Quote:
Originally Posted by dnanian
Given the cost of disks these days, though, it seems like giving yourself a bit more headroom in the meantime would be both cost and time effective...
|
This is my usual practice. However, all three disk failures that have threatened my data have involved laptop drives that died without warning. I have bought the largest drive that will fit in my laptop, already, and I am cloning to an emergency spare. This is "sneakernet mirrored raid"; if I could find a larger 2.5" spare, I'd buy two and use one as my primary drive.
So more headroom is NOT an option, for my most critical application of SD!
If this is rare, then efficiency is not a consideration. What should SD! do if it hits this error?
1. Give up in despair.
2. Start a new pass where it only deletes, then try again with a normal pass.
I am a programmer, and I can't imagine this being more than 20 lines of code, with no impact on performance until the alternative would be 1.