My name is David Cook and I'm new to the Deja Dup list. I'm a systems librarian at Prosentient Systems in Sydney, Australia where I provide support for library technology. I primarily develop in Perl for the open source library management system Koha, but occasionally write in PHP, read Java from the open source digital repository DSpace, and read C from the open source indexing engine Zebra.
I've been using Deja Dup to make encrypted backups for my desktop at home for about a year, but I've recently started thinking about doing rotating backups where I'd keep one drive at home throughout the week and the other at the office, then swap once a week.
However, I noticed that swapping out the drives at the same mount point wouldn't work as Deja Dup was using something more unique than that (perhaps the UUID, I couldn't tell at a glance).
So I started thinking about possibilities for using multiple target locations. Naturally, I found the bug for "Profiles for different target locations" (
https://bugs.launchpad.net/deja-dup/+bug/324631). I agree with Michael Terry that having different profiles might be over complicating things. However, I think it makes sense to be able to have the same settings but backup to multiple locations. Michael mentioned that Time Machine doesn't offer profiles, and that might be right, but my friend who uses it mentions that it can back-up to multiple locations (he backs up to a pair of drives). While maintaining separate profiles could make the UI overly complex, specifying more than one location could be very straightforward.
That's all I'd want. I'd specify two USB hard drives (or rather their mount points that I specify in /etc/fstab) and then Deja Dup would backup to them if they're available. If one wasn't available (ie it's at the office), it would just do a backup for the drive that is available. Judging from the comments on the bug, it seems like the kind of thing that a lot of people would be interested in.
Honestly, I've thought about writing my own scripts to utilize Duplicity, but that leads to a few different problems. So my friend suggested that I fork Deja Dup. Not a bad idea, but I figure... why fork it when I can contribute code back?
I suppose the problem might not be as simple as I see it. The timing of when to do full backups is one thing I haven't thought about too much. Is that done weekly or based on a set number of previous incremental updates? I suppose I can go look...
In any case... I've taken a look at
https://wiki.gnome.org/Apps/DejaDup/GettingInvolved/Coding. I'm willing to set up my environment and learn Vala. I suppose I'm just curious... if I do make this change... would it be something that people would be interested in? Michael, would you be willing to merge it back in?
I don't think it would be very complex, and I think it would help Deja Dup to meet the needs of even more people than it already does. Personally, why I am a developer for work, I don't really do much technical stuff at home. I'm just someone who wants to backup his personal data in the event of a robbery (it's already happened once, although fortunately they didn't take my backup drive :P), meteor, or whatever else.
Anyway, sorry for the long email. I just wanted to give some context, and note that I am willing to contribute, rather than just saying "please give me this feature".
Cheers,
-David Cook
