Currently we are trying use DLTK to implement our ML scenarios,but there are some questions that need to be fixed. 1. When create users and assign permissions, I had to add "Power" to users to make them can using the "fit" command, but they still cannot access Containers, the error message is " HTTP 403 Forbidden -- You do not have the capability: admin_all_objects " How can I make non-admin users can start their own containers ,and make sure every user's data is totally separated? 2. If we use Docker as Dev and Kubernetes as Production, how can I build the workflow that a user can easily deploy code in these 2 envs, do you have a more detailed manual? thanks.
... View more