I’m running Duplicati 2.0.5.1_beta_2020-01-18 on Linux Mint 19.3. I recently bought a new external HDD because my old one died, and I’m trying dual backups using rsync (via Timeshift) and Duplicati because there’s plenty of space. (Yeah, I know it’s not truly redundant when I’m using the same device for both kinds of backup, but it’s at least a little insurance in case one set gets corrupted. I’m also doing cloud backups with Duplicati for essential stuff.)
My home directory (according to the Duplicati log) has 576,113 files totaling 331 GB. The initial backup of all those files took 1.25 hours with rsync and 9.25 hours with Duplicati. That’s quite a difference! Of course, I’d expect subsequent incremental backups to be much faster in both cases.
Other than the pseudo-redundancy, what are other advantages of a Duplicati backup vs rsync? I’m calling on the community’s wisdom for advice. Would I be better off sticking with rsync alone for the local external HDD and using Duplicati for the cloud destinations?
They are fundamentally different tools. rsync is for synchronization, and Duplicati is for backups.
With Duplicati you get deduplication and efficient storage of multiple versions. With rsync you only get one version. If a file gets corrupted and you don’t discover it before your next rsync, the file will be unrecoverable.
Features (and to a bit, Overview above it) talk about what Duplicati tries to do. Althugh I don’t use it, Timeshift (if this is what you are using) sounds like it’s for OS backups, actively excluding user data, pointing to other tools for general backup. Possibly some use the same hard-linked clone approach.
I am using Timeshift, and you’re right about the default settings. They exclude user data. Anyone who lets it run as configured in a standard Linux Mint installation, for example, will just get system backups. (As some have said, that’s more or less equivalent to Windows restore points.)
But it’s possible to change the settings to include user data as well, and that’s what I did in the comparison I described in the OP. It seems to make a direct copy on the target drive of each file in the source set.
I can’t give any opinion based on Timeshift use, but many of Duplicati’s features have a lot more value in a case where you’re putting the backup on a remote site that you may or may not trust 100%, connected by a limited link, and having limited storage space. Duplicati does keep a close watch on the remote files compared to what they should look like. This can notice some issues, but the DB itself can suffer issues.
I’d say try them both, pick one. Maybe someone else will stop by, but a lot of times participation is low…
Direct copy is nice and easy if space is ample. If you find it gets tight, then consider heavier processing.