Re: Have you considered using the "multiple data files" mode of par2?

Yes, I didn't mean to imply that "shared parity files mode", which had 
been suggested, is being used.

> still, per duplicity file (volume, manifest etc.) put on/got from the 
> backend.
> 
> see source
>   https://gitlab.com/duplicity/duplicity/-/blob/dev/duplicity/backends/ 
> par2backend.py?ref_type=heads#L147-176
> 
> looks like additional volumes help to repair corrupted files.
> 
> ..sunny regards ede
> 
> On 07.10.2024 16:04, Jasmin68k via Duplicity-talk wrote:
>> The number of par2 files is configurable (default: 1) with the --par2- 
>> volumes option btw.
>>
>>> one *parity file per file that is put* on the backend
>>>
>>> On 07.10.2024 15:36, edgar.soldin--- via Duplicity-talk wrote:
>>>> hey Jacob,
>>>>
>>>> i'm sure par2-backend only uses the one parity file tat is put on 
>>>> the backend. what are you hinting at?
>>>>
>>>> sunny regards ..ede
>>>>
>>>> On 07.10.2024 15:22, Jakob Bohm via Duplicity-talk wrote:
>>>>> Dear Duplicity backup format designer,
>>>>>
>>>>> According to the par2 man page (at least on Debian 12), the par2 
>>>>> command
>>>>> has two distinct operating modes.  One mode creates one or more small
>>>>> parity files for a single data file stored on an imperfect medium or
>>>>> service.  The other, processes a group of N data files (such as
>>>>> Duplicity volumes) to produce a few shared parity files.  This second
>>>>> mode is described in the man page under the subheading "CREATING PAR2
>>>>> FILES FOR MULTIPLE DATA FILES" .
>>>>>
>>>>> Switching to the "multiple data files" mode constitutes a change in
>>>>> the backup format and thus needs to be guarded with appropriate
>>>>> compatibility logic in order to restore backups made with either
>>>>> method, even from damaged storage (where par2 files are actually
>>>>> needed) .  To keep a desired robustness level and a manageable temp
>>>>> volume size on the client, it is probably necessary to set an internal
>>>>> limit on the number of volumes for each par2 run .  For example if
>>>>> that limit is 99, the temp volume needs to store about 100 times the
>>>>> volume size even if not pipelining the creation of later volumes in
>>>>> parallel with par2 processing and upload.
>>>>>
>>>>> Enjoy
>>>>>
>>>>> Jakob
>>>>
>>>> _______________________________________________
>>>> Duplicity-talk mailing list
>>>> [email protected]
>>>> https://lists.nongnu.org/mailman/listinfo/duplicity-talk
>>>
>>> _______________________________________________
>>> Duplicity-talk mailing list
>>> [email protected]
>>> https://lists.nongnu.org/mailman/listinfo/duplicity-talk
>>
>>
>> _______________________________________________
>> Duplicity-talk mailing list
>> [email protected]
>> https://lists.nongnu.org/mailman/listinfo/duplicity-talk
> 
> _______________________________________________
> Duplicity-talk mailing list
> [email protected]
> https://lists.nongnu.org/mailman/listinfo/duplicity-talk


_______________________________________________
Duplicity-talk mailing list
[email protected]
https://lists.nongnu.org/mailman/listinfo/duplicity-talk