLflow added.

There is a valid use case for deploying a model stored in Minio without MLflow (ie sklearn model hosting using Seldon).

Barteus commented 1 year ago

Workaround - manually create a secret:

apiVersion: v1
kind: Secret
metadata:
  name: seldon-init-container-secret
type: Opaque
stringData:
  RCLONE_CONFIG_S3_TYPE: s3
  RCLONE_CONFIG_S3_PROVIDER: minio
  RCLONE_CONFIG_S3_ACCESS_KEY_ID: minio123
  RCLONE_CONFIG_S3_SECRET_ACCESS_KEY: minio123
  RCLONE_CONFIG_S3_ENDPOINT: http://minio.kubeflow.svc.cluster.local:9000
  RCLONE_CONFIG_S3_ENV_AUTH: "false"