I’m using the Nextcloud AIO backup tool. The backups are stored on a separate drive (different from the one the active Nextcloud data is stored on). I’d like to start using a second backup drive to swap them from time to time. My reasoning is this: One is plugged into the server and a new backup is written to it every night. The other is at a secure location outside my home. Therefore, if the active backup drive (or the whole home) gets destroyed, the data still exists.
Is that even possible? The backup path will presumably change depending on which drive is active. I’d have to mount both drives to the same path, wouldn’t I?
How does Borg Backup handle different states of previous backup? Wouldn’t it presume the data state of last night but find the state that may be weeks old when I swapped the drives during the day? Would incrementing the backups still work?
Without any dependency to nextcloud, you could buy one of those two slots SATA adapters that can clone the disk in Slot 1 to the second. That way, from time to time you can clone your primary drive whenever you like.
That’s a good idea. Thank you, @sven1234. I hadn’t thought about that. But I would need to change my hardware, I guess. I’m using usb connected drives and I haven’t seen raid type bays where you can connect usb drives so far. But maybe I’m wrong.
If you have your permanent storage attached via usb, you could attach the clone docking station to another usb port and put your backup disk in it. That one stays in there. The disk you want to store off-site, can be put in the second slot, cloned from the permanent backup disk and then it is put back off-site.
Per borg docs, they don’t recommend cloning a repo as a second backup. They recommend having 2 separate repos and doing backups to both. If one of the backups is corrupt and you clone, then you clone the corrupted backup. If you do 2 separate backups and one is corrupt, the other is not affected. I don’t use AIO so unsure how that works. I use linuxserver nextcloud and do my backups manually using borg scripts I write myself. Setup 2 repos, one local and one remote. Then in the script simply loop through your repos and do 2 backups.
Look, your main idea is correct. However, if it were in my case, I would increase one more backup copy. This is because the HD that receives the first backup is on all the time. In other words, if for some reason it burns, both HDs would be affected, not just one. The most correct thing would be for your backup HD to be connected just to make the backup and then removed and stored.
For example, in my Nextcloud installation, I don’t use AIO or the default backup app. Instead, I use Borg. My backup routine is as follows: every day at 2 am I mount my external storage (Microsoft Family Onedrive with 6 accounts of 1 TB each) as a local drive using Rclone. I then back up all Nextcloud data. At 04:00 in the morning, I run a backup of the server settings. After backup, the script waits 3 hours to unmount the remote drive. On the weekends, when I’m at home, I turn on the local HD and synchronize the backup that was made on Onedrive with Rclone.
As for Borg not liking 2 repositories with the same ID and key, you could just use one repository. If your default repository becomes corrupt, simply clear Borg’s cache and point to the current repository