A backup that has never been restored is only a copy.
We set up, monitor and test backups for Linux VPS and dedicated servers, then restore sites, databases and mail when the server, the account or the files are gone.
Your server. Your provider. Our Linux expertise.
Provider snapshots stay on the same platform
Hetzner, OVHcloud and the others can snapshot a disk. That helps before an upgrade. It does not help if the account is closed, the snapshot is corrupted, or you need last night’s database rather than last week’s disk image. Production servers need a copy off the machine, with a restore that has been tried.
The backup account should not be the same login as the server. If the server is compromised, that copy has to remain out of reach.
Files, databases and mail, on a schedule
Web roots, MariaDB or MySQL dumps, mailboxes, panel config and the NGINX or Apache files needed to rebuild. Not the whole disk, unless a disk image is actually the right restore.
Another server, object storage, or a backup host you control. Retention is agreed: daily for databases, longer for files. The job reports failure instead of failing quietly.
A restore of a database and a site onto a temporary path, not a checksum alone. On a monthly plan, backup status is part of the health check.
Recovery starts from whatever copy still exists
A bad deploy, a dropped table, a deleted site. We recover the database or the files from the last good copy and leave the rest of the server alone.
Disk dead, account locked, or the machine was wiped. We bring the copy up on a new VPS or dedicated server, then point DNS. This is a quoted recovery, close to a migration.
Setup is quoted. Checking it can sit on the monthly plan.
A restore, a failed job, or a first look at a backup that nobody has tested. Minimum 1 hour, then 30-minute steps. Emergency restore is €95/hour, subject to availability.
Essential at €79 already includes a backup status check. Managed and Managed Plus add a review of whether the job is still restorable, not only whether it ran.
We cannot restore a copy that was never made
If the only copy lived on the disk that failed, or the backup user could write to the backup store, there may be nothing to restore. We will say that after the first look, before billing a search through empty storage.
The backup should survive a change of provider. That is the point of keeping it off the server you are paying us to administer.
Your server. Your provider. Our Linux expertise.
Need a backup that can actually be restored?
Tell us what is copied today, and where it goes. We reply with the gap and the work before changing the job.