Reamer edited a comment on pull request #4097: URL: https://github.com/apache/zeppelin/pull/4097#issuecomment-827815927
You are right, putting the Conda environment in a cloud storage will be the best option. Do you have any idea what possibilities for integration `spark.archives` supports? Local mounting via filesystem is not an option in Kubernetes. I am hoping for an HTTP endpoint, which is very flexible and should work for most users. YARN should also be fine with an HTTP endpoint, so that the Conda environment can be dynamically loaded by the Zeppelin interpreter when the Python or PySpark interpreter is started. I think it would be great if the Zeppelin interpreter, independent of YARN and K8s, load the conda environment via the same mechanism. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org
