The automatic restore test and its report
Sefthy restores a copy of the server on its own in an isolated environment, checks that it boots and leaves you a report on the DR page.
A backup that has never been powered back on is a promise, not a certainty. So Sefthy periodically takes the latest backup of a DR, restores it in an isolated environment in the cloud, looks at whether the machine boots and then deletes everything. What is left is the Automatic Restore Test, the report you read from the DR page.
The customer server never notices. The restored copy is not connected to the company network and it is gone when the test ends.
Which plans, how often
| Plan | Automatic test |
|---|---|
| Sefthy PRO | once a week |
| Sefthy | once a month |
| Sefthy MINI | not included |
| Sefthy TRIAL | not included |
The test only runs if there is a backup to try. If the DR has no backups yet, or the latest one had already been verified in the previous round, the test is skipped and the page says so with the amber Autotest skipped tile. That is a warning, not a failure.
What it checks
- VM restart. Once the machine is up, Sefthy captures the console screen and hands it to the automatic analysis, which says whether the system finished booting.
- Network reachability. Sefthy runs a network reachability check to see whether the restored copy answers on the address the server has at the customer site.
- Application checks. If you defined some on the DR, the agent runs them inside the restored machine and reports each outcome.
Where to read it
Anteprima video temporaneamente non disponibile.
Apri il file direttamente →- Open the DR page and scroll down to the Latest Activities card.
- Below the grid, click the Automatic Restore Test tile, the one that reads Report.
- View Automatic Restore Report opens. On the left is the history of the last 3 months, on the right the selected test.
- At the top you find the three outcome tiles, VM restart, Network reachability and Validation. Below them are the application checks with their detail, the screen captured from the console and the explanation written by Sefthy AI.
- With Prev and Next, or by clicking a date in the history, you compare one test with the previous ones.
If the tile is not there, that DR has no test on record yet.
When the verdict says “unavailable”
The judgement on the boot comes from the automatic analysis of the screenshot. If that service does not answer, the VM restart tile stays neutral and reads AI verdict unavailable. It does not mean the server failed to boot, it means nobody was able to judge the screen. The screenshot stays in the report, look at it and decide yourself. A test like that is marked as one to review, never as failed.
When something goes wrong
If the machine does not boot, or does not answer on the network, the test does not pass. Sefthy raises the DRs with problems alert on your home page and sends an email with the subject Restore AutoTest failed.
The email reaches the users of your company who turned on the AutoRestore test failed alert in their own Profile, card Alert / Notifications. If nobody turned it on, you still see the outcome in the Console. Alerts for the same DR do not keep repeating, so a problem that lasts does not flood your inbox.
Application checks
The test verifies that the machine powers on. To know whether the application works too you need application checks, up to ten per DR, among path or file, TCP port, service and web address. You define them from the DR page menu under Post-restore validation, and from there you try them right away on the production server with Try now.
Anteprima video temporaneamente non disponibile.
Apri il file direttamente →A check defined after a test does not rewrite the past. The report freezes the checks as they were at test time, new ones run from the next restore onwards.
Taking it away as a PDF
The history of the tests downloads as a document, together with the backup outcomes, from the Reports entry in the three dot menu. The full round is in Reporting, on demand PDFs and automatic reports.