On 2014-03-30 17:16, Carl Hartung wrote:
On Sun, 30 Mar 2014 16:53:11 +0200 Carlos E. R. wrote:
:-) You forgot this one: http://en.wikipedia.org/wiki/Murphy%27s_law
LOL, yes.
In my experience, there seems to be an inverse relationship between how critical one's data is and how long it can hang around uncorrupted and undisturbed.
I managed to corrupt both the original and backup copies of a paper, in different floppies, that I had to hand to my teacher for examination, hours before the time limit. Yes, I know about Murphy...
The more important it is, the more frequently you should be backing it up. IOW, frequent and redundant backups -- serving both archival and operational needs -- is for all practical purposes the only defense.
The first link explains an interesting feature of btrfs native raid, however it is called. It can automatically repair a file with errors. Traditional raid can not. I'd like a similar repair capability self contained in normal files, without the need for raid type structures (double or triple hardware). Just with "forward recovery data" could work. At least on (long term) backups. -- Cheers / Saludos, Carlos E. R. (from 13.1 x86_64 "Bottle" at Telcontar)