[Guest]

Luckily i run linux and zfs  btw if you check the smart data on your hard disk periodically and see that blocks are reallocated then you are almost guaranteed that at least one block of a file was zeroed (corrupted).

[popo]

Disc C is an SSD and it became corrupted!  Now I put any image on separate HDs and link it to the desktop with a shortcut.

I'd be more concerned about how it got corrupted and preventing a repeat in future. How do you know the other hard disks couldn't get similarly corrupted?

 

I decided this approach when I discovered that an old MBP had silently corrupted a few photos - fortunately nothing important. Disk checks didn't detect the problem: you had to open the files to see it.

 

Some advanced filesystems, such as ZFS, do basically the same kind of check in background, but they are not available on Mac OS X or Windows. This feature might be included in a future new filesystem for Mac OS, but I'm not sure about when.

I only had corrupted files once, eventually traced down to ram instability. Before anyone says run memtest, it didn't detect it. The frequency of the error was of the magnitude of once a week to one a quarter, and unless you run memtest 24/7 for a similar period, you're unlikely to find it. It was only found as I do continuous computation in the background, and I managed to manually recover the corrupted files from a backup. For this reason, I'm thinking my next main system will have ECC ram for an extra layer of safety.

 

ZFS on paper sounds great. For a platform agnostic way to have its safety features, I'm thinking of a NAS using it, but I never got around to implementing it.