A backup you have never restored is a hypothesis, not a backup. Verification is the step that turns it into a fact, and it is the step almost everyone skips — usually until the day it matters.
This guide covers what verification actually means, what the free Windows backup tools on this site can and cannot check for you, and a schedule you can realistically keep.
Three levels of verification
| Level | What it proves | Effort |
|---|---|---|
| The job reported success | Almost nothing — only that the software did not crash | None |
| Checksum or archive test | The backup file is readable and not corrupted | Automatic, adds run time |
| Test restore | The data comes back and is usable | Manual, the only one that counts |
Most failures hide between the first and the third row: the job ran, the file exists, and the contents are unusable — an empty database, a locked file that was silently skipped, an archive nobody can open because the password was never written down.
What the free tools give you
SyncBackFree
SyncBackFree is the strongest free option for verification specifically. It can compare source and destination by size, date or contents, and it produces a detailed per-file log of what it copied, skipped and failed on. Commercial use is explicitly permitted.
Set comparison to file contents on your most important job. It is slower, but it is the difference between “the file exists at the destination” and “the file at the destination is the same file”.
FBackup
FBackup writes standard ZIP archives and can test the archive after the job. Because the format is standard, you are never dependent on the tool itself to read your data back — a real advantage that proprietary formats do not offer. Commercial use is allowed.
Personal Backup
Personal Backup keeps thorough logs and can verify copied files after writing. It is aimed at file-level backup of user data and does that job precisely, with a lot of control over filters and scheduling.
Iperius Backup Free
Iperius Backup Free handles scheduled file and folder backup to disk, NAS and external drives, with email reports on the result of each job. Those reports are what make unattended jobs trustworthy — a silent job is an unverified job.
A verification routine that survives contact with reality
- Every run — automatic checks. Turn on content comparison or archive testing in the tool. Accept the extra runtime on your important jobs; use fast comparison on the rest.
- Every run — a report you will actually see. Email on failure at minimum. A job that only logs to a file on the machine that died has told you nothing.
- Every week — look at the numbers. Open the log and check the file count and total size against last week. A backup that suddenly got 90% smaller ran “successfully”.
- Every month — restore something. Pick a file at random, restore it to a scratch folder, open it. Ten minutes.
- Every quarter — restore something hard. A whole folder tree, a database, a user profile. Time it, and write down how long it took. That number is your real recovery time.
- Every quarter — test the off-site copy too. The copy you never touch is the one most likely to be broken.
The failures verification actually catches
- Open and locked files skipped. Databases, mail stores and anything in use may be silently omitted by file-level tools. The log will say so; nobody reads the log.
- A source path that moved. The job keeps succeeding, backing up a folder that no longer contains anything.
- A full destination. Some tools report success after writing a partial set.
- Permission changes. A service account loses access to one share and that share quietly disappears from the backup.
- An archive password nobody recorded. Encrypted backups are only as good as the key management around them.
- A dead scheduled task. The task fails to start after a reboot and nothing complains, because nothing is watching for absence.
That last one is worth a separate control: monitor that the job ran, not just that it succeeded. A missing report should raise an alarm. EventSentry Light can watch Windows event logs and scheduled task results on your servers and tell you when the expected event does not appear.
Keep the evidence
If you are audited, “we back up nightly” is a claim. Dated logs, restore test records and email reports are evidence. Keep backup logs somewhere separate from the machine being backed up — Graylog for centralised collection, or LogExpert if you just need to read large log files quickly.
What free tools cannot verify
File-level tools cannot prove that an application will start from a restored dataset. A consistent database or mail-store backup needs either the application’s own dump tools or a backup product with the matching agent. If you are backing up a live database by copying its files, verification will tell you the files copied — not that they are restorable. Handle those with the application’s own export, and back up the export.
FAQ
How often should I test a restore?
A single file monthly, a meaningful dataset quarterly. Anything less frequent and you are trusting software you have not checked in a season.
Which free backup tool verifies copied files?
SyncBackFree compares by content, FBackup tests its ZIP archives, and Personal Backup can verify after copying.
Can I back up a database with a free file-level tool?
Only if the database is stopped or you back up an export produced by the database itself. Copying live database files usually produces something that will not restore.
Which free backup tools allow commercial use?
SyncBackFree, FBackup and Personal Backup all permit it.
What is the single most useful check?
Comparing the file count and total size of this run against the last one. It catches silent partial backups better than anything else that takes ten seconds.
