Backup repositories
A backup repository is where backups go. HivePaaS keeps them with Kopia, which:
- encrypts every backup with the repository's password;
- deduplicates: a backup stores only what changed since the last one, so a daily backup of a large volume stays small;
- compresses what it stores.
Where a repository lives
- A cloud storage: a bucket on AWS S3, Cloudflare R2, Backblaze B2, MinIO or any S3-compatible service. Off your servers, which is the point of a backup.
- A volume of the cluster: on one of your nodes' disks. Quick, but lost with the node.
A cloud storage is set up first, in Integrations → Cloud Storages, with a key auth holding its access key.
Create a repository
In Integrations → Backup Repos, globally or in a project, create one:
| Setting | What to give |
|---|---|
| Storage | The cloud storage, or the volume. |
| Storage Path | A directory inside it, such as hivepaas/prod, for several repositories to share a bucket. |
| Password | The key every backup in it is encrypted with. |
| Compression Level | How hard to compress. |
| Retention Policy | Which backups to keep. |
Without the repository's password, its backups cannot be read, by anyone. Keep a copy of it outside HivePaaS.
The storage, the storage path and the password are fixed once the repository exists. The password can be changed later, with its own action.
An existing repository
Import Existing, in place of creating a new one, connects to a repository already in the storage, such as one of a previous installation: give its storage path and its password.
Retention
The Retention Policy keeps:
- Keep Last: the newest backups, however many;
- Keep Hourly, Keep Daily, Keep Weekly, Keep Monthly: the last backup of each of that many hours, days, weeks and months.
A backup is kept if any rule keeps it. A rule left empty adds no limit: it deletes nothing.
Cleanup
The Backup Repo Cleanup job, in Settings, runs daily: it deletes the backups the retention policy no longer keeps, and frees their space. Run Cleanup Now runs it at once.