NEMO2/Containers/Apptainer

Apptainer (formerly Singularity) is a container runtime designed for HPC systems. It is installed system-wide on NEMO and available without loading a module.

Key properties

• Runs as the invoking user — no root required and no privilege escalation
• Uses SIF (Singularity Image Format), a single portable file
• Can convert Docker images to SIF at build time
• No native Slurm plugin — call apptainer exec/run inside your batch script

Image sources

Source	Example
Apptainer Hub	apptainer pull hello-world.sif library://lolcow
Docker Hub (converted)	apptainer pull ubuntu.sif docker://ubuntu:24.04
Local .sif file	copy to $HOME or a workspace and use directly

Image storage

SIF files are plain files — store them wherever you like. To avoid filling your home quota, keep images in a workspace and use a symlink so your scripts don't need to change:

# create a workspace (100 days)
ws_allocate apptainer 100

# create an images directory in the workspace and symlink it from home
ln -s $(ws_find apptainer) ~/images

Pull images directly into the workspace:

apptainer pull ~/images/ubuntu.sif docker://ubuntu:24.04
apptainer pull ~/images/pytorch.sif docker://nvcr.io/nvidia/pytorch:24.01-py3

In batch scripts, reference images via the symlink:

IMAGE=$HOME/images/pytorch.sif

Basic commands

Pull / build an image

# from Docker Hub → produces ubuntu.sif in ~/images/
apptainer pull ~/images/ubuntu.sif docker://ubuntu:24.04

# from NVIDIA NGC
apptainer pull ~/images/pytorch.sif docker://nvcr.io/nvidia/pytorch:24.01-py3

Run a command inside a container

# exec: run a specific command
apptainer exec ubuntu.sif python3 script.py

# run: execute the container's default runscript
apptainer run ubuntu.sif

# shell: interactive shell
apptainer shell ubuntu.sif

GPU access

Pass the --nv flag to enable NVIDIA GPU passthrough:

apptainer exec --nv pytorch.sif python3 train.py

Bind-mount paths

/home and /work are available inside the container by default (Apptainer automatically binds a set of paths; check apptainer config for the system defaults).

To mount additional directories explicitly:

apptainer exec --bind /work/classic/myWs:/data ubuntu.sif bash

Using Apptainer in a Slurm batch job

Unlike Enroot/Pyxis there is no Slurm plugin; simply call apptainer from your batch script:

CPU job:

#!/bin/bash
#SBATCH -p cpu
#SBATCH --ntasks=1
#SBATCH --time=01:00:00

IMAGE=$HOME/images/ubuntu.sif

apptainer exec "$IMAGE" python3 /work/classic/myWs/train.py

GPU job (use --nv to pass through the allocated GPUs):

#!/bin/bash
#SBATCH -p l40s
#SBATCH --gres=gpu:1
#SBATCH --ntasks=1
#SBATCH --time=01:00:00

IMAGE=$HOME/images/pytorch.sif

apptainer exec --nv "$IMAGE" python3 /work/classic/myWs/train.py

Building images

Building a new SIF image from a definition file requires root (or a system that supports fakeroot/user namespaces). On NEMO you cannot build images directly on the compute nodes.

Recommended workflow:

1. Build on a machine where you have root (local workstation, VM, GitHub Actions, …)
2. Transfer the .sif file to NEMO (e.g. scp, rsync)
3. Run on NEMO as usual

Store images in a workspace (symlinked to ~/images/ as described above) — not in $TMPDIR, which is job-local and deleted after the job.

Tips

• Use apptainer exec --writable-tmpfs if the container tries to write to its own filesystem (without persisting changes).
• The --containall flag gives a fully isolated environment (no automatic bind-mounts); useful for reproducibility tests.
• To check what Apptainer version is installed: apptainer --version