[BUG] Downloading (from the client side) shouldn't happen when we deploying on Kubbernetes
- Dominant language
- C++
- Stars
- 3.6k
- Forks
- 468
- Avg merge
- 29m
- Merged PRs (30d)
- 1
Description
**Describe the bug**
We have documentation for how to loading builtin dataset on Kubernetes [loading_graph.rst#load-built-in-datasets](https://github.com/alibaba/GraphScope/blob/main/docs/loading_graph.rst#load-built-in-datasets), but if user forget to mount the dataset when deploying, and trying to loading builtin dataset, we still do download operation on the client side. It doesn't make sense and user would be thrown a more misleading error message that says "IOError: failed to open file"...
There should be an error when user trying to loading the builtin dataset on kubernetes if no prefix is provided.
Besides, why didn't we `~/.graphscope/dataset` for the default of value `mount_dataset`? Introducing `/dataset` here can only leads to more confusion...
**Additional context**
Add any other context about the problem here.
Contributor guide
Assessment
This issue has not been assessed yet.