url fetching failed

[damez2000]

Yes, when i have nzb association with sab, and try to download an nzb from nzbindex.nl, it is interrupted. Not a fail, but interrupts it. If i use a browser such as IE, which I dont normally use, the nzb downloads fine. If i take off the association, the browser downloads the nzb fine. Has anyone ever seen the picture from this link before? And if so, how do you fix it. https://docs.google.com/leaf?id=0B19uJi ... 0&hl=en_US

[shypike]

It looks like something is wrong in the browser itself.
Chrome?
Have you tried re-installing it?

[damez2000]

If i uninstall sab, and then download from nzbindex, i get no interrupted. Maybe the nzb association isnt that great with sab. Maybe the watched folder is a better way to go.

[damez2000]

in fact, i just uninstalled, reinstalled, cleared registry, and same problem. Is anyone else having this issue with chrome?

[shypike]

Is this the standard Chrome of the Dev channel?
We don't support the Dev channel, because that's a moving target.

[damez2000]

Stable version 14 of chrome.

[Shane112358]

I am having the same problem. It only happens when I get an email from the NZBmatrix release watcher. I click the bookmark link to add it to my bookmarks RSS, and SabNZBd is supposed to pull the NZB from the RSS and download it. But it has a ~50% failure rate. If I log into SABnzbd I can click on retry and that will work, but it is annoying.

The weird thing is, if I add something to my bookmarks directly in NZBmatrix, SABnzbd pulls it successfully 99% of the time. It's only on the release watcher email links.

Any suggestions?

[Shane112358]

Here is the error,

URL Fetching failed; (Problem accessing nzbmatrix server (error:no_nzb_found)),

[shypike]

It's an nzbmatrix problem.
It happens frequently that NZB files cannot be retrieved through the API
while they are available through the web site.
Sometime this is temporary, sometimes a permanent problem.

[Shane112358]

Thanks - is there any way to get around this problem? I'm surprised that Sabnzbd doesn't have a feature where you can tell it to automatically try a failed fetch (up to say, 3 times, 5 minutes apart) so I don't have to constantly go in and see if the fetch succeeded or not.

[shypike]

It does retry quite a bit, but we're not going to compensate
for a basically unreliable service.

[Shane112358]

It does retry quite a bit, but we're not going to compensate
for a basically unreliable service.

...nor would I expect you to - but given the popularity of the service it seems advantageous to have that feature in there nonetheless. I'm surprised to hear that it does retry - I'm assuming several times within a short time, before it displays the corresponding error. I was thinking multiple attempts that are more temporally spaced, like 5 or 10 minutes apart. Because then at least I would be much more confident that one of those attempts would catch the API working - perhaps 95% instead of 50%. The only reason I say this is most of the time (90+%) manually retrying one time in the Sabnzbd interface does the trick.

[shypike]

SABnzbd doesn't retry when it gets the "not present" message,
because you're not supposed to hammer the nzbmatrix API after receiving it.
The "temporary" situation may last hours, the permanent one forever.
So how much retries are reasonable?
There is no good choice.

[bytemuncher]

So how much retries are reasonable?
There is no good choice.

I think a good choice would be to retry once every hour for three hours, then give up. That is hardly hammering.

If there is not going to be a retry option then is it possible to send an email notification when the "URL Fetching failed; (), Try again " error is returned?

Everything else is automated, but because of this I need to log in and manually check the logs to see if something was skipped over.

[krisg1984]

I'd hoped this problem would get fixed at either the nzbmatrix end or the sabnzbd end, but it's been a few months and still annoying the hell out of me.
I've written a little command line app that queries the history for errors and re-adds the bad jobs to the queue. You can use Windows Scheduler to run it however often you like.

If anyone's interested in the fix you can find it here...
https://github.com/krisg1984/SABnzbdRet ... /downloads