search

project-codeflare / rayvens

Rayvens makes it possible for data scientists to access hundreds of data services within Ray with little effort.
Apache License 2.0
43 stars 7 forks source link

Eliminate dependency on Camel-K CLI in operator mode #8

Open tardieu opened 3 years ago

tardieu commented 3 years ago

Rayvens currently uses the Camel-K CLI to deploy Camel integrations with the Camel-K operator. Therefore, even in operator mode, the Ray image has to be augmented with the CLI to run Rayvens program. If we were to deploy integrations by directly interfacing with Kubernetes using the kubernetes Python library (or kubectl) already included in the base Ray image, we could eliminate the need for including the Camel-K CLI in the Ray image in operator mode.

tardieu commented 3 years ago

To be clear, we still need the Camel-K operator to be deployed to the Kubernetes cluster but this is typically done beforehand, independently from deploying and configuring the Ray cluster itself.