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 chain breaks when one file that later restore points depend on is missing, unreadable or marked corrupt. Restore points taken before the gap still work. The restore point that needed the missing file, and every later one on the same chain, does not. The chain is usable again only when a new full backup starts a new chain.
So the first job is to find exactly where the gap is. That tells you which points you still have, how far back you must go, and whether a rebuild of the chain is needed.
What the error means
Veeam describes a chain as "the first full backup file, incremental backup files, metadata files and some additional files". It states the dependency plainly: "To roll back a VM to a specific point in time, you need a chain of backup files: a full backup file plus a set of incremental backup files dependent on this full backup file. If some file in the backup chain is missing, you will not be able to roll back to the necessary state." For forward incremental chains it adds that if you delete the full file "the entire chain of incremental backup files will become useless", and that deleting any incremental file before the point you want to reach prevents the restore, "since later incremental backup files depend on earlier incremental backup files".
On the page about missing restore points, Veeam says chains with missing restore points "get corrupted — you cannot perform backup or restore VM data from the missing restore point and restore points that depend on the missing restore point." It does not say that earlier points stay usable. That follows from the dependency rule: earlier points do not need the missing file. Treat it as reasoning from the documentation, and test an earlier point with a real restore before relying on it.
Bareos has no chain file, but a restore still needs a set of jobs. The console's "most recent backup" option selects "the most recent Full backup and all subsequent incremental and differential backups", and never selects a job that failed. If the File records of needed jobs have been pruned, Bareos says "the restore will be incomplete" and that it "does not correctly detect this condition".
Likely causes
| Cause | What you see | How to confirm |
|---|---|---|
| A backup file was deleted, moved or renamed by hand | Missing restore point; the next job sessions fail | Veeam: "If you delete backup files manually, subsequent backup or replication job sessions will fail." Compare the job folder with the restore point list |
| Repository unavailable, or in maintenance mode | Points shown as missing, files may still exist | Veeam names maintenance mode (scale-out repository) and an unavailable repository as causes |
| Health check found corrupt data | Restore point marked corrupted; job ends with an error | The job session result; the restore point state in Backup Properties |
| Files copied or archived and the originals deleted | Chain incomplete after a tidy-up | Veeam's rescan page: backups archived to tape and then deleted from the repository need a rescan |
| Bareos: a job in the set failed or was cancelled | Restore leaves out data you expected | Job status in the catalogue; failed jobs are never selected automatically |
| Bareos: File records pruned | Restore lists jobs with zero files | The JobFiles column shows zero where files were backed up |
Prove the cause before changing anything
Everything in this section only reads.
Veeam console. Open the Home view, select Disk under Backups, select the backup, and open Properties. Veeam says missing restore points can be forgotten or removed from this window, and "Veeam Backup & Replication requires some time to update information in the configuration database", so a point may not show as missing at once. Rescanning a repository updates that information, but Veeam recommends stopping or disabling the jobs first.
Risk: disabling a job stops its schedule. Re-enable it afterwards, and check the next run starts.
Veeam PowerShell. The PowerShell Reference shows this pattern:
$backup = Get-VBRBackup -Name "<job name>"
Get-VBRRestorePoint -Backup $backup
It returns the restore points in the configuration database. Compare their count and dates with the files in the job folder on the repository, which Veeam says holds VBK (full), VIB or VRB (incremental) and VBM (metadata) files. A date with a restore point but no file is the gap. Veeam does not document a field that flags a missing file, so do the comparison yourself.
Bareos console. list job=<job-name> lists all jobs with that name. For a job in the set, list files jobid=<id> shows its File records, list jobmedia jobid=<id> its media, and list volumes jobid=<id> the volumes. Empty File records for a job that backed up data means its records were pruned. A job with status not T or W did not finish cleanly, and the Bareos catalogue documentation lists the other codes.
Read the results as follows. If every file the restore point list implies is on disk, the chain is intact and the fault lies elsewhere. If one file is absent, note its date: that restore point and all later ones on the chain are affected. If the repository was unreachable, check it before concluding anything is lost.
What still restores
| What is missing | Veeam forward incremental chain | Bareos |
|---|---|---|
| The full (VBK) | Nothing on that chain; use an older chain or a copy | Every job that needs that Full |
| One incremental | Points before it; not it or those after | The other jobs in the set; files whose latest version is only in the missing job come back from an older job, or not at all |
| The VBM only | All points on disk, but the import must use the VBK by hand | Not applicable |
| The repository (offline) | Nothing until it returns | Nothing from that Volume until it is mounted |
| File records pruned, data intact | Not applicable | A non-selective restore of the whole job |
Veeam says it "will not be able to import backups if VBM files are unavailable" through a rescan, which is why the VBM row needs a manual import of the VBK.
Fix
File exists elsewhere. Put it back in the same folder, with the same name, from a copy, a snapshot or a recycle bin. Veeam's import page says to "ensure you have all required incremental backup files (forward or reverse) in the same folder where the full backup file resides." Then rescan the repository.
Disruptive step: a rescan imports and updates records in the configuration database. It does not move files. Stop or disable jobs first, as Veeam recommends, and re-enable them after.
Chain copied to another server or folder. Put every file of the chain in one folder and import it. Import from the VBM when you have it, because Veeam says it is "notably faster". The PowerShell Reference documents the import:
$server = Get-VBRServer -Name "<server name>"
Import-VBRBackup -Server $server -FileName "<path>\<job>.vbm"
Imported backups appear under the original job name with the _imported suffix.
Risk: an import adds records to the configuration database. It does not change the files. Veeam shows the result under Backups > Imported. The pages cited do not document how to take an imported record out again, so check that the folder holds the right chain before you import it.
File is gone for good. Decide which points you will keep, then start a new chain. Veeam documents two actions for missing points. Forget removes the records only, and "the actual backup files will remain on disk". Remove from disk also deletes the files.
Disruptive step: Remove from disk deletes backup files, including files of points that depend on the missing one. Veeam documents no undo. Use Forget first if you are unsure. On a scale-out repository, Veeam says the file tied to the missing point will be deleted from the capacity and archive tiers on the next offload and archiving job run.
Then create a full backup that resets the chain. Veeam says an active full backup "resets a backup chain", and all later increments use it as the new starting point. It reads the whole source again, so plan the window.
Health check detected corruption. Veeam ends the job with the Error status and starts a health check retry as a separate job session. The retry transports data blocks from the source datastore and writes a new full or increment; it does not repair the corrupt file itself. Linux immutable repositories "do not support repair". There Veeam marks the point corrupted and says to run an active full, otherwise later increments finish with the Error status.
Bareos. Choose the restore set by JobId when the automatic selection leaves something out (option 3 takes a list of JobIds; option 1 lists the last 20 jobs). When all the File records of the chosen jobs are pruned and the volume is intact, the restore documentation says Bareos "will then propose doing a full restore (non-selective)" of those jobs, because it still knows where the job data begins. Then run a new Full, so the next Incremental has something to follow.
If a Volume exists but its records are missing from the catalogue, bscan can recreate them. Start with the read-only form. The Bareos documentation says bscan -v -V Vol001|Vol002 /dev/nst0 "will give you an idea of what is going to happen without changing your catalog". The page also says to try to recover your previous catalogue backup rather than bscanning.
The form that writes is bscan -V <volume> -v -s -m <device>. Bareos says to scan into a fresh catalogue, or into the existing one "after having made an ASCII copy". Scanned-in records of purged volumes are pruned again by the next run of a job with the same name unless you change the Job and File retention times for those jobs.
Risk: bscan writes to the catalogue. Back up the catalogue first, and restore that backup to undo. Without the backup there is no way back.
Confirm the fix
Restore a small set of files from the oldest and the newest point that you believe are usable, to a scratch location. A restore point list is not proof. Then check that the next scheduled job completes without the Error status, and that the point list shows an unbroken sequence from the new full.
Example (illustrative): job fs01-job writes a full on Sunday and an incremental every night at 01:00 to /srv/backup/fs01-job on backup01. After a disk repair, Wednesday's file is gone. Sunday, Monday and Tuesday still restore (3 of 7 points). Wednesday is missing and Thursday, Friday and Saturday depend on it (4 of 7). If fs01 fails on Saturday at 12:00, the newest usable point is Tuesday 01:00, which is 107 hours old (4 days plus 11 hours). With the chain intact, the newest point would be Saturday 01:00, 11 hours old. Forgetting the points and taking an active full on Saturday evening resets the chain from then on.
What to watch afterwards
- Alert on a job session that ends with an Error status after one that succeeded. A broken chain can appear first as a failed increment.
- Run Veeam's health check on a schedule. By default it runs at 5:00 on the last Saturday of the month, and it verifies only the latest restore point.
- Treat a repository move as a project. Copy the whole job folder, compare file counts and sizes, and rescan before you decommission the old repository.
- Never tidy a repository by hand. Veeam says "you must not manually delete separate backup files" and to use the retention policy.
- Restore-test an old point as well as the newest one.
Read next
Related articles you may want next.
- Backup & RecoveryRetention that outlasts your ability to restoreHow retention periods, backup chains, keys and obsolete formats come apart, and how to test the path you claim to keep.Read article
- Backup & RecoveryChoosing the least-untrusted recovery copy after encryptionA decision order for picking a recovery copy after encryption, using reachability, time, integrity, and keys rather than the newest green job.Read article
- Backup & RecoveryFinding a bad backup before the day you need itWhat hashes, signatures and sample restores detect in backups, which faults they miss, and when to treat a copy as unusable.Read article
Risks and limits
These commands have not been run or lab-tested. Veeam does not state that earlier points survive; it follows from the dependency rule. Veeam's behaviour for immutable repositories, per-machine chains and scale-out repositories differs in detail, so read your version's pages before acting. Bareos wording and behaviour come from its documentation and source branch, not from a test run.
Do not forget or remove a restore point until you know what depends on it. If the data matters, copy the remaining files elsewhere first.
NetLeverage includes a daily full disaster-recovery backup of each hosted server at no extra charge. Talk to an engineer.
Sources
- Backup Files — Veeam, page updated 2024-07-19, accessed 17 November 2026
- Forward Incremental Backup Retention Policy — Veeam, page updated 2026-01-12, accessed 17 November 2026
- Removing Missing Restore Points — Veeam, page updated 2026-02-05, accessed 17 November 2026
- Health Check for Backup Files — Veeam, page updated 2026-08-04, accessed 17 November 2026
- Active Full Backup — Veeam, page updated 2025-07-09, accessed 17 November 2026
- Deleting Backups from Disk — Veeam, page updated 2025-02-10, accessed 17 November 2026
- Importing Backups Manually — Veeam, page updated 2025-08-07, accessed 17 November 2026
- Rescanning Backup Repositories — Veeam, page updated 2025-10-20, accessed 17 November 2026
- Get-VBRRestorePoint — Veeam, page updated 2026-08-19, accessed 17 November 2026
- Import-VBRBackup — Veeam, page updated 2026-08-19, accessed 17 November 2026
- Bareos Programs (bscan) — Bareos project, footer: Revision November 06, 2026, accessed 17 November 2026
- The Restore Command — Bareos project, footer: Revision November 06, 2026, accessed 17 November 2026
- Bareos Console — Bareos project, footer: Revision November 06, 2026, accessed 17 November 2026
- Catalog Tables — Bareos project, footer: Revision November 06, 2026, accessed 17 November 2026
This article was originally published on .