Skip to content

Apptainer (formally known as Singularity)#

What is Apptainer?#

Apptainer is an open-source computer program that performs operating-system-level virtualization (also known as containerisation).

One of the main uses of Apptainer is to bring containers and reproducibility to scientific computing and the high-performance computing (HPC) world. Using Apptainer/Singularity containers, developers can work in reproducible environments of their choosing and design, and these complete environments can easily be copied and executed on other platforms.

For more general information about the use of Apptainer, please see the official documentation at https://apptainer.org/docs/.

This documentation only covers aspects of using Apptainer on the HPC-UGent infrastructure.

Restrictions on image location#

Some restrictions have been put in place on the use of Apptainer. This is mainly done for performance reasons and to avoid that the use of Apptainer impacts other users on the system.

The Apptainer/Singularity image file must be located on either one of the scratch filesystems, the local disk of the workernode you are using or /dev/shm. The centrally provided apptainer command will refuse to run using images that are located elsewhere, in particular on the $VSC_HOME, /apps or $VSC_DATA filesystems.

In addition, this implies that running containers images provided via a URL (e.g., shub://... or docker://...) will not work.

If these limitations are a problem for you, please let us know via hpc@ugent.be.

Available filesystems#

All HPC-UGent shared filesystems will be readily available in an Apptainer/Singularity container, including the home, data and scratch filesystems, and they will be accessible via the familiar $VSC_HOME, $VSC_DATA* and $VSC_SCRATCH* environment variables.

Apptainer/Singularity Images#

Creating Apptainer/Singularity images#

We recommend writing your Apptainer/Singularity image to a globally writable location, like /tmp, or /local directories. Once the image is created, you should move it to your desired destination. An example to make an Apptainer/Singularity container image:

# avoid that Apptainer uses $HOME/.cache
export APPTAINER_CACHEDIR=/tmp/$USER/apptainer/cache
# instruct Apptainer to use temp dir on local filesystem
export APPTAINER_TMPDIR=/tmp/$USER/apptainer/tmpdir
# specified temp dir must exist, so create it
mkdir -p $APPTAINER_TMPDIR
# convert Docker container to Apptainer container image
apptainer build /tmp/$USER/tf.sif docker://nvcr.io/nvidia/tensorflow:21.10-tf1-py3
# mv container image to $VSC_SCRATCH
mv /tmp/$USER/tf.sif $VSC_SCRATCH/tf.sif

Converting Docker images#

For more information on converting existing Docker images to Apptainer/Singularity images, see https://apptainer.org/docs/user/main/docker_and_oci.html.

We strongly recommend the use of Docker Hub, see https://hub.docker.com/ for more information.

Execute our own script within our container#

Copy testing image from /apps/gent/tutorials/Singularity to $VSC_SCRATCH:

cp /apps/gent/tutorials/Singularity/CentOS7_EasyBuild.img $VSC_SCRATCH/

Create a job script like:

#!/bin/sh

#PBS -o apptainer.output
#PBS -e apptainer.error
#PBS -l nodes=1:ppn=1
#PBS -l walltime=12:00:00


apptainer exec $VSC_SCRATCH/CentOS7_EasyBuild.img ~/my_script.sh

Create an example myscript.sh:

#!/bin/bash

# prime factors
factor 1234567

Tensorflow example#

We already have a Tensorflow example image, but you can also convert the Docker image (see https://hub.docker.com/r/tensorflow/tensorflow) to an Apptainer/Singularity image yourself

Copy testing image from /apps/gent/tutorials to $VSC_SCRATCH:

cp /apps/gent/tutorials/Singularity/Ubuntu14.04_tensorflow.img $VSC_SCRATCH/

#!/bin/sh
#
#
#PBS -o tensorflow.output
#PBS -e tensorflow.error
#PBS -l nodes=1:ppn=4
#PBS -l walltime=12:00:00
#

apptainer exec $VSC_SCRATCH/Ubuntu14.04_tensorflow.img python ~/linear_regression.py

You can download linear_regression.py from the official Tensorflow repository.

MPI example#

It is also possible to execute MPI jobs within a container, but the following requirements apply:

  • Mellanox IB libraries must be available from the container (install the infiniband-diags, libmlx5-1 and libmlx4-1 OS packages)

  • Use modules within the container (install the environment-modules or lmod package in your container)

  • Load the required module(s) before apptainer execution.

  • Set C_INCLUDE_PATH variable in your container if it is required during compilation time (export C_INCLUDE_PATH=/usr/include/x86_64-linux-gnu/:$C_INCLUDE_PATH for Debian flavours)

Copy the testing image from /apps/gent/tutorials/Singularity to $VSC_SCRATCH

cp /apps/gent/tutorials/Singularity/Debian8_UGentMPI.img $VSC_SCRATCH/

For example to compile an MPI example:

module load intel
apptainer shell $VSC_SCRATCH/Debian8_UGentMPI.img
export LANG=C
export C_INCLUDE_PATH=/usr/include/x86_64-linux-gnu/:$C_INCLUDE_PATH
mpiicc ompi/examples/ring_c.c -o ring_debian
exit

Example MPI job script:

#!/bin/sh

#PBS -N mpi
#PBS -o apptainermpi.output
#PBS -e apptainermpi.error
#PBS -l nodes=2:ppn=15
#PBS -l walltime=12:00:00

module load intel vsc-mympirun
mympirun --impi-fallback apptainer exec $VSC_SCRATCH/Debian8_UGentMPI.img ~/ring_debian