Running Jupyterlab on Poseidon - USGS-CMG/usgs-cmg-portal GitHub Wiki
- Login to Denali
ssh denali.whoi.edu
- Get off the master node:
srun -p compute --ntasks-per-node=1 --mem=8G --time=1:00:00 --pty bash
- Install Miniconda:
wget https://repo.continuum.io/miniconda/Miniconda3-latest-Linux-x86_64.sh bash Miniconda3-latest-Linux-x86_64.sh -b export PATH=$HOME/miniconda3/bin:$PATH conda activate base
- Use conda-forge channel for all packages:
conda config --add channels conda-forge --force conda config --remove channels defaults --force
- Install Mamba into base environment (conda equivalent written in C++, much faster at solving package requirements) and update base packages:
conda install mamba mamba update --all
- Install packages required to jupyterlab into base environment
mamba install jupyterlab jupyter-server-proxy dask-labextension ipywidgets nb_conda_kernels nodejs
- create password to login to jupyterlab
jupyter server --generate-config jupyter server password
- Create a custom science environment including ipykernel (include extra packages here):
mamba create -n pangeo python=3.8 xarray h5netcdf dask dask-jobqueue ipywidgets hvplot geoviews datashader ipykernel \ pyepsg rioxarray metpy zarr rechunker intake intake-xarray
When you want to use Jupyterlab on Poseidon:
- Login to Denali
ssh denali.whoi.edu
- Get off the master node:
srun -p compute --ntasks-per-node=4 --mem=32G --time=3:00:00 --pty bash
- Start jupyter with a
start_jupyter
scrip something like this:conda activate base cd /vortexfs1/usgs/rsignell JPORT=$(shuf -i 8400-9400 -n 1) echo "ssh -N -L 8889:`hostname`:$JPORT [email protected]" jupyter lab --no-browser --ip=`hostname` --port=$JPORT
- Copy the printed out line starting with
ssh -N -L 8889:
to your clipboard and paste into a new terminal - Open a local browser to
http://localhost:8889
, and login to jupyterlab using the password you created above
Continue with the instructions here: https://pangeo.io/setup_guides/hpc.html#launch-dask-with-dask-jobqueue