File Archive & tiering

The File Archive plugin adds post-backup copy and archive jobs to NGBackup: after a backup completes, a copy job selects files by policy and moves or copies them to lower-cost storage — local or removable disk, network shares, tape, or deduplicated public object storage. In archive mode, verified copies replace the source file with a same-name stub, tiering cold data off primary storage while keeping it one recall away.

Capabilities

Post-backup copy jobs — run automatically after the backup job, driven by the same scheduler.
Rich selection filters — pick files and folders, then filter the set before copying.
Five destination classes — local disk, removable disk, CIFS/NFS shares, tape, and S3/Azure Blob/Google Cloud Storage with deduplication enabled.
AES-256 encryption of all copied data, at rest on every destination.
Archive (tiering) mode — delete the source after a verified copy and leave a stub in its place.
Recall workflow — users and operators bring archived files back on demand.

Selecting what to copy

Copy jobs start from an explicit file and folder selection, then narrow it with filters — combinable, and evaluated per file:

Filter	Examples
Type / extension	`.pst, .bak, *.iso`; include or exclude lists
Attributes	read-only, hidden, system, offline
Age (creation)	created more than 180 days ago
Last-access date	not accessed in 90 days
Last-modified date	not modified in 1 year
Size	larger than 100 MB

Copy destinations

Destination	Notes
Local disk	Any mounted path on the Storage Daemon
Removable disk	USB/eSATA media; safe eject after the copy verifies
Network share	CIFS/SMB and NFS
Tape	Standalone drives and autochangers
Public object storage	Amazon S3, Azure Blob Storage, Google Cloud Storage — with deduplication enabled, so repeated archive runs upload only unique blocks

All copied data is encrypted with AES-256 before it leaves the Storage Daemon; keys stay under your control.

Archive mode & stub files

With Mode = archive, each file is copied to the destination, verified (checksum read-back), and only then deleted at the source. In its place the plugin leaves a same-name stub file containing recovery instructions and a pointer to the archived copy (job, destination, path), so users who open it know exactly how to request the file back — and the folder tree keeps its shape for browsing. With Mode = copy, sources are left untouched.

Configure

Job {
  Name = "ArchiveColdFiles"
  Type = Copy
  Plugin = "file-archive"
  Schedule = "NightlyAfterBackup"
}

Device {
  Name = "ArchiveTier"
  Device Type = FileArchive
  Archive Target = "s3://company-archive/fileshare"   # or /mnt/usb, //nas/archive, tape changer
  Dedup = yes                                          # object-storage dedup
  Encryption = AES256
  Mode = archive              # or: copy (keep the source)
  Stub = yes                  # leave recovery-pointer stubs (archive mode)
  # Selection & filters
  Include = "/srv/fileshare"
  Filter Type = "*.pst,*.iso,*.bak"
  Filter LastAccess = ">90d"
  Filter Modified = ">1y"
  Filter Size = ">100M"
}

Restore & recall

Archived files are recovered like any other NGBackup restore: pick the file in the console or Web Console restore browser and restore it to the original or an alternate path — the recall replaces the stub with the real file. Operators can also recall in bulk (per folder or per filter), and the stub itself tells end users whom to ask and which pointer to quote. Restores read only the blocks they need; a single-file recall from object storage does not download the whole archive set.