#velero
> "I guess this makes all the "old" backups that include #velero already unrestorable?"
> Yes
Guess who found out the wrong way—a significant flaw in their default configuration.
https://github.com/vmware-tanzu/velero/issues/7057#issuecomment-1794124642
The MariaDB Operator messed up, clearing the entire jorijn.com database. Velero messed up, leaving me unable to restore the backups.
I'm very annoyed by this.
Turns out the #kubernetes backups, taken using #velero, are failing on #digitalocean not supporting their plugin. Apparently, you can only take a snapshot once every 10 minutes, resulting in the snapshot provider failing the backup.
The repository looks abandoned.
Falling back to plain file system backups for now.