Re: [Duplicity-talk] Redundancy?

[edgar . soldin]

On 23.05.2013 18:35, Mikołaj Czyż wrote:
> Hello,
> 
> Are there any works on adding some form of archive redundancy to duplicity?

nobody's working on it. but it is requested from time to time.
 
> If I understand correctly one bit-flip can destroy the whole duplicity 
> backup, am I right?
> 

no, it "just" obstructs you from restoring files located in incremental volumes 
after (by vol number or time) that volume as duplicity will probably break 
everytime it tries to decrypt/unpack that volume. there is however a 
possibility to ignore errors, where you will have to look into the log output 
to find what was restored to which state, and what not.

additionally there is manual reconstruction routine. but essentially, all files 
that had incremental data in that volume can only properly be restored until 
one backup before the corrupt volume.

concluding, there is generally a need for recovery redundancy but no one 
tackled it so far. easiest would probably be utilizing par to checksum all 
files of one backup.


..ede/duply.net