What exactly is the backup scope of borgbackup?

Unfortunately I couldn’t really find the answer anywhere. What exactly is backed up with borgbackup? So actually just the data or also the respective Docker containers? If Nextcloud no longer starts after an update, does borgbackup restore all the configuration files right?
If I already have a backup that protects all of the container’s resistent datas, do I still need borgbackup?


Borgbackup has it’s own page :
slight_smile: GitHub - borgbackup/borg: Deduplicating archiver with compression and authenticated encryption.
Borg is used because nexcloud believes
it’s an easy recovery platform,
that is a very small footprint
that is designed to only backup changes.
It is designed for your particular case, according to the docs.
They say to create a backup right after a working install, to preserve that image, for any future failures.

Hi, you can read more about aios backup solution here: GitHub - nextcloud/all-in-one: The official Nextcloud installation method. Provides easy deployment and maintenance with most features included in this one Nextcloud instance.

Thank you both.
I think borgbackup is great but I’m not sure if I can use it. I have about 2TB of data, mainly large photo gallery. The backup, at least the initial one, takes estimated 20 hours. The containers would have been stopped all this time. I’m afraid incremental backups also take a long time… That wouldn’t be optimal.
So I probably wouldn’t have any advantage over my other backups.

1 Like

For Nextcloud recovery first read Backup and Restore.

I am not a backup expert for large amounts of data. But I think you can make the backup at Nextcloud runtime. Then you stop the Nextcloud briefly and just perform a quick incremental backup, which only takes a few seconds to minutes even with 2 TB, as only all checksums are checked and only changes are rewritten.

On my personal instance with more than 2TB, it only takes a few minutes to perform an incremental backup. So the downtime should be fine even with that amount of data.

Ok good to hear that incremental is faster. Then it’s just a matter of the initial backup, which takes a very long time. However, I can’t do this at runtime or?
As soon as I start it via the master container in the AIO, all containers are stopped for the time the backup is running. This runs for I think about 20 hours on my 2TB. The path to the backup is via a mounted directory on the host where the VM is running. I don’t know if it should take that long.

20h does not sounds unreasonable for 2tb of data for the initial backup…

OK thanks. Then it’s just really stupid that the containers are stopped during this time, but maybe I just have to go through that initially.