SABnzbd Forums

Currently, as far as I understand, SABnzbd checks the backup servers for missing articles, and if it can not find them, it logs “missing from all servers, discarding”, and completes the file without (!) that one article.

Well, this is certainly not a problem if there are enough PAR2 files available, but it becomes a problem if the poster doesn’t post any PAR2 files. In that latter case you are left with incomplete (and mostly useless) RAR partitions without an option (?) to “rescan for/redownload the missing articles” after you find new servers which might have the missing articles.

Currently I work around this limitation by filling in a dummy server as “backup server”. SABnzbd blocks on it and doesn’t finish the file, which gives me the time to search for alternate servers.

My question is: are there better ways to solve my problem?

If not, I would like to suggest hereby a feature, like a checkbox or something for each download, which would block the completion of files (i.e. merging from cache to temp folder) if they have missing articles.

Kind regards,
Davor

SABnzbd's purpose is to automate downloading as much as possible.
If this is the way you need tp scrape your downloads together,
I'm not sure if SABnzbd is the right solution.
The next release will have retry options, but not at the article level.

You say "gives me time to search for alternative servers".
Why don't you just list those as backup servers?

It's not about scrapping the downloads together, it's about automating their downloading with the focus on good results. And this has everything to do with SABnzbd.

Why? Because there is no point in downloading the RAR partitions if it is obvious from the beginning that it will not be possible for them to be fixed (among others if there are no PAR2 partitions available), like when there are missing articles, CRC failures, etc, right?

If this “proper downloading” already can be achieved with some workarounds, then a descent solution is probably not that hard to implement. A checkbox (on file and/or collection level) that constrains the completion of the file if and only if all articles are available would be a descent solution, no? You can automate this decision task (with a default) or leave it to the user, like you leave it to the user to choose the category.

Rescanning/redownloading simply shows that the automation process has issues. In other words, if the downloading process was flawless, then I shouldn’t have to rescan/redownload/scrape/etc. So I leave it open what SABnzbd would focus on: proper downloading or download fixing.

shypike wrote: You say "gives me time to search for alternative servers".
Why don't you just list those as backup servers?

If there is no prospect on fixing the file, and none of the servers have the right articles, then I search for servers that might have them. For example I might pay for temporary access to a server which might have the required articles.

PS I think SABnzbd is a wonderful tool. The best, or at least one of the best available. I take time to write this because I believe it can be made even better.

Implementing your suggestions in a useful way would require substantial redesign of the current software.
Everything is oriented towards completing files a.s.a.p. and letting par2 take care of repairs.

There will be a future redesign that will do away with the article cache and instead do
direct writes to the files as the articles come in.
This redesign may be augmented with some of your suggestions.
This is quite some time away.
First we concentrate on finally finishing 0.6.0 and after that a quality drive addressing
the memory problems reported for some platforms.