Same repository, same storage, different snapshot id?

Support
#1

Hi all.

My setup, which I use for photography among other things, is just a linux laptop with limited internal storage and an external hard drive which I never carry with me. In my home directory I have a ‘Photos’ directory inside which I have one directory per year. Older years are soft links to directories in the external drive.

I’m setting up duplicacy for backups. My backups will sometimes be done with the external drive mounted, but most of the times without. A straightforward approach is simply ‘duplicacy init’ and run backups regardless if the external drive is mounted or not (if it’s not I get, e.g., ‘Failed to read the symlink: 2019’ but the process otherwise completes successfully).

If I ever needed a full restore I’d restore the last revision that includes the external data plus the most recent revision.

But… I figure that a bit of bad luck with ‘duplicacy prune -keep <n:m>’ could leave me only with revisions that do not include external data.

First question: is that a valid concern?

To prevent that, I’m planning on backing up with two separate snapshot ids, one that I’d use when the external drive is mounted, and another when it isn’t, and prune both sets separately.

Second question: Do you foresee any issues with this approach?

(Yes, alternatively I could backup the external data and internal data separately, but I’d rather not).

Thanks in advance.

Juan

#2

Yes.

But, are you deleting photos often? Maybe you don’t need to prune in the first place.

Great idea! I don’t see any issues. I would go a step further and add pre-backup script to check if the data is mounted and fail backup otherwise. Just to be safe.

#3

Thanks, saspus.

True, I might not need to prune. I’ll go with the 2-ids approach anyway. It’ll make restoring, if I ever need it, easier. And yes, it’ll all be scripted.

1 Like