Backups All plans, including Free
Scheduled Backups to Storage You Choose
Schedule website and database backups to local storage or to remote destinations configured through rclone, then restore from the same panel. Local/cloud scheduled backup (rclone) is part of every plan, Free included.
On this page: A backup job · Destinations · Restore · Backup is not the same as… · Plans
What a backup job holds
- What to back up — a website, its databases, or both. Databases can be MySQL, PostgreSQL or MongoDB instances managed by Nodekeeper. Virtual machines can be backed up from Pro.
- Where it is saved first — a local directory inside the owner's own file root. Paths outside it are refused, and the job runs as the website's own system user (or the database instance's), not as root.
- When — a 6-field cron schedule (seconds, minutes, hours, day, month, weekday), for example every day at 03:00.
- How many to keep — a retention count; when it is exceeded, the oldest run and its files are deleted.
- Optional upload — after a successful run, the files are copied to an rclone remote, each run into its own folder.
Every run is recorded with its result in the operation log.
Destinations
Local disk
Fast to restore from, but on the same machine — use it together with a remote copy.
S3-compatible and OSS
Object storage buckets you own, added as rclone instances in the panel.
Other rclone remotes
rclone supports many providers. Treat a provider as supported only after you have run a test restore from it.
- Our tests cover local disk, S3-compatible and OSS targets. Supporting a provider in rclone does not mean Nodekeeper has been tested against it.
- Storage, traffic and request fees of the destination are yours; Nodekeeper does not include cloud storage.
- A remote copy is not ransomware protection by itself: anything that can write to the bucket can also delete from it. Use bucket versioning or object lock on the provider side if you need that.
Restore, then check
Pick a run of the job and restore it from the panel. A successful job is only half of the evidence: open the restored site and query the restored data before you rely on a schedule. The product tour will show this end to end — change, restore, verify — rather than a list of green job results.
The restore recording is not published yet; the tour page says so.
A backup is not the same as…
| Mechanism | What it protects against | What it does not | Plan |
|---|---|---|---|
| Backup job (this page) | Mistakes, deleted files, a broken update — restore an earlier run | Data written after the last run | Free |
| Distributed storage snapshots | Fast recovery points of a disk inside the cluster; rollback needs 2FA and the disk out of use | Loss of the cluster itself; snapshots are crash-consistent and are not backups | Business |
| Distributed storage replicas / erasure coding | A failed disk or host | Deletion or corruption — it is replicated too | Business |
| Database replication | A failed database server (a replica is ready) | A bad query — it reaches the replica as well | Business |
Off-cluster backup of disk images on distributed storage is not available yet; back up the data inside them with a backup job.
Plans
| From plan | Backup and recovery |
|---|---|
| Free | Scheduled website and database backups, local and rclone destinations, retention, restore |
| Pro | Virtual machine backup and restore |
| Business | Distributed storage snapshots (manual and scheduled), rollback with 2FA, database replication |
Related: Web hosting · Storage and snapshots · Site access