problem with retriving NZB's (Trying to fetch NZB from XXXXXXXXX)

[weekender]

I spoke with Flash, the dev of nzbmatrix and he seems to have changed some stuff to fix the issue.
Since then, grabbing nzbs from their website's RSS has been working fine for me.

https://nzbmatrix.com/forums.php?action ... 770&page=1

[MiG]

Bump. Getting the same behaviour from both TVnzb.org and NZBs.org at the moment. The first one is probably just them being inaccessible again, but oddly enough NZBs.org RSS feeds seem to result in SABnzbd trying to fetch non existent URLs?

example: http://nzbs.org/index.php_action=getnzb&nzbid=330179

Deleted the SABnzbd cache, started it and allowed it to fetch a fresh feed (from http://nzbs.org/rss.php?catid=14). The link above resulted from that.



EDIT: found the problem. The first link is what SAB tries to fetch, the second is what I grab from nzbs.org itself.
http://nzbs.org/index.php_action=getnzb&nzbid=330179
http://nzbs.org/index.php?action=getnzb&nzbid=330179

Somehow an underscore ends up where a question mark should be.

[shypike]

This only what the UI shows, the question mark is sent to nzbs.org.
The display part uses the filename cleanup filter.
So far I have never had any problems with nzbs.org,
which makes it hard to fix anything.

[MiG]

I seem to have overlooked an authentication key for the NZBs.org RSS feeds, but unfortunately that still resulted in the same behaviour.

Uninstalling then reinstalling (0.5.1RC2) seems to have sorted it though, will keep adding my old settings until it either breaks or continues to work normally (i.e. the cause having previously been file corruption somewhere)

[Yarbles]

I was encountering this error too and figured out it was because I don't have an Nzbmatrix API key setup (because I don't have a VIP account). I don't have an Matrix RSS set up, but I did try adding some nzbs with nzbStatus, not realizing it wouldn't work without a key. Those items just remained in the queue "Fetching ... " and the log has the restarting crashed urlgrabber error.

I manually removed the items from the queue and then downloaded the nzb files to the blackhole, which worked fine. BUT I didn't notice until later that "Restarting crashed urlgrabber" must not have worked because even my nzbs.org RSS items stayed "fetching" until a restart, after which they were fetched fine.

[shypike]

Without a paid account you can only get NZB files from the website.
SABnzbd gets no access without the apikey.
The endless retries were put in because often nzbmatrix is very unresponsive.

[Yarbles]

Maybe SABnzbd shouldn't bother fetching if there isn't a Matrix API key in the config?

The reason I posted was just to let you and others know that the missing key problem could effect other fetch attempts, even those not fetching from nzbmatrix. At least that appeared to be what happened to SABnzbd until I restarted.

[Phunky]

Without a paid account you can only get NZB files from the website.
SABnzbd gets no access without the apikey.
The endless retries were put in because often nzbmatrix is very unresponsive.

I have a paid account and still get this problem - is there anything I can supply you with so you can debug this a bit more?

[shypike]

Maybe SABnzbd shouldn't bother fetching if there isn't a Matrix API key in the config?

SABnzbd logs an error and kills the job.

[shypike]

I have a paid account and still get this problem - is there anything I can supply you with so you can debug this a bit more?

First test the feed URL by pasting it into the browser's address bar (include username and apikey parms).
You'll get the feed content , then click on a nzb fetch link and see what happens.

[Phunky]

It works fine because I actually copy it from sabnzb's "attempting" bit in the que and it downloads fine - also noticed this happening for some other RSS feeds too.

[bbaldino]

I've got this same problem (on 0.5.2) with nzbs.org, and no files from the RSS will properly download.  If I copy the link it's trying to grab and paste it into chrome, the nzb downloads without any problem...but sabnzbd can never seem to pull it down.  I think this is the interesting part of the log, but let me know if you want more.  Happens to me every time. (Specific nzb ID yanked)

Code: Select all

2010-05-12 19:06:33,967::INFO::[_cplogging:55] [12/May/2010:19:06:33] ENGINE Started monitor thread '_TimeoutMonitor'.
2010-05-12 19:06:34,069::INFO::[_cplogging:55] [12/May/2010:19:06:34] ENGINE Serving on 127.0.0.1:8080
2010-05-12 19:06:34,070::INFO::[_cplogging:55] [12/May/2010:19:06:34] ENGINE Bus STARTED
2010-05-12 19:06:34,071::INFO::[misc:490] Lauching browser with http://localhost:8080/sabnzbd
2010-05-12 19:06:34,457::WARNING::[nzbstuff:572] Invalid NZB file index.php!action=getnzb&nzbid=XXXXXX.nzb, skipping (reason=NzbParser instance has no attribute 'groups', line=0)
2010-05-12 19:06:34,458::ERROR::[misc:833] Error getting url http://nzbs.org/index.php?action=getnzb&nzbid=XXXXXX (cause= "")
2010-05-12 19:07:07,238::INFO::[__init__:751] Restarting crashed urlgrabber
2010-05-12 19:07:20,094::INFO::[__init__:641] Saving data for rss_data.sab in /home/brian/.sabnzbd/cache/rss_data.sab
2010-05-12 19:07:20,322::INFO::[newzbin:366] Fetching Newzbin bookmarks
2010-05-12 19:07:40,272::INFO::[__init__:751] Restarting crashed urlgrabber
2010-05-12 19:08:13,308::INFO::[__init__:751] Restarting crashed urlgrabber
2010-05-12 19:08:46,344::INFO::[__init__:751] Restarting crashed urlgrabber

[shypike]

We discovered a problem in the error handling of NZB retrieval.
Do you have an explicit post processing value (+R, +U, +D) set for individual feeds
or filters in the feed. Maybe in the categories used?
If you always have "Default" for all post-processing, the problem should not occur.

This will be solved in 0.5.3

[bbaldino]

Yep...I've got +D set by default. I'll set the default to nothing and give it a try.

EDIT: I went into the defaults and change the processing to none, and the setting I have in the category is default...but the same thing still happened.  Is there a setting somewhere I missed?  The processing drop down on the queue page was empty.  The logs do appear different, though.  Here's a snippet:

Code: Select all

2010-05-13 20:40:07,879::INFO::[articlecache:143] Flushing <Article: [email protected], bytes=393740, partnum=27, art_id=SABnzbd_article_3ZGRdb> to disk
2010-05-13 20:40:07,880::INFO::[__init__:641] Saving data for SABnzbd_article_3ZGRdb in /home/XXXXX/.sabnzbd/cache/SABnzbd_article_3ZGRdb
2010-05-13 20:40:07,884::INFO::[decoder:92] Decoding <Article: [email protected], bytes=393832, partnum=32, art_id=None>
2010-05-13 20:40:07,893::INFO::[articlecache:143] Flushing <Article: [email protected], bytes=393832, partnum=32, art_id=SABnzbd_article_QG2Hrk> to disk
2010-05-13 20:40:07,894::INFO::[__init__:641] Saving data for SABnzbd_article_QG2Hrk in /home/XXXX/.sabnzbd/cache/SABnzbd_article_QG2Hrk
2010-05-13 20:40:07,898::INFO::[__init__:662] Loading data for SABnzbd_article_tBmPyi from /home/XXXX/.sabnzbd/cache/SABnzbd_article_tBmPyi

A message like the one above basically repeats over and over.

[shypike]

I don't think I explained it properly.
You can use the +D setting.
In Config->Switches, set the default value to +D.
In all other places (categories and feeds) you should select "Default".
The crash occurs when SABnzbd tries to create a retry link to send to the History page.
It will fail if an explicit post-proc value is set anywhere, except in Config->Switches.