Data Recovery

Deleted the wrong files? SeaweedFS Enterprise Data Recovery lets you find and restore deleted objects and files straight from the Admin UI, within a retention window you choose — no snapshot restore and no separate backup tool. Turn on a window, and recent deletes stay recoverable.

A file is deleted report.parquet Its bytes are retained report.parquet kept for the retention window Restored in one click report.parquet back at its original path
Deleted bytes stay on disk for the retention window, so a restore just puts the file back — no backup restore, no snapshot rollback.

It complements (doesn’t replace) backups, replication, and Object Lock — it’s the fast, in-place fix for the most common cause of data loss: deletion.

When to use it

  • An AI agent or script deletes the wrong files — recover the affected paths without rolling back unrelated work.
  • A bad rm, lifecycle rule, or sync removes the wrong prefix — restore one file, a batch, or a filtered set.
  • One user or key deleted across many paths — filter by owner and restore just that account’s deletions, without touching anyone else’s work.
  • A malicious actor or compromised key sweeps data — if caught inside the window, inspect the blast radius and restore.
  • An S3 delete marker hides an object — clear it in one click so the previous version returns.
  • A specific object version was permanently deleted — recover it from retained bytes (bucket versioning alone can’t).

How to use it

1. Turn on a retention window on the master — how long deletes stay recoverable:

weed server -master.deletionRetention=72h

The default 0 disables it. The free 25 TB trial caps the window at 1 hour; a full license unlocks longer windows.

The retention window decides what is recoverable retention window (e.g. 72h) deleted before the window gone for good deleted inside the window recoverable now
One flag sets how far back you can reach: deletes inside the window are recoverable, older ones are gone.

2. Recover from the Admin UI:

  1. Open the Admin UI → Recovery → Data Recovery.
  2. Click Scan, then filter by path prefix, glob, owner (user), or time range.
  3. Restore one file with the row action, or select many and Batch Restore.
  4. Pick where it lands and what happens if the target exists: keep, overwrite, or rename.
The Admin UI Data Recovery page: retention window active, scan complete, and recoverable deletions listed with per-row restore buttons.
The Data Recovery page after a scan: filter by prefix, glob, owner, or time, then restore per row or in batch.

Restores reuse the data already retained on disk, so there’s nothing to roll back.

Benefits

  • Minutes, not snapshots — recover directly from the Admin UI instead of restoring a backup.
  • Scoped — restore only the affected files, by path or by the user who deleted them — not the whole cluster.
  • S3-aware — clears delete markers and recovers permanently-deleted object versions.
  • No new moving parts — uses retained data and the existing metadata log; nothing extra to deploy.

Want the internals — how retention, the metadata log, and restore work, plus the full HTTP API and limits? See the Data Recovery technical reference.