As a user working across 3 different High-Performance Computing (HPC) platforms provided by my school, I encounter a workflow issue with the VSCode Remote SSH extension and Jupyter server management. My typical workflow is as follows:
Connect to the HPC management node using VSCode's Remote SSH extension.
Submit a job that invokes a Jupyter server on a computation node, granting me access to GPU resources.
Open a Jupyter notebook and connect to the Jupyter server I just launched on the computation node to begin my work.
However, I've observed that the list of Jupyter remote servers seems to be stored locally across all environments. For instance, the remote server from HPC1 appears on HPC2 and also on my local laptop's VSCode. This is problematic and confusing because Jupyter servers are not accessible outside their host HPC environment.
Suggested Feature:
To resolve this confusion and enhance user experience, I propose adding an option to manage where Jupyter remote server entries are stored:
Local Storage Option: For Jupyter servers that are publicly accessible on the network, provide an option to store server details locally.
Host-Specific Storage Option: For servers like those in HPC environments that aren't accessible outside their network, offer an option to store the server details only on the host machine.
This feature would significantly streamline workflows for users like me who operate across multiple HPC environments, reducing confusion and improving efficiency by ensuring that Jupyter server lists are relevant to their specific environment and accessible status.
Reason for Request:
As a user working across 3 different High-Performance Computing (HPC) platforms provided by my school, I encounter a workflow issue with the VSCode Remote SSH extension and Jupyter server management. My typical workflow is as follows:
However, I've observed that the list of Jupyter remote servers seems to be stored locally across all environments. For instance, the remote server from HPC1 appears on HPC2 and also on my local laptop's VSCode. This is problematic and confusing because Jupyter servers are not accessible outside their host HPC environment.
Suggested Feature:
To resolve this confusion and enhance user experience, I propose adding an option to manage where Jupyter remote server entries are stored:
This feature would significantly streamline workflows for users like me who operate across multiple HPC environments, reducing confusion and improving efficiency by ensuring that Jupyter server lists are relevant to their specific environment and accessible status.