I have one request that I would find cool. I think it could be hard to implement, but maybe it is possible to add when implementing backup server priority.
Most people that use a backupserver have a main server for which they have some sort of flatrate and one or more backup severs where every download costs additional money (blockaccount).
So what would be a good possibility IMO is a option to use the backupserver as little as possible. At the moment, if a release is incomplete at my main server but available at my backup, sabnzbd loads what is available from the main server and the rest from the backup.
As an option, it would be nice to be able to set it so that it loads all files and par files available at the main server and then only what is stil needed to repair the files from the backup server.
This would use less traffic from the backup server (=less costs) at the cost of more time and CPU ressources to repair the file.
Prefer par2 over backup servers?
Re: Prefer par2 over backup servers?
That's a) very complicated and b) not very likely to save that much data.
The described solution on the roadmap is already challenging (from a perspective of required implementation time).
The described solution on the roadmap is already challenging (from a perspective of required implementation time).
Re: Prefer par2 over backup servers?
Wouldn't it be enough to download all available data like there is no backup server available and then load what is needed for par2 repair from the backup server(s)?shypike wrote:That's a) very complicated and
Re: Prefer par2 over backup servers?
Not really, repeatability cannot predicted accurately due to shortcomings of
the NZB format and the fact that articles and par2 blocks do not match at all.
Chances are that you end up with unrepairable set of files and it is very very hard
to fill missing articles in files later on.
Like I said, it is just too much complexity for too little gain. Not going to happen.
the NZB format and the fact that articles and par2 blocks do not match at all.
Chances are that you end up with unrepairable set of files and it is very very hard
to fill missing articles in files later on.
Like I said, it is just too much complexity for too little gain. Not going to happen.