jupyter

jupyter lab

Tools for working with jupyter lab on eagle

There are a couple of bash scripts that can be used for launching a jupyter lab instance on an eagle compute node and then connecting via your local web browser.

One example workflow is to use the sbatch_jupyter.sh, auto_launch_jupyter.sh scripts.

Create your conda environment:

conda create -n jupyterenv python=3.8 jupyterlab

Update the sbatch_jupyter.sh script with your account information and desired run time:

#!/bin/bash  --login

#SBATCH --account=<project handle>
#SBATCH --time=01:00:00

#SBATCH --nodes=1
#SBATCH --tasks-per-node=1
#SBATCH --cpus-per-task=36

Then, you can submit and monitor the job automatically with:

bash auto_launch_jupyter.sh

This will submit the sbatch_jupyter.sh script to eagle and then monitor the status, eventually providing you will the commands you need to connect. The output should look something like:

[<hpc-username>@el2 bash]$ bash auto_launch_jupyter.sh
Checking job status..
job still pending..
job is running!
getting jupyter information, hang tight..
okay, now run the follwing on your local machine:
ssh -L 7878:r1i6n1:7878 el2.hpc.nrel.gov
then, navigate to the following on your local browser:
http://127.0.0.1:7878/?token=ed23bab2d938f52122c8cccf833dbee185d62adbfe56d02b