Quick Start
This guide is intended for running Singularity on a computer where you have root (administrative) privileges, and will install Singularity from source code. Other installation options, including building an RPM package and installing Singularity without root privileges are discussed in the installation section of the admin guide.
If you need to request an installation on your shared resource, see the requesting an installation section for information to send to your system administrator.
For any additional help or support contact the Singularity Community: https://singularity.hpcng.org/help
Quick Installation Steps
You will need a Linux system to run Singularity natively. Options for using Singularity on Mac and Windows machines, along with alternate Linux installation options are discussed in the installation section of the admin guide.
Install system dependencies
You must first install development libraries to your host. Assuming Ubuntu (apply similar to RHEL derivatives):
$ sudo apt-get update && sudo apt-get install -y \
build-essential \
libssl-dev \
uuid-dev \
libgpgme11-dev \
squashfs-tools \
libseccomp-dev \
wget \
pkg-config \
git \
cryptsetup
Note
Note that squashfs-tools is only a dependency for commands that build
images. The build command obviously relies on squashfs-tools, but
other commands may do so as well if they are ran using container images
from Docker Hub for instance.
There are 3 broad steps to installing Singularity:
Install Go
Singularity v3 and above is written primarily in Go, so you will need Go installed to compile it from source.
This is one of several ways to install and configure Go.
Note
If you have previously installed Go from a download, rather than an
operating system package, you should remove your go directory,
e.g. rm -r /usr/local/go before installing a newer version.
Extracting a new version of Go over an existing installation can
lead to errors when building Go programs, as it may leave old
files, which have been removed or replaced in newer versions.
Visit the Go Downloads page and pick a package
archive suitable to the environment you are in. Once the Download is complete,
extract the archive to /usr/local (or use other instructions on go installation
page). Alternatively, follow the commands here:
$ export VERSION=1.17.2 OS=linux ARCH=amd64 && \ # Replace the values as needed
wget https://dl.google.com/go/go$VERSION.$OS-$ARCH.tar.gz && \ # Downloads the required Go package
sudo tar -C /usr/local -xzvf go$VERSION.$OS-$ARCH.tar.gz && \ # Extracts the archive
rm go$VERSION.$OS-$ARCH.tar.gz # Deletes the ``tar`` file
Set the Environment variable PATH to point to Go:
$ echo 'export PATH=/usr/local/go/bin:$PATH' >> ~/.bashrc && \
source ~/.bashrc
Download Singularity from a release
You can download Singularity from one of the releases. To see a full list, visit the GitHub release page. After deciding on a release to install, you can run the following commands to proceed with the installation.
$ export VERSION=3.8.5 && # adjust this as necessary \
wget https://github.com/hpcng/singularity/releases/download/v${VERSION}/singularity-${VERSION}.tar.gz && \
tar -xzf singularity-${VERSION}.tar.gz && \
cd singularity-${VERSION}
Compile the Singularity source code
Now you are ready to build Singularity. Dependencies will be automatically downloaded. You can build Singularity using the following commands:
$ ./mconfig && \
make -C builddir && \
sudo make -C builddir install
Singularity must be installed as root to function properly.
Overview of the Singularity Interface
Singularity’s command line interface allows you to build and interact with containers transparently. You can run programs inside a container as if they were running on your host system. You can easily redirect IO, use pipes, pass arguments, and access files, sockets, and ports on the host system from within a container.
The help command gives an overview of Singularity options and subcommands as
follows:
$ singularity help
Linux container platform optimized for High Performance Computing (HPC) and
Enterprise Performance Computing (EPC)
Usage:
singularity [global options...]
Description:
Singularity containers provide an application virtualization layer enabling
mobility of compute via both application and environment portability. With
Singularity one is capable of building a root file system that runs on any
other Linux system where Singularity is installed.
Options:
-d, --debug print debugging information (highest verbosity)
-h, --help help for singularity
--nocolor print without color output (default False)
-q, --quiet suppress normal output
-s, --silent only print errors
-v, --verbose print additional information
Available Commands:
build Build a Singularity image
cache Manage the local cache
capability Manage Linux capabilities for users and groups
exec Run a command within a container
help Help about any command
inspect Show metadata for an image
instance Manage containers running as services
key Manage OpenPGP keys
oci Manage OCI containers
plugin Manage singularity plugins
pull Pull an image from a URI
push Upload image to the provided library (default is "cloud.sylabs.io")
remote Manage singularity remote endpoints
run Run the user-defined default command within a container
run-help Show the user-defined help for an image
search Search a Container Library for images
shell Run a shell within a container
sif siftool is a program for Singularity Image Format (SIF) file manipulation
sign Attach a cryptographic signature to an image
test Run the user-defined tests within a container
verify Verify cryptographic signatures attached to an image
version Show the version for Singularity
Examples:
$ singularity help <command> [<subcommand>]
$ singularity help build
$ singularity help instance start
For additional help or support, please visit https://www.sylabs.io/docs/
Information about subcommand can also be viewed with the help command.
$ singularity help verify
Verify cryptographic signatures attached to an image
Usage:
singularity verify [verify options...] <image path>
Description:
The verify command allows a user to verify cryptographic signatures on SIF
container files. There may be multiple signatures for data objects and
multiple data objects signed. By default the command searches for the primary
partition signature. If found, a list of all verification blocks applied on
the primary partition is gathered so that data integrity (hashing) and
signature verification is done for all those blocks.
Options:
-a, --all verify all objects
-g, --group-id uint32 verify objects with the specified group ID
-h, --help help for verify
-j, --json output json
--legacy-insecure enable verification of (insecure) legacy signatures
-l, --local only verify with local keys
-i, --sif-id uint32 verify object with the specified ID
-u, --url string key server URL (default "https://keys.sylabs.io")
Examples:
$ singularity verify container.sif
For additional help or support, please visit https://www.sylabs.io/docs/
Singularity uses positional syntax (i.e. the order of commands and options
matters). Global options affecting the behavior of all commands follow the main
singularity command. Then sub commands are followed by their options
and arguments.
For example, to pass the --debug option to the main singularity command
and run Singularity with debugging messages on:
$ singularity --debug run library://lolcow
To pass the --containall option to the run command and run a
Singularity image in an isolated manner:
$ singularity run --containall library://lolcow
Singularity 2.4 introduced the concept of command groups. For instance, to list
Linux capabilities for a particular user, you would use the list command in
the capability command group like so:
$ singularity capability list dave
Container authors might also write help docs specific to a container or for an
internal module called an app. If those help docs exist for a particular
container, you can view them like so.
$ singularity inspect --helpfile container.sif # See the container's help, if provided
$ singularity inspect --helpfile --app=foo foo.sif # See the help for foo, if provided
Download pre-built images
You can use the search command to locate groups, collections, and
containers of interest on the Container Library .
singularity search tensorflow
Found 22 container images for amd64 matching "tensorflow":
library://ajgreen/default/tensorflow2-gpu-py3-r-jupyter:latest
Current software: tensorflow2; py3.7; r; jupyterlab1.2.6
Signed by: 1B8565093D80FA393BC2BD73EA4711C01D881FCB
library://bensonyang/collection/tensorflow-rdma_v4.sif:latest
library://dxtr/default/hpc-tensorflow:0.1
library://emmeff/tensorflow/tensorflow:latest
library://husi253/default/tensorflow:20.01-tf1-py3-mrcnn-2020.10.07
library://husi253/default/tensorflow:20.01-tf1-py3-mrcnn-20201014
library://husi253/default/tensorflow:20.01-tf2-py3-lhx-20201007
library://irinaespejo/default/tensorflow-gan:sha256.0c1b6026ba2d6989242f418835d76cd02fc4cfc8115682986395a71ef015af18
library://jon/default/tensorflow:1.12-gpu
Signed by: D0E30822F7F4B229B1454388597B8AFA69C8EE9F
...
You can use the pull and build commands to download pre-built images from an external resource like the Container Library or Docker Hub.
When called on a native Singularity image like those provided on the Container Library, pull
simply downloads the image file to your system.
$ singularity pull library://lolcow
You can also use pull with the docker:// uri to reference Docker images
served from a registry. In this case pull does not just download an image
file. Docker images are stored in layers, so pull must also combine those
layers into a usable Singularity file.
$ singularity pull docker://sylabsio/lolcow
Pulling Docker images reduces reproducibility. If you were to pull a Docker image today and then wait six months and pull again, you are not guaranteed to get the same image. If any of the source layers has changed the image will be altered. If reproducibility is a priority for you, try building your images from the Container Library.
You can also use the build command to download pre-built images from an
external resource. When using build you must specify a name for your
container like so:
$ singularity build ubuntu.sif library://ubuntu
$ singularity build lolcow.sif docker://sylabsio/lolcow
Unlike pull, build will convert your image to the latest Singularity
image format after downloading it.
build is like a “Swiss Army knife” for container creation. In addition to
downloading images, you can use build to create images from other images or
from scratch using a definition file. You can also
use build to convert an image between the container formats supported by
Singularity. To see a comparison of Singularity definition file with Dockerfile,
please see: this section.
Interact with images
You can interact with images in several ways, each of which can accept image URIs in addition to a local image path.
For demonstration, we will use a lolcow_latest.sif image that can be pulled
from the Container Library:
$ singularity pull library://lolcow
Shell
The shell command allows you to spawn a new shell within your container and interact with it as though it were a small virtual machine.
$ singularity shell lolcow_latest.sif
Singularity lolcow_latest.sif:~>
The change in prompt indicates that you have entered the container (though you should not rely on that to determine whether you are in container or not).
Once inside of a Singularity container, you are the same user as you are on the host system.
Singularity lolcow_latest.sif:~> whoami
david
Singularity lolcow_latest.sif:~> id
uid=1000(david) gid=1000(david) groups=1000(david),4(adm),24(cdrom),27(sudo),30(dip),46(plugdev),116(lpadmin),126(sambashare)
shell also works with the library://, docker://, and shub://
URIs. This creates an ephemeral container that disappears when the shell is
exited.
$ singularity shell library://lolcow
Executing Commands
The exec
command allows you to execute a custom command within a container by specifying
the image file. For instance, to execute the cowsay program within the
lolcow_latest.sif container:
$ singularity exec lolcow_latest.sif cowsay moo
_____
< moo >
-----
\ ^__^
\ (oo)\_______
(__)\ )\/\
||----w |
|| ||
exec also works with the library://, docker://, and shub://
URIs. This creates an ephemeral container that executes a command and
disappears.
$ singularity exec library://lolcow cowsay "Fresh from the library!"
_________________________
< Fresh from the library! >
-------------------------
\ ^__^
\ (oo)\_______
(__)\ )\/\
||----w |
|| ||
Running a container
Singularity containers contain runscripts. These are user defined scripts that define the actions a container should perform when someone runs it. The runscript can be triggered with the run command, or simply by calling the container as though it were an executable.
$ singularity run lolcow_latest.sif
______________________________
< Mon Aug 16 13:01:55 CDT 2021 >
------------------------------
\ ^__^
\ (oo)\_______
(__)\ )\/\
||----w |
|| ||
$ ./lolcow_latest.sif
______________________________
< Mon Aug 16 13:12:50 CDT 2021 >
------------------------------
\ ^__^
\ (oo)\_______
(__)\ )\/\
||----w |
|| ||
run also works with the library://, docker://, and shub:// URIs.
This creates an ephemeral container that runs and then disappears.
$ singularity run library://lolcow
______________________________
< Mon Aug 16 13:12:33 CDT 2021 >
------------------------------
\ ^__^
\ (oo)\_______
(__)\ )\/\
||----w |
|| ||
Working with Files
Files on the host are reachable from within the container.
$ echo "Hello from inside the container" > $HOME/hostfile.txt
$ singularity exec lolcow_latest.sif cat $HOME/hostfile.txt
Hello from inside the container
This example works because hostfile.txt exists in the user’s home directory.
By default Singularity bind mounts /home/$USER, /tmp, and $PWD into
your container at runtime.
You can specify additional directories to bind mount into your container with
the --bind option. In this example, the data directory on the host
system is bind mounted to the /mnt directory inside the container.
$ echo "Drink milk (and never eat hamburgers)." > /data/cow_advice.txt
$ singularity exec --bind /data:/mnt lolcow_latest.sif cat /mnt/cow_advice.txt
Drink milk (and never eat hamburgers).
Pipes and redirects also work with Singularity commands just like they do with normal Linux commands.
$ cat /data/cow_advice.txt | singularity exec lolcow_latest.sif cowsay
________________________________________
< Drink milk (and never eat hamburgers). >
----------------------------------------
\ ^__^
\ (oo)\_______
(__)\ )\/\
||----w |
|| ||
Build images from scratch
Singularity v3.0 and above produces immutable images in the Singularity Image File (SIF) format. This ensures reproducible and verifiable images and allows for many extra benefits such as the ability to sign and verify your containers.
However, during testing and debugging you may want an image format that is
writable. This way you can shell into the image and install software and
dependencies until you are satisfied that your container will fulfill your
needs. For these scenarios, Singularity also supports the sandbox format
(which is really just a directory).
Sandbox Directories
To build into a sandbox (container in a directory) use the
build --sandbox command and option:
$ sudo singularity build --sandbox ubuntu/ library://ubuntu
This command creates a directory called ubuntu/ with an entire Ubuntu
Operating System and some Singularity metadata in your current working
directory.
You can use commands like shell, exec , and run with this directory
just as you would with a Singularity image. If you pass the --writable
option when you use your container you can also write files within the sandbox
directory (provided you have the permissions to do so).
$ sudo singularity exec --writable ubuntu touch /foo
$ singularity exec ubuntu/ ls /foo
/foo
Converting images from one format to another
The build command allows you to build a container from an existing
container. This means that you can use it to convert a container from one format
to another. For instance, if you have already created a sandbox (directory) and
want to convert it to the default immutable image format (squashfs) you can do
so:
$ singularity build new-sif sandbox
Doing so may break reproducibility if you have altered your sandbox outside of the context of a definition file, so you are advised to exercise care.
Singularity Definition Files
For a reproducible, verifiable and production-quality container you should build a SIF file using a Singularity definition file. This also makes it easy to add files, environment variables, and install custom software, and still start from your base of choice (e.g., the Container Library).
A definition file has a header and a body. The header determines the base container to begin with, and the body is further divided into sections that perform things like software installation, environment setup, and copying files into the container from host system, etc.
Here is an example of a definition file:
BootStrap: library
From: ubuntu:16.04
%post
apt-get -y update
apt-get -y install date cowsay lolcat
%environment
export LC_ALL=C
export PATH=/usr/games:$PATH
%runscript
date | cowsay | lolcat
%labels
Author Sylabs
To build a container from this definition file (assuming it is a file named lolcow.def), you would call build like so:
$ sudo singularity build lolcow.sif lolcow.def
In this example, the header tells Singularity to use a base Ubuntu 16.04 image from the Container Library.
The
%postsection executes within the container at build time after the base OS has been installed. The%postsection is therefore the place to perform installations of new applications.The
%environmentsection defines some environment variables that will be available to the container at runtime.The
%runscriptsection defines actions for the container to take when it is executed.And finally, the
%labelssection allows for custom metadata to be added to the container.
This is a very small example of the things that you can do with a definition file. In addition to building a container from the Container Library, you can start with base images from Docker Hub and use images directly from official repositories such as Ubuntu, Debian, CentOS, Arch, and BusyBox. You can also use an existing container on your host system as a base.
If you want to build Singularity images but you don’t have administrative (root) access on your build system, you can build images using the Remote Builder.
This quickstart document just scratches the surface of all of the things you can do with Singularity!
If you need additional help or support, see https://singularity.hpcng.org/help.