Hi SubstantialElk6 ,
I would also remove:CLEARML__apiserver__pre_populate__enabled: "true" CLEARML__apiserver__pre_populate__zip_files: "/opt/clearml/db-pre-populate"
As these pre-populated examples depend on files stored in the fileserver (if you want these examples with missing files, you can keep it).
would they need the fileserver to route to minio then?
Correct :)
But if user forgot to do above, they will be saved on ClearML server. If I switch off file_server, the configuration above will break right?
Yeah, there's no way around that - if the SDK tries to access a non-existing storage service, you'll get an error.
However, if you can control each user's environment and make sure to set CLEARML_FILES_HOST=
s3://ecs.ai:80/clearml-data/default env var, than in that case, the users will only have to configure the credentials (and if they forget, they will get an error saying they don't have permissions to this bucket)