About this article: This article is reviewed against the cited sources. This is source research, not lab validation. Check the stated versions and your own requirements before making changes.
A backup job reads the live source and writes a backup. A backup copy job reads an existing backup and writes another one. Whether the second set can restore on its own, and whether it has its own chain of restore points, depends on the product. In Veeam Backup & Replication a copy job builds its own backup file chain with its own retention. In Bareos a copy job duplicates the data of one finished job to another volume, and the duplicate is a copy, not a backup, until the original is gone.
This matters for the second "2" and the "1" in 3-2-1. A copy that cannot be restored without the first set is a copy of the data but not a second way to recover.
How is a copy job different from a backup job?
A backup job talks to the machine being protected. It takes a snapshot, reads the data and writes backup files. A copy job does not. Veeam describes it this way: it "accesses backup files in the source backup repository, retrieves data blocks for a specific machine from the backup file, copies them to the target backup repository, and composes copied blocks into a backup file in the target backup repository." It "does not require creation of additional VM snapshots or VSS snapshots and does not produce load on machines whose backups are copied."
Bareos says the same about its copy jobs: they "do not involve the File daemon", which is the agent on the protected machine.
Two consequences follow. A copy job cannot capture anything the first backup missed, because it only moves what is already there. And a fault in the first backup can be copied into the second.
Does the copy get its own chain in Veeam?
Yes. Veeam documents that "the backup copy file has the same format as the primary backup, so you can restore necessary data directly from it in case of a disaster." The pages describe several properties of the separate chain:
- Its own retention. "The retention policy of a backup copy job does not depend on retention policy of the source backup job. The backup copy job has its own retention policy settings." Veeam offers short-term retention and a long-term "GFS" (weekly, monthly, yearly) policy.
- Its own full backups. For long-term retention Veeam "creates synthetic or active full backup files and marks them with GFS flags", and says "GFS backups are always full backup files". With GFS "instead of one backup chain consisting of one full backup and incremental backups, you will have several backup chains".
- Its own active full. An administrator can create an "active full backup copy" and add it to the target chain. Veeam notes this is useful to change settings such as encryption, which apply "starting from this full backup".
- Its own health check. On a schedule that by default starts at 5:00 on the last Sunday of each month, Veeam checks the latest restore point in the copy's chain by comparing stored checksums. If it finds corrupted blocks, it sends valid blocks again in the next restore point, so the chain gets "fixed".
Veeam also documents what the copy does not take. It copies "the most recent restore points", not necessarily all history, even on the first run. It does not copy "incomplete restore points", points "locked by the backup transformation process (merge, transform)", or points already copied. If the source job's block size changes, the copy job does not copy new points until an active full is made.
There are two modes. In immediate mode, restore points are copied "as soon as they appear in a source backup repository". In periodic mode the job "copies the latest source restore point according to schedule".
What does Bareos do differently?
Bareos documents Copy as "essentially identical to the Migration feature with the exception that the Job that is copied is left unchanged. This essentially creates two identical copies of the same backup."
The unit is a job, not a chain. "Migration takes place on a JobId by JobId basis. That is each JobId is migrated in its entirety and independently of other JobIds." Bareos adds:
- The copy "is treated as a copy rather than a backup job, and hence is not directly available for restore."
- "If Bareos finds a copy when a job record is purged (deleted) from the catalog, it will promote the copy as real backup and will make it available for automatic restore."
- A job is copied only if all volumes holding it are marked Full, Used or Error. Volumes still marked Append are not considered.
- The pool being copied from must name a
Next Poolwhere the copies go, and the original job's resource must still exist in the Director configuration, or the run fails. - A copy or migration job runs only when you define and schedule it; setting a pool water mark alone does not start one.
So in Bareos the second set is a faithful duplicate of each job, written to other media. It does not have an independent retention story unless you configure the pool that way. Treat a Bareos copy as a protected duplicate that becomes the restore source only if the original record is purged.
How do the options compare?
| Question | Veeam backup copy job | Bareos Copy job | One-time copy of backup files |
|---|---|---|---|
| Reads from | Existing backup files | Existing job data on a volume | The backup files |
| Touches the protected machine | No | No (no File daemon) | No |
| Own retention | Yes, independent of the source job | Follows pool and catalog rules you set | Veeam's one-time copy replaces the source repository's immutability or retention settings with the target's |
| Own chain of restore points | Yes, with its own fulls and increments | Copy of each job's data, kept as a copy | Copies "the whole backup chain" |
| Restorable alone | Yes, same format as the primary | Not directly until promoted | Veeam shows copied backups in a node with the "(Exported)" postfix |
| Copies everything the source holds | Recent points; skips incomplete ones | Selected jobs, per JobId | Whole chain |
Veeam separates the scheduled copy job from a one-time "copy backups" operation, which copies the whole chain and keeps its format. The second is a file copy of the chain, not a job with its own copy retention policy; Veeam lets the copy be given an automatic deletion time ("Delete this backup automatically in").
Example (illustrative): backup01 keeps 14 daily restore points of fs01, which is 1,000 GiB with 20 GiB of new data per day, using no compression or deduplication so the arithmetic is simple. The source chain holds about 1,000 + 13 × 20 = 1,260 GiB. A copy job to /srv/offsite keeps 30 daily points and 4 weekly GFS fulls. The 30-point chain holds 1,000 + 29 × 20 = 1,580 GiB, a floor: Veeam says that once GFS is enabled the chain switches to forward incremental and "the increments are no longer merged to the full backup file", so the copy can hold more points than the retention count at times. Each GFS weekly full is another whole machine image, 1,000 GiB, so four add 4,000 GiB. The copy side needs about 1,580 + 4,000 = 5,580 GiB, more than four times the source side. That is the price of a separate chain that survives the loss of the first, and it is why a long-retention copy needs more space than the primary.
Which one do you have, and does the copy restore on its own?
Check these on your own system or with your provider:
Is the copy made from the backup or from the source? A job that reads the server again is a second backup job, with its own load and snapshot. A copy job reads backups.
Is there separate retention for the copy? Ask what is kept, for how long, and what deletes it.
Is there a health check on the copy? In Veeam it is a setting on the job, "Perform backup files health check", under the target's advanced settings.
Can it restore without the first set? Disconnect or deny access to the primary repository in a test, and restore from the copy alone. For Veeam, look under Backups, Disk (Copy) in the console, which Veeam names as the node listing backup copies.
Risk: while the primary repository is disconnected, jobs that write to it fail. Run the test outside the backup window, then reconnect it and confirm that the next job succeeds.
In Bareos, find out how the copies were chosen. The Bareos page says the catalogue's Job table has a
PriorJobIdcolumn, set to the original JobId for migrated jobs, and that a debug level of 100 or more shows how a migration or copy job selected its jobs. The page does not say whatPriorJobIdholds for a copy, so check your own catalogue read-only.
Disruptive step: Veeam's "Delete from disk" removes backup files from the target repository; "Detach from job" makes the job "start a new backup chain" with an active full at the next run. There is no way back for files deleted from disk. Detached files are not deleted: Veeam keeps them in the repository under a node with the "(Orphaned)" postfix, retained by the background retention process, and the job starts a new chain with an active full at its next run, so expect extra space use. To back out of a detach, there is no documented way to re-attach them in the pages cited. In Bareos, the original of a migrated job has its file records purged, so a restore from it can only restore all files. Do not remove a source chain until a restore from the copy has been tested.
Read next
Related articles you may want next.
- Backup & RecoveryThe 3-2-1 backup rule in 2026: what each number must proveWhat the 3-2-1 backup rule says, what it misses, and how ASD's Essential Eight and ISM backup controls add access, immutability and testing to it.Read article
- Backup & RecoveryVerify a Backup Delivered to S3 or SFTP Before RecoveryCheck delivered backup identity, completeness and integrity, then rehearse extraction and recovery using independently available credentials.Read article
- Backup & RecoveryLogical air gap and physical air gap for backup repositoriesWhat offline and logically isolated backup copies each stop, where they fail, and how CISA and NIST describe a copy an attacker cannot reach.Read article
Risks and limits
- These steps have not been run or lab-tested. They are taken from the vendors' documentation.
- Veeam's pages are the version 13 user guide, and several note that earlier versions behave differently (for example periodic copy jobs created in version 11 or earlier are deprecated). Check your version.
- The Bareos page is from the bareos-25 documentation, revision 6 November 2026. It refers to behaviour "beginning from Bareos Version >= 13.2.0" for one feature and says some migration selection types are experimental.
- Other products use the words "copy", "replica" and "secondary" differently. A vendor's documentation is the source for its product.
- The example assumes no compression, deduplication or block cloning, which would shrink real figures.
- A copy shares any flaw in the first backup, such as a file missing from the job, because it only copies what the first backup holds.
NetLeverage includes a daily full disaster-recovery backup of each hosted server at no extra charge. Talk to an engineer.
Sources
- About Backup Copy — Veeam Backup & Replication 13 User Guide, page updated 2026-10-13, accessed 17 November 2026
- Backup Copy Modes — Veeam Backup & Replication 13 User Guide, page updated 2026-10-07, accessed 17 November 2026
- Restore Point Selection — Veeam Backup & Replication 13 User Guide, page updated 2025-03-17, accessed 17 November 2026
- Retention Policy for Backup Copy Jobs — Veeam Backup & Replication 13 User Guide, page updated 2026-11-16, accessed 17 November 2026
- Long-Term Retention Policy (GFS) — Veeam Backup & Replication 13 User Guide, page updated 2024-08-16, accessed 17 November 2026
- Active Full Backup Copies — Veeam Backup & Replication 13 User Guide, page updated 2025-08-31, accessed 17 November 2026
- Health Check for Backup Files — Veeam Backup & Replication 13 User Guide, page updated 2025-09-16, accessed 17 November 2026
- Removing Backups — Veeam Backup & Replication 13 User Guide, page updated 2026-07-23, accessed 17 November 2026
- Copying Backups — Veeam Backup & Replication 13 User Guide, page updated 2026-10-07, accessed 17 November 2026
- Maintenance Settings (Backup Copy) — Veeam Backup & Replication 13 User Guide, page updated 2026-07-27, accessed 17 November 2026
- Migration and Copy — Bareos documentation bareos-25, revision 6 November 2026, accessed 17 November 2026
This article was originally published on .