Rockfish Quick Start

What you need

Item

Notes

JHED and Rockfish account

Request one on the Coldfront Portal

Hopkins VPN (Pulse Secure)

Required from off-campus for some services

SSH client

macOS/Linux: built-in • Windows: OpenSSH or PuTTY

Create an Account

Log in

ssh <YourUserID>@login.rockfish.jhu.edu

First time? Type yes at the host-key prompt, then your Rockfish password.

Load software modules

Rockfish uses Lmod modules.

module avail          # list all software
module spider R       # list every R version
module load R/4.4.1   # make it your default R for this session

Tip: add frequently-used modules to ~/.bashrc with module load ….

Quick interactive work

Short, exploratory commands belong on an interactive compute node.

interact -n 4 -t 2:00:00        # 4 tasks, 2 h in the *express* partition
hostname                        # now you're on compute-XYZ, not a login node
python                          # fire up IPython, R, etc.

Submit your first batch job

Create hello.slurm :

#!/bin/bash
#SBATCH --job-name=hello
#SBATCH --output=hello.%j.out
#SBATCH --time=00:02:00
#SBATCH --ntasks=1
#SBATCH --cpus-per-task=1
#SBATCH --partition=express

echo "Hello from $(hostname)!"
sleep 30

Submit & monitor:

sbatch hello.slurm      # submit
squeue -u $USER         # check status
tail -f hello.<jobID>.out

If the job is PENDING for more than a minute, reduce resources or pick another partition (see table below).

Partitions at a glance

For available partitions, see: Available Partitions

Managing Python, Jupyter, RStudio

Tool / service

Where to start

Docs

JupyterLab

Portal → Jupyter Server (GUI)\ jupyterlab.sh helper (CLI)

JupyterLab on Rockfish

RStudio Server

Portal → RStudio Server (GUI)\ r-studio-server.sh helper (CLI)

R & RStudio

Conda environments

module load anaconda\ conda create -n myenv python=3.11

Environments & Packages

Storage locations

Path

Default quota

Intended for

/home/$USER

50 GB (backed up)

configs, notebooks, small scripts

/scratch4/$PI

1 TB

small files, working data

/scratch16/$PI

By request

large files, working data

/data/$PI

10 TB

long-term, high-value data

For more information on available filesystems, see here: Filesystems on Rockfish

House-keeping

  • Purge policy – anything in scratch > 30 days old is deleted. Move results you wish to keep to /data or download them.

  • Fair-share scheduler – large jobs may wait if your lab has used more CPU-hours than average recently.

  • Login nodes – no heavy compute. Use interact or sbatch instead.

Need help?