SABnzbd Forums

Looks like sabnzbd picks up quite randomly one of the par2 files
Eg. from the set:
abc.par2 abc.vol1+1.par2 .. abc.vol14+56.par2

par2 r abc.vol14+56.par2

is run. It happens quite often, that while abc.par2 is ok, some of other par2 are corrupted - in such case par2 fails to rocover the set.

Would it be possible to always try abc.par2 first instead?

It tries the smallest par2 file first.
abc.par2 is ignored altogether.
Old design flaw (from the original author) which we haven't tackled yet.

BTW, even abc.par2 may be corrupted. So some more action is needed.

When a corrupt par2 file is found, sabnzbd should then download another par2 file. I'm not sure why this wouldn't be happening, maybe it depends on what gets corrupted.

Ive been looking into this aswell, as unnecessary PAR2 volumes increase overhead for downloads when monthly quota is an issue.

Experienced just before a small nzb processing (30mb in total) and sabnzbd fetched all vol files that were <1mb in size (5 out of the 7 vol sets)

Previously I had observed whats mentioned above, where the smallest vol file was included in the queue.

Would very much prefer a process where the .par2 only is fetched, processed & IF repair blocks are needed, only then are they readded to the queue.

This is due to having many connections.
SABnzbd will try to use all slots, so it ends up downloading many small files at once.
The actual amount of extra data is rather low, as the smallest are tried first.