I apologize if this is not the right forum section, but I couldn't think where else would be a more appropriate section.

I make regular file backups to my own cloud, but this is a hassle to maintain (I used one local server and one remote server). Currently, also, I backup only select files.

I wanted to get some thoughts on a proposed backup scheme, and see if I've finally gone off the deep end.

- Buy a monstrous external (or 2nd internal) hard drive, set up Encfs container, clone (dd?) current HD there. Then, sync the container with, say, SpiderOak. I would mount the Encfs container nightly and perform regular rsyncs between primary HD and this Encfs container. Sync the container with SpiderOak (?) regularly.

Is it crazy to want to backup a clone of the HD instead of select files? The reason for the entire HD backup is in case my current HD dies, I'd have a working backup ready to go.

I used to sync truecrypt with Dropbox a while ago, and that worked ok. But I switched to using my own cloud because I wanted greater control, and Dropbox's 2GB size restriction.

The reason why I want to move away from my current cloud setup is because, to perform a backup, I now have to decrypt the truecrypt container *on* the remote machine, and sometimes the re-encryption fails and so the files are just sitting there exposed. I'd love to use SSHFS and Encfs (and have tried), but the backup speed is atrocious. Perhaps this is because the remote machine is one of these fan-less low horsepower ones, but it just doesn't work well.

I'd greatly appreciate any advice/thoughts/feedback on the proposed structure.

Thanks!!!

You can clone your drive, compress it and also encrypt it using clonezilla. Depending on your drive size and the compressibility of your data (did I invent a new word here?), this might be a practical solution - for disaster recovery.

But not a good solution for the perhaps more frequent case of restoring an accidentally deleted file. You don't want to restore a full backup just to find that file. So, in addition to the drive clone, you should also perform a file-by-file backup.

Thanks for the response. I was hoping for some specifics. I'm aware of file by file backups, there are definitely methods to doing that, but it doesn't really go to the question at hand.

For sure I want to clone my drive and keep it off-site. If I go the drive backup route, I won't bother with encrypting with software since my hard drives are all FDE. Having a full clone also allows me to get back up and running in no time.

I'm also not as concerned about accidentally deleted files, since I operate usually within the GUI, and anything deleted goes to the recycle bin. In the even I'm not, my syncs are usually not scheduled to run until each night, and I'd likely notice that same day if a file is missing. If I were really concerned about this, I would use duplicati or something of that nature as the backup mechanism.

Any concerns about the following:
1) Full clone onto a drive with EnCFS, with nightly rsync. Then sync the EnCFS container with SpiderOak
OR
2) Full clone onto a drive with FDE, and keep that offsite. Then copy of most important files (there are a lot!) onto a drive and into an EnCFS container, with nightly rsync. Then sync the EnCFS container with SpiderOak.

Essentially: is it bad to go all the way and keep the entire clone in the EnCFS container and rsync that each night + sync to Spider, rather than just doing it for important files?

Thanks very much for any feedback or guidance.