Delete CFS data

how-to · Reviewed 2026-07-31 · FEAT-0120 Public Beta

CFS has no trash. Deletion is immediate and permanent. This page covers both ways it happens.

The two ways data leaves CFS

Trigger	You initiated it	Recoverable
Deleting from a live workstation or job	Yes	No
Archiving a job or workstation	Indirectly	No

Delete manually

Mount CFS in a workstation and delete through the file explorer, or use rm from a terminal in a running job or workstation.

There is no undo

CFS deletion cannot be reversed. Confirm before you remove anything.

CFS backups don't protect against deletion

If your workspace has optional backups enabled, don't read them as an undo button.

The backup is a mirror, not a snapshot

CFS backups mirror the filesystem rather than capturing point-in-time snapshots. A file you create, change, or delete is created, changed, or deleted in the backup too.

That makes backups useful against infrastructure failure, and useless against someone deleting the wrong directory. There's no point in time to roll back to.

The only real protection against accidental deletion is a copy somewhere else: Rescale Files, or off-platform.

Archival deletes CFS data automatically

Archiving permanently deletes CFS files, and it's on by default

Since September 2025, archiving a job or workstation automatically and permanently deletes its associated files from both Rescale Files and CFS. This is enabled by default: not something you opt into.

Deleted files cannot be recovered even if you later unarchive the job.

What archival removes:

• The job folder and all its files from CFS, including all output files.
• The job folder from any project directory it was shared under — so teammates lose access at the same time.

The behavior is identical across interfaces:

Interface	Behavior
UI	Visual confirmation of deletion in the archival drawer
CLI	delete job does the same, with no warning
API	Applies to single and batch deletions

The CLI gives you no warning

Scripted cleanup that archives jobs will silently destroy CFS data. If you automate archival, download or relocate anything worth keeping in the same script: before the archive call.

Data belonging to deactivated users

Deactivating a user does not delete their CFS data, and the two kinds of data they leave behind behave differently:

Their data	What happens
Job data	Stays until the job is archived. Transfer job ownership so the new owner can archive it
Personal library data	Stays, with no cleanup or ownership-transfer path today

Library data left by departed users is a known way for a filesystem to fill up with data nobody can claim. Two things help: have people keep shared reference data in project folders rather than personal libraries, and check the usage dashboard for names you don't recognize.

Before you archive anything on CFS

1. Download what you need, or copy it to Rescale Files or a project library.
2. Check whether a project depends on it: see organize and share.
3. Confirm with teammates if the data is shared.

Note the asymmetry: on Rescale Files, archiving sends files to the trash and you typically have about a day to change your mind. On CFS you have none.

Related

• Archive and delete jobs — the job-level workflow
• The file lifecycle — how Rescale Files differs
• Monitor CFS usage — spotting the need to clean up
• Organize and share — what else archival affects