Selective NZB download will not autoqueue

[ziddey]

Version: 0.5.6
OS: Windows 7 x64
Install-type: Windows Installer
Skin: Smpl
Firewall Software: Windows 7 Firewall
Are you using IPV6? IPV6 and IPV4
Is the issue reproducible? yes

I remember reading that this issue was resolved in the past, but I might be mistaken. In any case, I have an RSS feed set up to download NZBs. For most instances, the NZB will download, and will immediately be introduced into queue with the same criteria as specified in the RSS feed (e.g. TV category). It will subsequently download as usual. I prefer this method over an RSS feed for the actual video, since a lot of times, the par set will be a separate listing, and allow for the possibility of crc errors. In comparison, the posted nzb will include everything.

I'm not sure on etiquette in terms of posting specific examples, but there is one show in particular, where it will download the nzb, and then it will just sit in the incomplete folder. I then have to manually load the nzb, and it downloads as usual.

I am using one rss feed from nzbindex.nl, and have filters for individual shows. It works extremely well in most circumstances.

I am seeing now that 0.6.0 has been released, so I will try that and see how it works, since I understand the RSS has been revamped (at least front end?).

[shypike]

I'm not sure I understand the issue.
Please keep us posted with your experiences with 0.6.0

[ziddey]

Thanks, will do. The next time the typical weekly failure will be next Sunday. Basically, I have a nzbindex.nl feed of nzbs (for an efnet release of a show say, there'll be a set that usually includes the nfo, rars, pars, another set for the sample, and another set that's just the nzb as posted by whatever posting agent). So my feed is only for the nzb's. Sabnzbd downloads the nzb from nzbindex via the rss feed. The contents of that nzb is simply the nzb created by the posting agent of the releaser. Then, usually, sabnzbd automatically requeues that nzb it just downloaded under the same criteria as the original nzb it downloaded from the index site. But occasionally, the nzb it downloads from the nzb will be stuck in the incomplete folder, and will never be queued up.

I know this will clarify things entirely, but if it's not kosher, feel free to remove it:

Feed: http://rss.binsearch.net/rss.php?max=50 ... multimedia
I then have a filter to require: #a.b.teevee@EFNet*xvid*hdtv*nzb
From then on, I just need to add individual filters to accept each show I want.

So this only downloads the nzb from binsearch to download the nzb as posted by the uploader. Then, that nzb it downloads should be automatically parsed again, and that's when the "actual" download begins.

[shypike]

There is a known problem in 0.5.6: it doesn't know what to do with an empty NZB file
and it will be stuck in the queue.
0.6.0 will refuse an empty NZB.

There may be some issue with chained NZBs in 0.5.6,
but I cannot recall the details currently.

[ziddey]

The nzb is not empty. I can add the nzb that's in the incomplete folder and everything will download as is usual.

Chained nzb sounds like what I'm talking about though. It works most of the time. I haven't figured out what the common factor is when it doesn't though. So far, I've only found that one show repeatedly reproduces the problem. Not sure if it's a file length issue or some other issue with how the poster made the nzb.

I suppose I'll switch my feed setup so it won't use the chained-nzb approach, since it leaves behind empty folders from the first download of only the nzb. I remember I did it that way to begin with since sometimes the par files are not grouped with the rar files on index sites. The chained-nzb approach ensures the par files will be accessible if needed.

and as it goes, I can't reproduce it at the moment.

looking through the history, when everything goes according to plan, it'll show:

Download:
Downloaded in 2 seconds at an average of 25.12 KB/s
Sent ['abcdefg.nzb'] to queue

vs when it just downloads the nzb and leaves it in the incomplete folder:

Download:
Downloaded in 3 seconds at an average of 51.96 KB/s

then I'll have to manually add the chain-downloaded nzb


According to the path, it saves the nzb to the target location and then sends it to the queue. But in instances when I have to manually do it, the nzb will still be in the incomplete folder. My suspicion is that the filename/folder/nesting is too long and beyond the limits of ntfs, so the nzb is never able to be moved to the completed location, so sabnzbd is unable to find the nzb to resend it. We're talking in excess of 200 characters for folders and filenames.

So in conclusion, it's not a problem with sabnzbd at all. Sorry I didn't look into it more before posting.

[shypike]

There is a (hidden) option in both 0.5.6 and 0.6.0
that allows you to set a maximum size for path elements.
This may help you.
See: http://wiki.sabnzbd.org/configure-special element "folder_max_length".