VS Code Remote Development

This guide provides information on running VS Code Remote Development on HiPerGator, including running Jupyter notebooks. This has been updated in December 2022 when Microsoft integrated Remote Development into VS Code.

Introduction

Why might you want to do this?

• VS Code is a nice IDE with many great features, including GitHub Copilot (which most UF users should be able to access for free--via either student or faculty/staff GitHub Education plans).
• HiPerGator provides a powerful compute environment with thousands of cores, petabytes of storage, and powerful GPUs.
• But, combining the two can be problematic...
  • Remote SSH setup is complicated by MFA and the need to tunnel into a compute server with your running job. This can be worked out, but is a bit cumbersome.
  • You can accidentally end up running scripts on the login servers and have your account suspended if you connect via the remote ssh extension and run a notebook or other code outside of Development_and_Testing limits as those processes would be running on login nodes.
  • Running VS Code on HiPerGator through OOD is kind of slow and limiting as vscode environment module updates can lag behind the rapid vscode release schedule.
• With VS Code Remote Development, you can startup the server within an interactive job with whatever resources you have requested and connect to it in your web browser from https://vscode.dev/. There are no MFA issues and you can install your desired extensions, connect to GitHub Copilot, run and debug your code easily!

VSCode Setup

Using hpg env module

If you don't mind staying on our update schedule you can put the following commands into the job script that starts the vscode server and the tunnel in a SLURM job:

export XDG_RUNTIME_DIR=${SLURM_TMPDIR}
module load vscode
code tunnel

= Personal VSCode install

This may become somewhat optional as Research Computing updates VS Code, but this does give you full control over the version to use.

If missing ~/bin, create it with

mkdir ~/bin

1. Download the x64 CLI for Linux package from the VS Code Downloads site .
2. Transfer the download to HiPerGator.
3. Extract the tar.gz file and copy the binary to ~/bin:

tar -xvf vscode_cli_alpine_x64_cli.tar.gz
cp code ~/bin/

Test VSCode Tunnel

See this page for information on development sessions HiPerGator, but something like the following should work:

1. srundev
2. Start VS Code Server: code tunnel
3. Use 'Ctrl=c' key combination to stop the tunnel process.
4. Exit the srun session

Regular Use

Now that VS Code is setup, it should be relatively easy to start. For the most part, the process is:

• Log into HiPerGator
• Submit a tunnel job. E.g. job script (adjust the resource request as needed for your analysis):

CPU only job

#!/bin/bash
#SBATCH --job-name=vscode
#SBATCH --nodes=1
#SBATCH --ntasks=1
#SBATCH --cpus-per-task=2
#SBATCH --mem=8gb
#SBATCH --time=08:00:00 # example 8 hrs
hostname;date;pwd
export XDG_RUNTIME_DIR=${SLURM_TMPDIR}
# HPG install of vscode. Comment the next line out with '#' if using a personal install
module load vscode
# Both hpg and personal install of vscode
code tunnel

GPU job

#!/bin/bash
#SBATCH --job-name=vscode
#SBATCH --partition=gpu
#SBATCH --gpus=a100:1
#SBATCH --nodes=1
#SBATCH --ntasks=1
#SBATCH --cpus-per-task=2
#SBATCH --mem=8gb
#SBATCH --time=08:00:00 # example 8 hrs
hostname;date;pwd
export XDG_RUNTIME_DIR=${SLURM_TMPDIR}
# HPG install of vscode. Comment the next line out with '#' if using a personal install
module load vscode
# Both hpg and personal install of vscode
code tunnel

For an interactive test: Note: srun session will exit if your network connection to HPG is broken or if you exit.

# Start a development session, [https://help.rc.ufl.edu/doc/GPU_Access#Interactive_Access requesting GPUs if needed], and other resources for the time you want to work.
#* For example, for a session requesting 4 cores, 15GB of RAM, an A100 GPU for 1 hour: <code>'''srun -c 4 --mem 15gb -p gpu --gres gpu:a100:1 -t 60 --pty bash -i'''</code>
# For Jupyter notebooks it is important to:
#* Load Jupyter  and the module you need for the kernel you want to use, e.g. for Tensorflow: <code>module load jupyter tensorflow/2.7.0</code>
#* Export the <code>XDG_RUNTIME_DIR</code> to set temp directory (otherwise it tries to use <code>/run/user/</code>, which you can't be written to!): <code>export XDG_RUNTIME_DIR=${SLURM_TMPDIR}</code>
# Start VS Code tunnel: <code>code tunnel</code>
#*This can all be done with this line: <code>'''module load jupyter tensorflow/2.7.0; export XDG_RUNTIME_DIR=${SLURM_TMPDIR}; code tunnel'''</code>

Connect To The Tunnel

• Find the authentication information. Look at the SLURM job log file e.g. with 'less slurm-JOBID.out'. At the top you will see text similar to the following example

* Visual Studio Code Server
*
* By using the software, you agree to
* the Visual Studio Code Server License Terms (https://aka.ms/vscode-server-license) and
* the Microsoft Privacy Statement (https://privacy.microsoft.com/en-US/privacystatement).
*
[2024-02-19 16:57:03] info Using Github for authentication, run `code tunnel user login --provider <provider>` option to change this.
To grant access to the server, please log into https://github.com/login/device and use code 14AJ-AYFD
[2024-02-19 17:08:11] info Creating tunnel with the name: c0700a-s4ufhpc

Open this link in your browser https://vscode.dev/tunnel/c0700a-s4ufhpc

• Go either to the link shown in the above example

or go to https://vscode.dev/ and click on the 'Connect to tunnel' button.

• Authenticate with your GitHub account using the above code.
• Connect to the URL provided in your browser.
• Once you open VS Code in the browser, open the command palette (ctrl+shift+p) and select the appropriate Jupyter Interpreter.
• Code away!