StrongLink-to-Versity

file version: 9 Jul 2026

Overview

This page is meant for users who wish to quickly update their archival/retrieval scripts so that they work with the new HSM software by Versity. Getting Started provides a broader introduction.

slk is replaced by atlas which is the ATLAS command line interface. Running only atlas starts an interactive shell whereas running atlas <command> just executes the command without entering the interactive shell. The latter behaves similar to slk.

Additionally, Versity is accessible via the ATLAS web interface which has the same features as the atlas CLI: https://tapearchive.dkrz.de (available from calendar week 29 onwards).

Important changes and new features are:

• all paths have the prefix /hsm: /hsm/arch, /hsm/double, …

• automatic packing of small files (invisible to users); but, 60 000 files of 1 MB each should still be packed in advance

• ATLAS can read directly from /work and /scratch (archival). It can write directly to /scratch (retrieval)

• async background data transfer via ATLAS: no need for users to run a long SLURM job for archivals or retrievals; each transfer has its own job id; please check job status: atlas jobs get <job id>

• integrity check of data transfers (copy and move) via sha1 checksums by default

• very efficient internal organization of the tape access: no need to sort retrieval requests by tape

• Python library: currently under development

Replacing slk commands

If you need to replace slk commands other than archive and retrieve, please have a look into the Getting Started.

Load module

slk:

$ module load slk

atlas:

$ module load tape

Login

slk:

# using your default DKRZ user credentials
$ slk login

atlas:

# using your default DKRZ user credentials
$ atlas login

mount point:

no login required

Archive one file

case:

• source: /work/ab1234/c567890/file.txt

• destination, StrongLink: /arch/ab1234/c567890/

• destination, Versity: /hsm/arch/ab1234/c567890/

slk:

# transfer file
$ slk archive /work/ab1234/c567890/file.txt /arch/ab1234/c567890/

atlas:

# create destination folder
$ atlas mkdir -p /hsm/arch/ab1234/c567890
# transfer file
$ atlas cp /work/ab1234/c567890/file.txt /hsm/arch/ab1234/c567890/
copy job submitted: 95
# get job status
$ atlas jobs get 95
...


# ALTERNATIVE transfer file => set destination filename
# e.g. renaming possible
$ atlas cp /work/ab1234/c567890/file.txt /hsm/arch/ab1234/c567890/renamed_file.txt
copy job submitted: 96
# get job status
$ atlas jobs get 96
...

Note: If `` /hsm/arch/ab1234`` exists but `` /hsm/arch/ab1234/c567890`` not, the first command will create a file c567890 in `` /hsm/arch/ab1234`` with the content of file.txt.

mount point:

not possible

Archive folder recursively

case:

• source: /work/ab1234/c567890/output

• destination, StrongLink: /arch/ab1234/c567890/output

• destination, Versity: /hsm/arch/ab1234/c567890/output

slk:

# transfer files
$ slk archive -R /work/ab1234/c567890/output/ /arch/ab1234/c567890/output/

atlas:

# create destination folder
$ atlas mkdir -p  /hsm/arch/ab1234/c567890/output
# transfer files
$ atlas cp -R /work/ab1234/c567890/output/  /hsm/arch/ab1234/c567890/output/
copy job submitted: 97
# get job status
$ atlas jobs get 97
...

mount point:

not possible

Retrieve file

case:

• source, StrongLink: /arch/ab1234/c567890/file.txt

• source, Versity: `` /hsm/arch/ab1234/c567890/file.txt``

• destination: /scratch/c/c567890/parent/child

slk:

# transfer file
$ slk retrieve /arch/ab1234/c567890/file.txt /scratch/c/c567890/parent/child/

atlas:

# create destination folder
$ mkdir -p /scratch/c/c567890/parent/child
# stage file (tape to cache)
$ atlas stage /hsm/arch/ab1234/c567890/file.txt
# transfer file
$ atlas cp  /hsm/arch/ab1234/c567890/file.txt /scratch/c/c567890/parent/child/
copy job submitted: 98
# get job status
$ atlas jobs get 98
...

Note: If only the second command was run and /scratch/c/c567890/parent existed but /scratch/c/c567890/parent/child not, then the first command would create a file child in /scratch/c/c567890/parent with the content of file.txt.

mount point:

# run staging command beforehand to be sure that file is actually in the cache
$ atlas stage  /hsm/arch/ab1234/c567890/file.txt
$ rsync -av  /hsm/arch/ab1234/c567890/file.txt /scratch/c/c567890/parent/child/
# will exit with an error if file.txt is not cached

Note 1: The mount point is meant for easy access to individual files - e.g. to check the header of a netCDF file or to peek into a dataset. It is not meant for the transfer of large amounts of data. Please use cp of ATLAS for this.

Note 2: We recommand to always run atlas stage prior to accessing files via the mount point. If a cache cleanup process starts before the file was accessed, it might be removed from the cache causing the transfer to fail. Running atlas stage is similar to a touch command with respect to the time stamp relevant for the cache cleanup.

Retrieve folder recursively

case:

• source: /arch/ab1234/c567890/output

• destination: /scratch/c/c567890/output

slk:

# run slk_helpers watcher setup for recall + retrieval:
... watchers ...

# OR retrieve all files at once (not recommanded, when files are not cached)
$ slk retrieve -R /arch/ab1234/c567890/output/ /scratch/c/c567890/output

atlas:

# create destination folder
$ mkdir -p /scratch/c/c567890/output
# stage file (tape to cache)
$ atlas stage /hsm/arch/ab1234/c567890/output
# transfer files
$ atlas cp --recursive  /hsm/arch/ab1234/c567890/output/* /scratch/c/c567890/output/
copy job submitted: 99
# get job status
$ atlas jobs get 99
...

mount point:

Please use ATLAS for this purpose

Replacing pyslk

Work in progress

Replacing fsspec + slkspec

Work in progress

Details on atlas

Location of token

• slk: ~/.slk/config.json

• Versity: ~/.config/atlas/session.json

Commands

** StrongLink**	Versity	Meaning
slk login	atlas login	generate a login token
slk archive	atlas cp/ atlas upload	file transfer to HSM [1,2]
slk retrieve	atlas cp / atlas download	file transfer from HSM [1,3]
slk recall	atlas stage	transfer file from tape to cache
slk move	atlas mv	move files within HSM
slk rename	atlas mv	rename files in HSM
slk delete	atlas rm	delete file
slk mkdir [-R]	atlas mkdir [-p]	create directory
slk chmod	atlas chmod	change permissions
slk search	atlas search	run search for files

[1] atlas cp transfers data in the background between tape archive and Lustre filesystem

[2] atlas upload transfer data from current system/node directly to the tape archive

[3] atlas download transfer data from tape archive directly to the current system/node

Warning

atlas mv performs a real move between filesystems. After the transfer completes successfully, it attempts to remove the source data.

Be careful when using atlas mv to archive data to HSM.

• atlas mv /work/b123456/mydata.nc /hsm/archive/b123456/mydata.nc copies the file to HSM, but the original file remains in /work because the filesystem is mounted read-only and cannot be modified.

• atlas mv /scratch/b123456/mydata.nc /hsm/archive/b123456/mydata.nc copies the file to HSM and removes the original from /scratch, because /scratch is writable.

If you want to keep the source data, use atlas cp instead of atlas mv.

Glossary

• online: a file is cached

• offline: a file is not cached

StrongLink	Versity	Meaning
recall	staging	transfer files from tape to cache
	release	allow removal of file from cache
		if it is already on tape