Re: [Duplicity-talk] Announ cing Déjà Dup

[Peter Schuller]

> One of the reasons: If one file in the incremental chain develops even 
> just one bit flip, the chances rather high are your whole backup is 
> useless. Another one is possibly just caution towards duplicity itself. 

However, on this topic is should be noted that if you are backing up
data whose pattern is that it is very large compared to the changes
(say 100 GB that generates 1 GB of delta per week), the statistical
probability of a bitflip disaster increases fairly slowly, since for
the first several weeks the most likely bitflip is in the original
backup.

So in terms of "best practices", one possible recommendation is to
keep at least two full backups rather than just one.

Hopefully future duplicities will have par2/fec support to help a
little here though.

-- 
/ Peter Schuller

PGP userID: 0xE9758B7D or 'Peter Schuller <address@hidden>'
Key retrieval: Send an E-Mail to address@hidden
E-Mail: address@hidden Web: http://www.scode.org

pgpoBE3Sb5Wnh.pgp
Description: PGP signature